minio/cmd/erasure-healing-common.go

378 lines
11 KiB
Go
Raw Normal View History

// Copyright (c) 2015-2021 MinIO, Inc.
//
// This file is part of MinIO Object Storage stack
//
// This program is free software: you can redistribute it and/or modify
// it under the terms of the GNU Affero General Public License as published by
// the Free Software Foundation, either version 3 of the License, or
// (at your option) any later version.
//
// This program is distributed in the hope that it will be useful
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU Affero General Public License for more details.
//
// You should have received a copy of the GNU Affero General Public License
// along with this program. If not, see <http://www.gnu.org/licenses/>.
package cmd
import (
"bytes"
"context"
"time"
"github.com/minio/madmin-go/v3"
)
func commonETags(etags []string) (etag string, maxima int) {
etagOccurrenceMap := make(map[string]int, len(etags))
// Ignore the uuid sentinel and count the rest.
for _, etag := range etags {
if etag == "" {
continue
}
etagOccurrenceMap[etag]++
}
maxima = 0 // Counter for remembering max occurrence of elements.
latest := ""
// Find the common cardinality from previously collected
// occurrences of elements.
for etag, count := range etagOccurrenceMap {
if count < maxima {
continue
}
// We are at or above maxima
if count > maxima {
maxima = count
latest = etag
}
}
// Return the collected common max time, with maxima
return latest, maxima
}
// commonTime returns a maximally occurring time from a list of time.
func commonTimeAndOccurence(times []time.Time, group time.Duration) (maxTime time.Time, maxima int) {
timeOccurenceMap := make(map[int64]int, len(times))
groupNano := group.Nanoseconds()
// Ignore the uuid sentinel and count the rest.
for _, t := range times {
if t.Equal(timeSentinel) || t.IsZero() {
continue
}
nano := t.UnixNano()
if group > 0 {
for k := range timeOccurenceMap {
if k == nano {
// We add to ourself later
continue
}
diff := k - nano
if diff < 0 {
diff = -diff
}
// We are within the limit
if diff < groupNano {
timeOccurenceMap[k]++
}
}
}
// Add ourself...
timeOccurenceMap[nano]++
}
maxima = 0 // Counter for remembering max occurrence of elements.
latest := int64(0)
// Find the common cardinality from previously collected
// occurrences of elements.
for nano, count := range timeOccurenceMap {
if count < maxima {
continue
}
// We are at or above maxima
if count > maxima || nano > latest {
maxima = count
latest = nano
}
}
// Return the collected common max time, with maxima
return time.Unix(0, latest).UTC(), maxima
}
// commonTime returns a maximally occurring time from a list of time if it
// occurs >= quorum, else return timeSentinel
func commonTime(modTimes []time.Time, quorum int) time.Time {
if modTime, count := commonTimeAndOccurence(modTimes, 0); count >= quorum {
return modTime
}
return timeSentinel
}
func commonETag(etags []string, quorum int) string {
if etag, count := commonETags(etags); count >= quorum {
return etag
}
return ""
}
// Beginning of unix time is treated as sentinel value here.
var (
timeSentinel = time.Unix(0, 0).UTC()
timeSentinel1970 = time.Unix(0, 1).UTC() // 1970 used for special cases when xlmeta.version == 0
)
// Boot modTimes up to disk count, setting the value to time sentinel.
func bootModtimes(diskCount int) []time.Time {
modTimes := make([]time.Time, diskCount)
// Boots up all the modtimes.
for i := range modTimes {
modTimes[i] = timeSentinel
}
return modTimes
}
func listObjectETags(partsMetadata []FileInfo, errs []error, quorum int) (etags []string) {
etags = make([]string, len(partsMetadata))
vidMap := map[string]int{}
for index, metadata := range partsMetadata {
if errs[index] != nil {
continue
}
vid := metadata.VersionID
if metadata.VersionID == "" {
vid = nullVersionID
}
vidMap[vid]++
etags[index] = metadata.Metadata["etag"]
}
for _, count := range vidMap {
// do we have enough common versions
// that have enough quorum to satisfy
// the etag.
if count >= quorum {
return etags
}
}
return make([]string, len(partsMetadata))
}
// Extracts list of times from FileInfo slice and returns, skips
// slice elements which have errors.
func listObjectModtimes(partsMetadata []FileInfo, errs []error) (modTimes []time.Time) {
modTimes = bootModtimes(len(partsMetadata))
for index, metadata := range partsMetadata {
if errs[index] != nil {
continue
}
// Once the file is found, save the uuid saved on disk.
modTimes[index] = metadata.ModTime
}
return modTimes
}
func filterOnlineDisksInplace(fi FileInfo, partsMetadata []FileInfo, onlineDisks []StorageAPI) {
for i, meta := range partsMetadata {
if fi.XLV1 == meta.XLV1 {
continue
}
onlineDisks[i] = nil
}
}
// Notes:
// There are 5 possible states a disk could be in,
// 1. __online__ - has the latest copy of xl.meta - returned by listOnlineDisks
//
// 2. __offline__ - err == errDiskNotFound
//
// 3. __availableWithParts__ - has the latest copy of xl.meta and has all
// parts with checksums matching; returned by disksWithAllParts
//
// 4. __outdated__ - returned by outDatedDisk, provided []StorageAPI
// returned by diskWithAllParts is passed for latestDisks.
// - has an old copy of xl.meta
// - doesn't have xl.meta (errFileNotFound)
// - has the latest xl.meta but one or more parts are corrupt
//
// 5. __missingParts__ - has the latest copy of xl.meta but has some parts
// missing. This is identified separately since this may need manual
// inspection to understand the root cause. E.g, this could be due to
// backend filesystem corruption.
// listOnlineDisks - returns
// - a slice of disks where disk having 'older' xl.meta (or nothing)
// are set to nil.
// - latest (in time) of the maximally occurring modTime(s), which has at least quorum occurrences.
func listOnlineDisks(disks []StorageAPI, partsMetadata []FileInfo, errs []error, quorum int) (onlineDisks []StorageAPI, modTime time.Time, etag string) {
onlineDisks = make([]StorageAPI, len(disks))
// List all the file commit ids from parts metadata.
modTimes := listObjectModtimes(partsMetadata, errs)
// Reduce list of UUIDs to a single common value.
modTime = commonTime(modTimes, quorum)
if modTime.IsZero() || modTime.Equal(timeSentinel) {
etags := listObjectETags(partsMetadata, errs, quorum)
etag = commonETag(etags, quorum)
if etag != "" { // allow this fallback only if a non-empty etag is found.
for index, e := range etags {
if partsMetadata[index].IsValid() && e == etag {
onlineDisks[index] = disks[index]
} else {
onlineDisks[index] = nil
}
}
return onlineDisks, modTime, etag
}
}
// Create a new online disks slice, which have common uuid.
for index, t := range modTimes {
if partsMetadata[index].IsValid() && t.Equal(modTime) {
onlineDisks[index] = disks[index]
} else {
onlineDisks[index] = nil
}
}
return onlineDisks, modTime, ""
}
// disksWithAllParts - This function needs to be called with
// []StorageAPI returned by listOnlineDisks. Returns,
//
// - disks which have all parts specified in the latest xl.meta.
//
// - slice of errors about the state of data files on disk - can have
// a not-found error or a hash-mismatch error.
func disksWithAllParts(ctx context.Context, onlineDisks []StorageAPI, partsMetadata []FileInfo,
errs []error, latestMeta FileInfo, bucket, object string,
scanMode madmin.HealScanMode) ([]StorageAPI, []error, time.Time,
) {
availableDisks := make([]StorageAPI, len(onlineDisks))
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
dataErrs := make([]error, len(onlineDisks))
inconsistent := 0
for i, meta := range partsMetadata {
if !meta.IsValid() {
// Since for majority of the cases erasure.Index matches with erasure.Distribution we can
// consider the offline disks as consistent.
continue
}
if !meta.Deleted {
if len(meta.Erasure.Distribution) != len(onlineDisks) {
// Erasure distribution seems to have lesser
// number of items than number of online disks.
inconsistent++
continue
}
if meta.Erasure.Distribution[i] != meta.Erasure.Index {
// Mismatch indexes with distribution order
inconsistent++
}
}
}
erasureDistributionReliable := true
if inconsistent > len(partsMetadata)/2 {
// If there are too many inconsistent files, then we can't trust erasure.Distribution (most likely
// because of bugs found in CopyObject/PutObjectTags) https://github.com/minio/minio/pull/10772
erasureDistributionReliable = false
}
for i, onlineDisk := range onlineDisks {
if errs[i] != nil {
dataErrs[i] = errs[i]
continue
}
if onlineDisk == OfflineDisk {
dataErrs[i] = errDiskNotFound
continue
}
meta := partsMetadata[i]
if !meta.ModTime.Equal(latestMeta.ModTime) || meta.DataDir != latestMeta.DataDir {
dataErrs[i] = errFileCorrupt
partsMetadata[i] = FileInfo{}
continue
}
if erasureDistributionReliable {
if !meta.IsValid() {
partsMetadata[i] = FileInfo{}
dataErrs[i] = errFileCorrupt
continue
}
if !meta.Deleted {
if len(meta.Erasure.Distribution) != len(onlineDisks) {
// Erasure distribution is not the same as onlineDisks
// attempt a fix if possible, assuming other entries
// might have the right erasure distribution.
partsMetadata[i] = FileInfo{}
dataErrs[i] = errFileCorrupt
continue
}
}
}
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
// Always check data, if we got it.
if (len(meta.Data) > 0 || meta.Size == 0) && len(meta.Parts) > 0 {
checksumInfo := meta.Erasure.GetChecksumInfo(meta.Parts[0].Number)
dataErrs[i] = bitrotVerify(bytes.NewReader(meta.Data),
int64(len(meta.Data)),
meta.Erasure.ShardFileSize(meta.Size),
checksumInfo.Algorithm,
checksumInfo.Hash, meta.Erasure.ShardSize())
if dataErrs[i] == nil {
// All parts verified, mark it as all data available.
availableDisks[i] = onlineDisk
} else {
// upon errors just make that disk's fileinfo invalid
partsMetadata[i] = FileInfo{}
}
continue
}
meta.DataDir = latestMeta.DataDir
switch scanMode {
case madmin.HealDeepScan:
// disk has a valid xl.meta but may not have all the
// parts. This is considered an outdated disk, since
// it needs healing too.
if !meta.Deleted && !meta.IsRemote() {
dataErrs[i] = onlineDisk.VerifyFile(ctx, bucket, object, meta)
}
case madmin.HealNormalScan:
if !meta.Deleted && !meta.IsRemote() {
dataErrs[i] = onlineDisk.CheckParts(ctx, bucket, object, meta)
}
}
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 dataErrs[i] == nil {
// All parts verified, mark it as all data available.
availableDisks[i] = onlineDisk
} else {
// upon errors just make that disk's fileinfo invalid
partsMetadata[i] = FileInfo{}
}
}
return availableDisks, dataErrs, timeSentinel
}