minio/cmd/admin-handlers.go

1481 lines
41 KiB
Go
Raw Normal View History

/*
* Minio Cloud Storage, (C) 2016, 2017, 2018 Minio, Inc.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package cmd
import (
"archive/zip"
"bytes"
"context"
"encoding/base64"
"encoding/json"
"errors"
"fmt"
"io"
"net/http"
"os"
"strconv"
"strings"
"sync"
"time"
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
"github.com/gorilla/mux"
"github.com/minio/minio/cmd/logger"
"github.com/minio/minio/pkg/auth"
"github.com/minio/minio/pkg/disk"
"github.com/minio/minio/pkg/handlers"
"github.com/minio/minio/pkg/iam/policy"
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
"github.com/minio/minio/pkg/madmin"
"github.com/minio/minio/pkg/quick"
"github.com/tidwall/gjson"
"github.com/tidwall/sjson"
)
const (
maxEConfigJSONSize = 262272
)
// Type-safe query params.
type mgmtQueryKey string
// Only valid query params for mgmt admin APIs.
const (
mgmtBucket mgmtQueryKey = "bucket"
mgmtPrefix = "prefix"
mgmtClientToken = "clientToken"
mgmtForceStart = "forceStart"
mgmtForceStop = "forceStop"
)
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
var (
// This struct literal represents the Admin API version that
// the server uses.
adminAPIVersionInfo = madmin.AdminAPIVersionInfo{
Version: "1",
}
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
)
// VersionHandler - GET /minio/admin/version
// -----------
// Returns Administration API version
func (a adminAPIHandlers) VersionHandler(w http.ResponseWriter, r *http.Request) {
ctx := newContext(r, w, "Version")
adminAPIErr := checkAdminRequestAuthType(ctx, r, "")
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
if adminAPIErr != ErrNone {
writeErrorResponseJSON(w, adminAPIErr, r.URL)
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
return
}
jsonBytes, err := json.Marshal(adminAPIVersionInfo)
if err != nil {
writeErrorResponseJSON(w, toAdminAPIErrCode(ctx, err), r.URL)
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
return
}
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
writeSuccessResponseJSON(w, jsonBytes)
}
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
// ServiceStatusHandler - GET /minio/admin/v1/service
// ----------
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
// Returns server version and uptime.
func (a adminAPIHandlers) ServiceStatusHandler(w http.ResponseWriter, r *http.Request) {
ctx := newContext(r, w, "ServiceStatus")
adminAPIErr := checkAdminRequestAuthType(ctx, r, "")
if adminAPIErr != ErrNone {
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
writeErrorResponseJSON(w, adminAPIErr, r.URL)
return
}
// Fetch server version
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
serverVersion := madmin.ServerVersion{
Version: Version,
CommitID: CommitID,
}
2017-02-08 03:13:02 -05:00
// Fetch uptimes from all peers. This may fail to due to lack
// of read-quorum availability.
uptime, err := getPeerUptimes(globalAdminPeers)
if err != nil {
writeErrorResponseJSON(w, toAdminAPIErrCode(ctx, err), r.URL)
2017-02-08 03:13:02 -05:00
return
}
// Create API response
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
serverStatus := madmin.ServiceStatus{
ServerVersion: serverVersion,
2017-02-08 03:13:02 -05:00
Uptime: uptime,
}
// Marshal API response
jsonBytes, err := json.Marshal(serverStatus)
if err != nil {
writeErrorResponseJSON(w, toAdminAPIErrCode(ctx, err), r.URL)
return
}
// Reply with storage information (across nodes in a
// distributed setup) as json.
writeSuccessResponseJSON(w, jsonBytes)
}
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
// ServiceStopNRestartHandler - POST /minio/admin/v1/service
// Body: {"action": <restart-action>}
// ----------
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
// Restarts/Stops minio server gracefully. In a distributed setup,
// restarts all the servers in the cluster.
func (a adminAPIHandlers) ServiceStopNRestartHandler(w http.ResponseWriter, r *http.Request) {
ctx := newContext(r, w, "ServiceStopNRestart")
adminAPIErr := checkAdminRequestAuthType(ctx, r, "")
if adminAPIErr != ErrNone {
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
writeErrorResponseJSON(w, adminAPIErr, r.URL)
return
}
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
var sa madmin.ServiceAction
err := json.NewDecoder(r.Body).Decode(&sa)
if err != nil {
logger.LogIf(ctx, err)
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
writeErrorResponseJSON(w, ErrRequestBodyParse, r.URL)
return
}
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
var serviceSig serviceSignal
switch sa.Action {
case madmin.ServiceActionValueRestart:
serviceSig = serviceRestart
case madmin.ServiceActionValueStop:
serviceSig = serviceStop
default:
writeErrorResponseJSON(w, ErrMalformedPOSTRequest, r.URL)
logger.LogIf(ctx, errors.New("Invalid service action received"))
return
}
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
// Reply to the client before restarting minio server.
writeSuccessResponseHeadersOnly(w)
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
sendServiceCmd(globalAdminPeers, serviceSig)
}
2017-02-15 13:45:45 -05:00
// ServerProperties holds some server information such as, version, region
// uptime, etc..
type ServerProperties struct {
Uptime time.Duration `json:"uptime"`
Version string `json:"version"`
CommitID string `json:"commitID"`
Region string `json:"region"`
SQSARN []string `json:"sqsARN"`
}
// ServerConnStats holds transferred bytes from/to the server
type ServerConnStats struct {
TotalInputBytes uint64 `json:"transferred"`
TotalOutputBytes uint64 `json:"received"`
Throughput uint64 `json:"throughput,omitempty"`
}
// ServerHTTPMethodStats holds total number of HTTP operations from/to the server,
// including the average duration the call was spent.
type ServerHTTPMethodStats struct {
Count uint64 `json:"count"`
AvgDuration string `json:"avgDuration"`
}
// ServerHTTPStats holds all type of http operations performed to/from the server
// including their average execution time.
type ServerHTTPStats struct {
TotalHEADStats ServerHTTPMethodStats `json:"totalHEADs"`
SuccessHEADStats ServerHTTPMethodStats `json:"successHEADs"`
TotalGETStats ServerHTTPMethodStats `json:"totalGETs"`
SuccessGETStats ServerHTTPMethodStats `json:"successGETs"`
TotalPUTStats ServerHTTPMethodStats `json:"totalPUTs"`
SuccessPUTStats ServerHTTPMethodStats `json:"successPUTs"`
TotalPOSTStats ServerHTTPMethodStats `json:"totalPOSTs"`
SuccessPOSTStats ServerHTTPMethodStats `json:"successPOSTs"`
TotalDELETEStats ServerHTTPMethodStats `json:"totalDELETEs"`
SuccessDELETEStats ServerHTTPMethodStats `json:"successDELETEs"`
}
// ServerInfoData holds storage, connections and other
// information of a given server.
type ServerInfoData struct {
2017-02-15 13:45:45 -05:00
StorageInfo StorageInfo `json:"storage"`
ConnStats ServerConnStats `json:"network"`
HTTPStats ServerHTTPStats `json:"http"`
2017-02-15 13:45:45 -05:00
Properties ServerProperties `json:"server"`
}
// ServerInfo holds server information result of one node
type ServerInfo struct {
Error string `json:"error"`
Addr string `json:"addr"`
Data *ServerInfoData `json:"data"`
}
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
// ServerInfoHandler - GET /minio/admin/v1/info
2017-02-15 13:45:45 -05:00
// ----------
// Get server information
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
func (a adminAPIHandlers) ServerInfoHandler(w http.ResponseWriter, r *http.Request) {
ctx := newContext(r, w, "ServerInfo")
2017-02-15 13:45:45 -05:00
// Authenticate request
// Setting the region as empty so as the mc server info command is irrespective to the region.
adminAPIErr := checkAdminRequestAuthType(ctx, r, "")
2017-02-15 13:45:45 -05:00
if adminAPIErr != ErrNone {
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
writeErrorResponseJSON(w, adminAPIErr, r.URL)
2017-02-15 13:45:45 -05:00
return
}
// Web service response
reply := make([]ServerInfo, len(globalAdminPeers))
2017-02-15 13:45:45 -05:00
var wg sync.WaitGroup
2017-02-15 13:45:45 -05:00
// Gather server information for all nodes
for i, p := range globalAdminPeers {
wg.Add(1)
2017-02-15 13:45:45 -05:00
// Gather information from a peer in a goroutine
go func(idx int, peer adminPeer) {
defer wg.Done()
2017-02-15 13:45:45 -05:00
// Initialize server info at index
reply[idx] = ServerInfo{Addr: peer.addr}
serverInfoData, err := peer.cmdRunner.ServerInfo()
if err != nil {
reqInfo := (&logger.ReqInfo{}).AppendTags("peerAddress", peer.addr)
ctx := logger.SetReqInfo(ctx, reqInfo)
logger.LogIf(ctx, err)
reply[idx].Error = err.Error()
return
}
reply[idx].Data = &serverInfoData
}(i, p)
2017-02-15 13:45:45 -05:00
}
wg.Wait()
2017-02-15 13:45:45 -05:00
// Marshal API response
jsonBytes, err := json.Marshal(reply)
2017-02-15 13:45:45 -05:00
if err != nil {
writeErrorResponseJSON(w, toAdminAPIErrCode(ctx, err), r.URL)
2017-02-15 13:45:45 -05:00
return
}
2017-02-15 13:45:45 -05:00
// Reply with storage information (across nodes in a
// distributed setup) as json.
writeSuccessResponseJSON(w, jsonBytes)
}
// ServerDrivesPerfInfo holds informantion about address, performance
// of all drives on one server. It also reports any errors if encountered
// while trying to reach this server.
type ServerDrivesPerfInfo struct {
Addr string `json:"addr"`
Error string `json:"error,omitempty"`
Perf []disk.Performance `json:"perf"`
}
// PerfInfoHandler - GET /minio/admin/v1/performance?perfType={perfType}
// ----------
// Get all performance information based on input type
// Supported types = drive
func (a adminAPIHandlers) PerfInfoHandler(w http.ResponseWriter, r *http.Request) {
ctx := newContext(r, w, "PerfInfo")
// Authenticate request
// Setting the region as empty so as the mc server info command is irrespective to the region.
adminAPIErr := checkAdminRequestAuthType(ctx, r, "")
if adminAPIErr != ErrNone {
writeErrorResponseJSON(w, adminAPIErr, r.URL)
return
}
vars := mux.Vars(r)
perfType := vars["perfType"]
if perfType == "drive" {
// Get drive performance details from local server's drive(s)
dp := localEndpointsPerf(globalEndpoints)
// Notify all other Minio peers to report drive performance numbers
dps := globalNotificationSys.DrivePerfInfo()
dps = append(dps, dp)
// Marshal API response
jsonBytes, err := json.Marshal(dps)
if err != nil {
writeErrorResponseJSON(w, toAdminAPIErrCode(ctx, err), r.URL)
return
}
// Reply with performance information (across nodes in a
// distributed setup) as json.
writeSuccessResponseJSON(w, jsonBytes)
} else {
writeErrorResponseJSON(w, ErrNotImplemented, r.URL)
}
return
}
// StartProfilingResult contains the status of the starting
// profiling action in a given server
type StartProfilingResult struct {
NodeName string `json:"nodeName"`
Success bool `json:"success"`
Error string `json:"error"`
}
// StartProfilingHandler - POST /minio/admin/v1/profiling/start?profilerType={profilerType}
// ----------
// Enable server profiling
func (a adminAPIHandlers) StartProfilingHandler(w http.ResponseWriter, r *http.Request) {
ctx := newContext(r, w, "StartProfiling")
adminAPIErr := checkAdminRequestAuthType(ctx, r, "")
if adminAPIErr != ErrNone {
writeErrorResponseJSON(w, adminAPIErr, r.URL)
return
}
vars := mux.Vars(r)
profiler := vars["profilerType"]
startProfilingResult := make([]StartProfilingResult, len(globalAdminPeers))
// Call StartProfiling function on all nodes and save results
wg := sync.WaitGroup{}
for i, peer := range globalAdminPeers {
wg.Add(1)
go func(idx int, peer adminPeer) {
defer wg.Done()
result := StartProfilingResult{NodeName: peer.addr}
if err := peer.cmdRunner.StartProfiling(profiler); err != nil {
result.Error = err.Error()
return
}
result.Success = true
startProfilingResult[idx] = result
}(i, peer)
}
wg.Wait()
// Create JSON result and send it to the client
startProfilingResultInBytes, err := json.Marshal(startProfilingResult)
if err != nil {
writeCustomErrorResponseJSON(w, http.StatusInternalServerError, err.Error(), r.URL)
return
}
writeSuccessResponseJSON(w, []byte(startProfilingResultInBytes))
}
// dummyFileInfo represents a dummy representation of a profile data file
// present only in memory, it helps to generate the zip stream.
type dummyFileInfo struct {
name string
size int64
mode os.FileMode
modTime time.Time
isDir bool
sys interface{}
}
func (f dummyFileInfo) Name() string { return f.name }
func (f dummyFileInfo) Size() int64 { return f.size }
func (f dummyFileInfo) Mode() os.FileMode { return f.mode }
func (f dummyFileInfo) ModTime() time.Time { return f.modTime }
func (f dummyFileInfo) IsDir() bool { return f.isDir }
func (f dummyFileInfo) Sys() interface{} { return f.sys }
// DownloadProfilingHandler - POST /minio/admin/v1/profiling/download
// ----------
// Download profiling information of all nodes in a zip format
func (a adminAPIHandlers) DownloadProfilingHandler(w http.ResponseWriter, r *http.Request) {
ctx := newContext(r, w, "DownloadProfiling")
adminAPIErr := checkAdminRequestAuthType(ctx, r, "")
if adminAPIErr != ErrNone {
writeErrorResponseJSON(w, adminAPIErr, r.URL)
return
}
profilingDataFound := false
// Initialize a zip writer which will provide a zipped content
// of profiling data of all nodes
zipWriter := zip.NewWriter(w)
defer zipWriter.Close()
for i, peer := range globalAdminPeers {
// Get profiling data from a node
data, err := peer.cmdRunner.DownloadProfilingData()
if err != nil {
logger.LogIf(ctx, fmt.Errorf("Unable to download profiling data from node `%s`, reason: %s", peer.addr, err.Error()))
continue
}
profilingDataFound = true
// Send profiling data to zip as file
header, err := zip.FileInfoHeader(dummyFileInfo{
name: fmt.Sprintf("profiling-%d", i),
size: int64(len(data)),
mode: 0600,
modTime: time.Now().UTC(),
isDir: false,
sys: nil,
})
if err != nil {
continue
}
writer, err := zipWriter.CreateHeader(header)
if err != nil {
continue
}
if _, err = io.Copy(writer, bytes.NewBuffer(data)); err != nil {
return
}
}
if !profilingDataFound {
writeErrorResponseJSON(w, ErrAdminProfilerNotEnabled, r.URL)
return
}
}
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
// extractHealInitParams - Validates params for heal init API.
func extractHealInitParams(r *http.Request) (bucket, objPrefix string,
hs madmin.HealOpts, clientToken string, forceStart bool, forceStop bool,
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
err APIErrorCode) {
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
vars := mux.Vars(r)
bucket = vars[string(mgmtBucket)]
objPrefix = vars[string(mgmtPrefix)]
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
if bucket == "" {
if objPrefix != "" {
// Bucket is required if object-prefix is given
err = ErrHealMissingBucket
return
}
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
} else if !IsValidBucketName(bucket) {
err = ErrInvalidBucketName
return
}
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
// empty prefix is valid.
if !IsValidObjectPrefix(objPrefix) {
err = ErrInvalidObjectName
return
}
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
qParms := r.URL.Query()
if len(qParms[string(mgmtClientToken)]) > 0 {
clientToken = qParms[string(mgmtClientToken)][0]
}
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
if _, ok := qParms[string(mgmtForceStart)]; ok {
forceStart = true
}
if _, ok := qParms[string(mgmtForceStop)]; ok {
forceStop = true
}
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
// ignore body if clientToken is provided
if clientToken == "" {
jerr := json.NewDecoder(r.Body).Decode(&hs)
if jerr != nil {
logger.LogIf(context.Background(), jerr)
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
err = ErrRequestBodyParse
return
}
}
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
err = ErrNone
return
}
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
// HealHandler - POST /minio/admin/v1/heal/
// -----------
// Start heal processing and return heal status items.
//
// On a successful heal sequence start, a unique client token is
// returned. Subsequent requests to this endpoint providing the client
// token will receive heal status records from the running heal
// sequence.
//
// If no client token is provided, and a heal sequence is in progress
// an error is returned with information about the running heal
// sequence. However, if the force-start flag is provided, the server
// aborts the running heal sequence and starts a new one.
func (a adminAPIHandlers) HealHandler(w http.ResponseWriter, r *http.Request) {
ctx := newContext(r, w, "Heal")
// Get object layer instance.
objLayer := newObjectLayerFn()
if objLayer == nil {
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
writeErrorResponseJSON(w, ErrServerNotInitialized, r.URL)
return
}
// Validate request signature.
adminAPIErr := checkAdminRequestAuthType(ctx, r, "")
if adminAPIErr != ErrNone {
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
writeErrorResponseJSON(w, adminAPIErr, r.URL)
return
}
// Check if this setup has an erasure coded backend.
if !globalIsXL {
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
writeErrorResponseJSON(w, ErrHealNotImplemented, r.URL)
return
}
bucket, objPrefix, hs, clientToken, forceStart, forceStop, apiErr := extractHealInitParams(r)
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
if apiErr != ErrNone {
writeErrorResponseJSON(w, apiErr, r.URL)
return
}
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
type healResp struct {
respBytes []byte
errCode APIErrorCode
errBody string
}
// Define a closure to start sending whitespace to client
// after 10s unless a response item comes in
keepConnLive := func(w http.ResponseWriter, respCh chan healResp) {
ticker := time.NewTicker(time.Second * 10)
defer ticker.Stop()
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
started := false
forLoop:
for {
select {
case <-ticker.C:
if !started {
// Start writing response to client
started = true
setCommonHeaders(w)
w.Header().Set("Content-Type", string(mimeJSON))
// Set 200 OK status
w.WriteHeader(200)
}
// Send whitespace and keep connection open
w.Write([]byte("\n\r"))
w.(http.Flusher).Flush()
case hr := <-respCh:
switch hr.errCode {
case ErrNone:
if started {
w.Write(hr.respBytes)
w.(http.Flusher).Flush()
} else {
writeSuccessResponseJSON(w, hr.respBytes)
}
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
default:
apiError := getAPIError(hr.errCode)
var errorRespJSON []byte
if hr.errBody == "" {
errorRespJSON = encodeResponseJSON(getAPIErrorResponse(apiError, r.URL.Path, w.Header().Get(responseRequestIDKey)))
} else {
errorRespJSON = encodeResponseJSON(APIErrorResponse{
Code: apiError.Code,
Message: hr.errBody,
Resource: r.URL.Path,
RequestID: w.Header().Get(responseRequestIDKey),
HostID: "3L137",
})
}
if !started {
setCommonHeaders(w)
w.Header().Set("Content-Type", string(mimeJSON))
w.WriteHeader(apiError.HTTPStatusCode)
}
w.Write(errorRespJSON)
w.(http.Flusher).Flush()
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
}
break forLoop
}
}
}
// find number of disks in the setup
info := objLayer.StorageInfo(ctx)
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
numDisks := info.Backend.OfflineDisks + info.Backend.OnlineDisks
healPath := pathJoin(bucket, objPrefix)
if clientToken == "" && !forceStart && !forceStop {
nh, exists := globalAllHealState.getHealSequence(healPath)
if exists && !nh.hasEnded() && len(nh.currentStatus.Items) > 0 {
b, err := json.Marshal(madmin.HealStartSuccess{
ClientToken: nh.clientToken,
ClientAddress: nh.clientAddress,
StartTime: nh.startTime,
})
if err != nil {
writeErrorResponseJSON(w, toAdminAPIErrCode(ctx, err), r.URL)
return
}
// Client token not specified but a heal sequence exists on a path,
// Send the token back to client.
writeSuccessResponseJSON(w, b)
return
}
}
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
if clientToken != "" && !forceStart && !forceStop {
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
// Since clientToken is given, fetch heal status from running
// heal sequence.
respBytes, errCode := globalAllHealState.PopHealStatusJSON(
healPath, clientToken)
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
if errCode != ErrNone {
writeErrorResponseJSON(w, errCode, r.URL)
} else {
writeSuccessResponseJSON(w, respBytes)
}
return
}
respCh := make(chan healResp)
switch {
case forceStop:
go func() {
respBytes, errCode := globalAllHealState.stopHealSequence(healPath)
hr := healResp{respBytes: respBytes, errCode: errCode}
respCh <- hr
}()
case clientToken == "":
nh := newHealSequence(bucket, objPrefix, handlers.GetSourceIP(r), numDisks, hs, forceStart)
go func() {
respBytes, errCode, errMsg := globalAllHealState.LaunchNewHealSequence(nh)
hr := healResp{respBytes, errCode, errMsg}
respCh <- hr
}()
}
// Due to the force-starting functionality, the Launch
// call above can take a long time - to keep the
// connection alive, we start sending whitespace
keepConnLive(w, respCh)
}
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
// GetConfigHandler - GET /minio/admin/v1/config
// Get config.json of this minio setup.
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
func (a adminAPIHandlers) GetConfigHandler(w http.ResponseWriter, r *http.Request) {
ctx := newContext(r, w, "GetConfigHandler")
// Get current object layer instance.
objectAPI := newObjectLayerFn()
if objectAPI == nil {
writeErrorResponseJSON(w, ErrServerNotInitialized, r.URL)
return
}
// Validate request signature.
adminAPIErr := checkAdminRequestAuthType(ctx, r, "")
if adminAPIErr != ErrNone {
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
writeErrorResponseJSON(w, adminAPIErr, r.URL)
return
}
config, err := readServerConfig(ctx, objectAPI)
if err != nil {
writeErrorResponseJSON(w, toAdminAPIErrCode(ctx, err), r.URL)
return
}
configData, err := json.MarshalIndent(config, "", "\t")
if err != nil {
writeErrorResponseJSON(w, toAdminAPIErrCode(ctx, err), r.URL)
return
}
password := config.GetCredential().SecretKey
econfigData, err := madmin.EncryptData(password, configData)
if err != nil {
writeErrorResponseJSON(w, toAdminAPIErrCode(ctx, err), r.URL)
return
}
writeSuccessResponseJSON(w, econfigData)
}
// Disable tidwall json array notation in JSON key path so
// users can set json with a key as a number.
// In tidwall json, notify.webhook.0 = val means { "notify" : { "webhook" : [val] }}
// In Minio, notify.webhook.0 = val means { "notify" : { "webhook" : {"0" : val}}}
func normalizeJSONKey(input string) (key string) {
subKeys := strings.Split(input, ".")
for i, k := range subKeys {
if i > 0 {
key += "."
}
if _, err := strconv.Atoi(k); err == nil {
key += ":" + k
} else {
key += k
}
}
return
}
// GetConfigHandler - GET /minio/admin/v1/config-keys
// Get some keys in config.json of this minio setup.
func (a adminAPIHandlers) GetConfigKeysHandler(w http.ResponseWriter, r *http.Request) {
ctx := newContext(r, w, "GetConfigKeysHandler")
// Get current object layer instance.
objectAPI := newObjectLayerFn()
if objectAPI == nil {
writeErrorResponseJSON(w, ErrServerNotInitialized, r.URL)
return
}
// Validate request signature.
adminAPIErr := checkAdminRequestAuthType(ctx, r, "")
if adminAPIErr != ErrNone {
writeErrorResponseJSON(w, adminAPIErr, r.URL)
return
}
var keys []string
queries := r.URL.Query()
for k := range queries {
keys = append(keys, k)
}
config, err := readServerConfig(ctx, objectAPI)
if err != nil {
writeErrorResponseJSON(w, toAdminAPIErrCode(ctx, err), r.URL)
return
}
configData, err := json.Marshal(config)
if err != nil {
writeErrorResponseJSON(w, toAdminAPIErrCode(ctx, err), r.URL)
return
}
configStr := string(configData)
newConfigStr := `{}`
for _, key := range keys {
// sjson.Set does not return an error if key is empty
// we should check by ourselves here
if key == "" {
continue
}
val := gjson.Get(configStr, key)
2018-10-02 18:21:34 -04:00
if j, ierr := sjson.Set(newConfigStr, normalizeJSONKey(key), val.Value()); ierr == nil {
newConfigStr = j
}
}
password := config.GetCredential().SecretKey
econfigData, err := madmin.EncryptData(password, []byte(newConfigStr))
if err != nil {
writeErrorResponseJSON(w, toAdminAPIErrCode(ctx, err), r.URL)
return
}
writeSuccessResponseJSON(w, []byte(econfigData))
}
// toAdminAPIErrCode - converts errXLWriteQuorum error to admin API
// specific error.
func toAdminAPIErrCode(ctx context.Context, err error) APIErrorCode {
switch err {
case errXLWriteQuorum:
return ErrAdminConfigNoQuorum
default:
return toAPIErrorCode(ctx, err)
}
}
// RemoveUser - DELETE /minio/admin/v1/remove-user?accessKey=<access_key>
func (a adminAPIHandlers) RemoveUser(w http.ResponseWriter, r *http.Request) {
ctx := newContext(r, w, "RemoveUser")
// Get current object layer instance.
objectAPI := newObjectLayerFn()
if objectAPI == nil {
writeErrorResponseJSON(w, ErrServerNotInitialized, r.URL)
return
}
// Validate request signature.
adminAPIErr := checkAdminRequestAuthType(ctx, r, "")
if adminAPIErr != ErrNone {
writeErrorResponseJSON(w, adminAPIErr, r.URL)
return
}
// Deny if WORM is enabled
if globalWORMEnabled {
writeErrorResponseJSON(w, ErrMethodNotAllowed, r.URL)
return
}
vars := mux.Vars(r)
accessKey := vars["accessKey"]
if err := globalIAMSys.DeleteUser(accessKey); err != nil {
writeErrorResponseJSON(w, toAdminAPIErrCode(ctx, err), r.URL)
}
}
// ListUsers - GET /minio/admin/v1/list-users
func (a adminAPIHandlers) ListUsers(w http.ResponseWriter, r *http.Request) {
ctx := newContext(r, w, "ListUsers")
// Get current object layer instance.
objectAPI := newObjectLayerFn()
if objectAPI == nil {
writeErrorResponseJSON(w, ErrServerNotInitialized, r.URL)
return
}
// Validate request signature.
adminAPIErr := checkAdminRequestAuthType(ctx, r, "")
if adminAPIErr != ErrNone {
writeErrorResponseJSON(w, adminAPIErr, r.URL)
return
}
allCredentials, err := globalIAMSys.ListUsers()
if err != nil {
writeErrorResponseJSON(w, toAdminAPIErrCode(ctx, err), r.URL)
return
}
data, err := json.Marshal(allCredentials)
if err != nil {
writeErrorResponseJSON(w, toAdminAPIErrCode(ctx, err), r.URL)
return
}
password := globalServerConfig.GetCredential().SecretKey
econfigData, err := madmin.EncryptData(password, data)
if err != nil {
writeErrorResponseJSON(w, toAdminAPIErrCode(ctx, err), r.URL)
return
}
writeSuccessResponseJSON(w, econfigData)
}
2018-10-16 17:55:23 -04:00
// SetUserStatus - PUT /minio/admin/v1/set-user-status?accessKey=<access_key>&status=[enabled|disabled]
func (a adminAPIHandlers) SetUserStatus(w http.ResponseWriter, r *http.Request) {
ctx := newContext(r, w, "SetUserStatus")
// Get current object layer instance.
objectAPI := newObjectLayerFn()
if objectAPI == nil {
writeErrorResponseJSON(w, ErrServerNotInitialized, r.URL)
return
}
// Validate request signature.
adminAPIErr := checkAdminRequestAuthType(ctx, r, "")
2018-10-16 17:55:23 -04:00
if adminAPIErr != ErrNone {
writeErrorResponseJSON(w, adminAPIErr, r.URL)
return
}
// Deny if WORM is enabled
if globalWORMEnabled {
writeErrorResponseJSON(w, ErrMethodNotAllowed, r.URL)
return
}
vars := mux.Vars(r)
accessKey := vars["accessKey"]
status := vars["status"]
// Custom IAM policies not allowed for admin user.
if accessKey == globalServerConfig.GetCredential().AccessKey {
writeErrorResponseJSON(w, ErrInvalidRequest, r.URL)
2018-10-16 17:55:23 -04:00
return
}
if err := globalIAMSys.SetUserStatus(accessKey, madmin.AccountStatus(status)); err != nil {
writeErrorResponseJSON(w, toAdminAPIErrCode(ctx, err), r.URL)
2018-10-16 17:55:23 -04:00
return
}
// Notify all other Minio peers to reload users
for host, err := range globalNotificationSys.LoadUsers() {
if err != nil {
logger.GetReqInfo(ctx).SetTags("peerAddress", host.String())
logger.LogIf(ctx, err)
}
}
2018-10-16 17:55:23 -04:00
}
// AddUser - PUT /minio/admin/v1/add-user?accessKey=<access_key>
func (a adminAPIHandlers) AddUser(w http.ResponseWriter, r *http.Request) {
ctx := newContext(r, w, "AddUser")
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
// Get current object layer instance.
objectAPI := newObjectLayerFn()
if objectAPI == nil {
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
writeErrorResponseJSON(w, ErrServerNotInitialized, r.URL)
return
}
// Validate request signature.
adminAPIErr := checkAdminRequestAuthType(ctx, r, "")
if adminAPIErr != ErrNone {
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
writeErrorResponseJSON(w, adminAPIErr, r.URL)
return
}
// Deny if WORM is enabled
if globalWORMEnabled {
writeErrorResponseJSON(w, ErrMethodNotAllowed, r.URL)
return
}
vars := mux.Vars(r)
accessKey := vars["accessKey"]
// Custom IAM policies not allowed for admin user.
if accessKey == globalServerConfig.GetCredential().AccessKey {
writeErrorResponseJSON(w, ErrInvalidRequest, r.URL)
return
}
if r.ContentLength > maxEConfigJSONSize || r.ContentLength == -1 {
// More than maxConfigSize bytes were available
writeErrorResponseJSON(w, ErrAdminConfigTooLarge, r.URL)
return
}
password := globalServerConfig.GetCredential().SecretKey
configBytes, err := madmin.DecryptData(password, io.LimitReader(r.Body, r.ContentLength))
if err != nil {
logger.LogIf(ctx, err)
writeErrorResponseJSON(w, ErrAdminConfigBadJSON, r.URL)
return
}
var uinfo madmin.UserInfo
if err = json.Unmarshal(configBytes, &uinfo); err != nil {
logger.LogIf(ctx, err)
writeErrorResponseJSON(w, ErrAdminConfigBadJSON, r.URL)
return
}
if err = globalIAMSys.SetUser(accessKey, uinfo); err != nil {
writeErrorResponseJSON(w, toAdminAPIErrCode(ctx, err), r.URL)
return
}
// Notify all other Minio peers to reload users
for host, err := range globalNotificationSys.LoadUsers() {
if err != nil {
logger.GetReqInfo(ctx).SetTags("peerAddress", host.String())
logger.LogIf(ctx, err)
}
}
}
// ListCannedPolicies - GET /minio/admin/v1/list-canned-policies
func (a adminAPIHandlers) ListCannedPolicies(w http.ResponseWriter, r *http.Request) {
ctx := newContext(r, w, "ListCannedPolicies")
// Get current object layer instance.
objectAPI := newObjectLayerFn()
if objectAPI == nil {
writeErrorResponseJSON(w, ErrServerNotInitialized, r.URL)
return
}
// Validate request signature.
adminAPIErr := checkAdminRequestAuthType(ctx, r, "")
if adminAPIErr != ErrNone {
writeErrorResponseJSON(w, adminAPIErr, r.URL)
return
}
policies, err := globalIAMSys.ListCannedPolicies()
if err != nil {
writeErrorResponseJSON(w, toAdminAPIErrCode(ctx, err), r.URL)
return
}
if err = json.NewEncoder(w).Encode(policies); err != nil {
writeErrorResponseJSON(w, toAdminAPIErrCode(ctx, err), r.URL)
return
}
w.(http.Flusher).Flush()
}
// RemoveCannedPolicy - DELETE /minio/admin/v1/remove-canned-policy?name=<policy_name>
func (a adminAPIHandlers) RemoveCannedPolicy(w http.ResponseWriter, r *http.Request) {
ctx := newContext(r, w, "RemoveCannedPolicy")
// Get current object layer instance.
objectAPI := newObjectLayerFn()
if objectAPI == nil {
writeErrorResponseJSON(w, ErrServerNotInitialized, r.URL)
return
}
vars := mux.Vars(r)
policyName := vars["name"]
// Validate request signature.
adminAPIErr := checkAdminRequestAuthType(ctx, r, "")
if adminAPIErr != ErrNone {
writeErrorResponseJSON(w, adminAPIErr, r.URL)
return
}
// Deny if WORM is enabled
if globalWORMEnabled {
writeErrorResponseJSON(w, ErrMethodNotAllowed, r.URL)
return
}
if err := globalIAMSys.DeleteCannedPolicy(policyName); err != nil {
writeErrorResponseJSON(w, toAdminAPIErrCode(ctx, err), r.URL)
return
}
// Notify all other Minio peers to reload users
for host, err := range globalNotificationSys.LoadUsers() {
if err != nil {
logger.GetReqInfo(ctx).SetTags("peerAddress", host.String())
logger.LogIf(ctx, err)
}
}
}
// AddCannedPolicy - PUT /minio/admin/v1/add-canned-policy?name=<policy_name>
func (a adminAPIHandlers) AddCannedPolicy(w http.ResponseWriter, r *http.Request) {
ctx := newContext(r, w, "AddCannedPolicy")
// Get current object layer instance.
objectAPI := newObjectLayerFn()
if objectAPI == nil {
writeErrorResponseJSON(w, ErrServerNotInitialized, r.URL)
return
}
vars := mux.Vars(r)
policyName := vars["name"]
// Validate request signature.
adminAPIErr := checkAdminRequestAuthType(ctx, r, "")
if adminAPIErr != ErrNone {
writeErrorResponseJSON(w, adminAPIErr, r.URL)
return
}
// Deny if WORM is enabled
if globalWORMEnabled {
writeErrorResponseJSON(w, ErrMethodNotAllowed, r.URL)
return
}
// Error out if Content-Length is missing.
if r.ContentLength <= 0 {
writeErrorResponseJSON(w, ErrMissingContentLength, r.URL)
return
}
// Error out if Content-Length is beyond allowed size.
if r.ContentLength > maxBucketPolicySize {
writeErrorResponseJSON(w, ErrEntityTooLarge, r.URL)
return
}
iamPolicy, err := iampolicy.ParseConfig(io.LimitReader(r.Body, r.ContentLength))
if err != nil {
writeErrorResponseJSON(w, ErrMalformedPolicy, r.URL)
return
}
// Version in policy must not be empty
if iamPolicy.Version == "" {
writeErrorResponseJSON(w, ErrMalformedPolicy, r.URL)
return
}
if err = globalIAMSys.SetCannedPolicy(policyName, *iamPolicy); err != nil {
writeErrorResponseJSON(w, toAdminAPIErrCode(ctx, err), r.URL)
return
}
// Notify all other Minio peers to reload users
for host, err := range globalNotificationSys.LoadUsers() {
if err != nil {
logger.GetReqInfo(ctx).SetTags("peerAddress", host.String())
logger.LogIf(ctx, err)
}
}
}
// SetUserPolicy - PUT /minio/admin/v1/set-user-policy?accessKey=<access_key>&name=<policy_name>
func (a adminAPIHandlers) SetUserPolicy(w http.ResponseWriter, r *http.Request) {
ctx := newContext(r, w, "SetUserPolicy")
// Get current object layer instance.
objectAPI := newObjectLayerFn()
if objectAPI == nil {
writeErrorResponseJSON(w, ErrServerNotInitialized, r.URL)
return
}
vars := mux.Vars(r)
accessKey := vars["accessKey"]
policyName := vars["name"]
// Validate request signature.
adminAPIErr := checkAdminRequestAuthType(ctx, r, "")
if adminAPIErr != ErrNone {
writeErrorResponseJSON(w, adminAPIErr, r.URL)
return
}
// Deny if WORM is enabled
if globalWORMEnabled {
writeErrorResponseJSON(w, ErrMethodNotAllowed, r.URL)
return
}
// Custom IAM policies not allowed for admin user.
if accessKey == globalServerConfig.GetCredential().AccessKey {
writeErrorResponseJSON(w, ErrInvalidRequest, r.URL)
return
}
if err := globalIAMSys.SetUserPolicy(accessKey, policyName); err != nil {
writeErrorResponseJSON(w, toAdminAPIErrCode(ctx, err), r.URL)
}
// Notify all other Minio peers to reload users
for host, err := range globalNotificationSys.LoadUsers() {
if err != nil {
logger.GetReqInfo(ctx).SetTags("peerAddress", host.String())
logger.LogIf(ctx, err)
}
}
}
// SetConfigHandler - PUT /minio/admin/v1/config
func (a adminAPIHandlers) SetConfigHandler(w http.ResponseWriter, r *http.Request) {
ctx := newContext(r, w, "SetConfigHandler")
// Get current object layer instance.
objectAPI := newObjectLayerFn()
if objectAPI == nil {
writeErrorResponseJSON(w, ErrServerNotInitialized, r.URL)
return
}
// Validate request signature.
adminAPIErr := checkAdminRequestAuthType(ctx, r, "")
if adminAPIErr != ErrNone {
writeErrorResponseJSON(w, adminAPIErr, r.URL)
return
}
// Deny if WORM is enabled
if globalWORMEnabled {
writeErrorResponseJSON(w, ErrMethodNotAllowed, r.URL)
return
}
if r.ContentLength > maxEConfigJSONSize || r.ContentLength == -1 {
// More than maxConfigSize bytes were available
writeErrorResponseJSON(w, ErrAdminConfigTooLarge, r.URL)
return
}
password := globalServerConfig.GetCredential().SecretKey
configBytes, err := madmin.DecryptData(password, io.LimitReader(r.Body, r.ContentLength))
if err != nil {
logger.LogIf(ctx, err)
writeErrorResponseJSON(w, ErrAdminConfigBadJSON, r.URL)
return
}
// Validate JSON provided in the request body: check the
// client has not sent JSON objects with duplicate keys.
if err = quick.CheckDuplicateKeys(string(configBytes)); err != nil {
logger.LogIf(ctx, err)
writeErrorResponseJSON(w, ErrAdminConfigBadJSON, r.URL)
return
}
var config serverConfig
if err = json.Unmarshal(configBytes, &config); err != nil {
logger.LogIf(ctx, err)
writeCustomErrorResponseJSON(w, ErrAdminConfigBadJSON, err.Error(), r.URL)
return
}
// If credentials for the server are provided via environment,
// then credentials in the provided configuration must match.
if globalIsEnvCreds {
creds := globalServerConfig.GetCredential()
if config.Credential.AccessKey != creds.AccessKey ||
config.Credential.SecretKey != creds.SecretKey {
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
writeErrorResponseJSON(w, ErrAdminCredentialsMismatch, r.URL)
return
}
}
if err = config.Validate(); err != nil {
writeCustomErrorResponseJSON(w, ErrAdminConfigBadJSON, err.Error(), r.URL)
return
}
if err = config.TestNotificationTargets(); err != nil {
writeCustomErrorResponseJSON(w, ErrAdminConfigBadJSON, err.Error(), r.URL)
return
}
if err = saveServerConfig(ctx, objectAPI, &config); err != nil {
writeErrorResponseJSON(w, toAdminAPIErrCode(ctx, err), r.URL)
return
}
// Reply to the client before restarting minio server.
writeSuccessResponseHeadersOnly(w)
}
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
func convertValueType(elem []byte, jsonType gjson.Type) (interface{}, error) {
str := string(elem)
switch jsonType {
case gjson.False, gjson.True:
return strconv.ParseBool(str)
case gjson.JSON:
return gjson.Parse(str).Value(), nil
case gjson.String:
return str, nil
case gjson.Number:
return strconv.ParseFloat(str, 64)
default:
return nil, nil
}
}
// SetConfigKeysHandler - PUT /minio/admin/v1/config-keys
func (a adminAPIHandlers) SetConfigKeysHandler(w http.ResponseWriter, r *http.Request) {
ctx := newContext(r, w, "SetConfigKeysHandler")
// Get current object layer instance.
objectAPI := newObjectLayerFn()
if objectAPI == nil {
writeErrorResponseJSON(w, ErrServerNotInitialized, r.URL)
return
}
// Deny if WORM is enabled
if globalWORMEnabled {
writeErrorResponseJSON(w, ErrMethodNotAllowed, r.URL)
return
}
// Validate request signature.
adminAPIErr := checkAdminRequestAuthType(ctx, r, "")
if adminAPIErr != ErrNone {
writeErrorResponseJSON(w, adminAPIErr, r.URL)
return
}
// Load config
configStruct, err := readServerConfig(ctx, objectAPI)
if err != nil {
writeErrorResponseJSON(w, toAdminAPIErrCode(ctx, err), r.URL)
return
}
// Convert config to json bytes
configBytes, err := json.Marshal(configStruct)
if err != nil {
writeErrorResponseJSON(w, toAdminAPIErrCode(ctx, err), r.URL)
return
}
configStr := string(configBytes)
queries := r.URL.Query()
password := globalServerConfig.GetCredential().SecretKey
// Set key values in the JSON config
for k := range queries {
// Decode encrypted data associated to the current key
encryptedElem, dErr := base64.StdEncoding.DecodeString(queries.Get(k))
if dErr != nil {
reqInfo := (&logger.ReqInfo{}).AppendTags("key", k)
ctx = logger.SetReqInfo(ctx, reqInfo)
logger.LogIf(ctx, dErr)
writeErrorResponseJSON(w, ErrAdminConfigBadJSON, r.URL)
return
}
elem, dErr := madmin.DecryptData(password, bytes.NewBuffer([]byte(encryptedElem)))
if dErr != nil {
logger.LogIf(ctx, dErr)
writeErrorResponseJSON(w, ErrAdminConfigBadJSON, r.URL)
return
}
// Calculate the type of the current key from the
// original config json
jsonFieldType := gjson.Get(configStr, k).Type
// Convert passed value to json filed type
val, cErr := convertValueType(elem, jsonFieldType)
if cErr != nil {
writeCustomErrorResponseJSON(w, ErrAdminConfigBadJSON, cErr.Error(), r.URL)
return
}
// Set the key/value in the new json document
if s, sErr := sjson.Set(configStr, normalizeJSONKey(k), val); sErr == nil {
configStr = s
}
}
configBytes = []byte(configStr)
// Validate config
var config serverConfig
if err = json.Unmarshal(configBytes, &config); err != nil {
writeCustomErrorResponseJSON(w, ErrAdminConfigBadJSON, err.Error(), r.URL)
return
}
if err = config.Validate(); err != nil {
writeCustomErrorResponseJSON(w, ErrAdminConfigBadJSON, err.Error(), r.URL)
return
}
if err = config.TestNotificationTargets(); err != nil {
writeCustomErrorResponseJSON(w, ErrAdminConfigBadJSON, err.Error(), r.URL)
return
}
// If credentials for the server are provided via environment,
// then credentials in the provided configuration must match.
if globalIsEnvCreds {
creds := globalServerConfig.GetCredential()
if config.Credential.AccessKey != creds.AccessKey ||
config.Credential.SecretKey != creds.SecretKey {
writeErrorResponseJSON(w, ErrAdminCredentialsMismatch, r.URL)
return
}
}
if err = saveServerConfig(ctx, objectAPI, &config); err != nil {
writeErrorResponseJSON(w, toAdminAPIErrCode(ctx, err), r.URL)
return
}
// Send success response
writeSuccessResponseHeadersOnly(w)
}
// UpdateAdminCredsHandler - POST /minio/admin/v1/config/credential
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
// ----------
// Update admin credentials in a minio server
func (a adminAPIHandlers) UpdateAdminCredentialsHandler(w http.ResponseWriter,
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
r *http.Request) {
ctx := newContext(r, w, "UpdateCredentialsHandler")
// Get current object layer instance.
objectAPI := newObjectLayerFn()
if objectAPI == nil {
writeErrorResponseJSON(w, ErrServerNotInitialized, r.URL)
return
}
// Avoid setting new credentials when they are already passed
// by the environment. Deny if WORM is enabled.
if globalIsEnvCreds || globalWORMEnabled {
writeErrorResponseJSON(w, ErrMethodNotAllowed, r.URL)
return
}
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
// Authenticate request
adminAPIErr := checkAdminRequestAuthType(ctx, r, "")
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
if adminAPIErr != ErrNone {
writeErrorResponseJSON(w, adminAPIErr, r.URL)
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
return
}
if r.ContentLength > maxEConfigJSONSize || r.ContentLength == -1 {
// More than maxConfigSize bytes were available
writeErrorResponseJSON(w, ErrAdminConfigTooLarge, r.URL)
return
}
password := globalServerConfig.GetCredential().SecretKey
configBytes, err := madmin.DecryptData(password, io.LimitReader(r.Body, r.ContentLength))
if err != nil {
logger.LogIf(ctx, err)
writeErrorResponseJSON(w, ErrAdminConfigBadJSON, r.URL)
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
return
}
// Decode request body
var req madmin.SetCredsReq
if err = json.Unmarshal(configBytes, &req); err != nil {
logger.LogIf(ctx, err)
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
writeErrorResponseJSON(w, ErrRequestBodyParse, r.URL)
return
}
creds, err := auth.CreateCredentials(req.AccessKey, req.SecretKey)
if err != nil {
writeErrorResponseJSON(w, toAdminAPIErrCode(ctx, err), r.URL)
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
return
}
// Acquire lock before updating global configuration.
globalServerConfigMu.Lock()
defer globalServerConfigMu.Unlock()
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
// Update local credentials in memory.
globalServerConfig.SetCredential(creds)
// Set active creds.
globalActiveCred = creds
if err = saveServerConfig(ctx, objectAPI, globalServerConfig); err != nil {
writeErrorResponseJSON(w, toAdminAPIErrCode(ctx, err), r.URL)
return
}
// Notify all other Minio peers to update credentials
for host, err := range globalNotificationSys.LoadCredentials() {
if err != nil {
logger.GetReqInfo(ctx).SetTags("peerAddress", host.String())
logger.LogIf(ctx, err)
}
}
// Reply to the client before restarting minio server.
writeSuccessResponseHeadersOnly(w)
Move admin APIs to new path and add redesigned heal APIs (#5351) - Changes related to moving admin APIs - admin APIs now have an endpoint under /minio/admin - admin APIs are now versioned - a new API to server the version is added at "GET /minio/admin/version" and all API operations have the path prefix /minio/admin/v1/<operation> - new service stop API added - credentials change API is moved to /minio/admin/v1/config/credential - credentials change API and configuration get/set API now require TLS so that credentials are protected - all API requests now receive JSON - heal APIs are disabled as they will be changed substantially - Heal API changes Heal API is now provided at a single endpoint with the ability for a client to start a heal sequence on all the data in the server, a single bucket, or under a prefix within a bucket. When a heal sequence is started, the server returns a unique token that needs to be used for subsequent 'status' requests to fetch heal results. On each status request from the client, the server returns heal result records that it has accumulated since the previous status request. The server accumulates upto 1000 records and pauses healing further objects until the client requests for status. If the client does not request any further records for a long time, the server aborts the heal sequence automatically. A heal result record is returned for each entity healed on the server, such as system metadata, object metadata, buckets and objects, and has information about the before and after states on each disk. A client may request to force restart a heal sequence - this causes the running heal sequence to be aborted at the next safe spot and starts a new heal sequence.
2018-01-22 17:54:55 -05:00
}