minio/cmd/metacache-walk.go

449 lines
12 KiB
Go
Raw Normal View History

// Copyright (c) 2015-2023 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 (
"context"
"io"
"sort"
"strings"
perf: websocket grid connectivity for all internode communication (#18461) This PR adds a WebSocket grid feature that allows servers to communicate via a single two-way connection. There are two request types: * Single requests, which are `[]byte => ([]byte, error)`. This is for efficient small roundtrips with small payloads. * Streaming requests which are `[]byte, chan []byte => chan []byte (and error)`, which allows for different combinations of full two-way streams with an initial payload. Only a single stream is created between two machines - and there is, as such, no server/client relation since both sides can initiate and handle requests. Which server initiates the request is decided deterministically on the server names. Requests are made through a mux client and server, which handles message passing, congestion, cancelation, timeouts, etc. If a connection is lost, all requests are canceled, and the calling server will try to reconnect. Registered handlers can operate directly on byte slices or use a higher-level generics abstraction. There is no versioning of handlers/clients, and incompatible changes should be handled by adding new handlers. The request path can be changed to a new one for any protocol changes. First, all servers create a "Manager." The manager must know its address as well as all remote addresses. This will manage all connections. To get a connection to any remote, ask the manager to provide it given the remote address using. ``` func (m *Manager) Connection(host string) *Connection ``` All serverside handlers must also be registered on the manager. This will make sure that all incoming requests are served. The number of in-flight requests and responses must also be given for streaming requests. The "Connection" returned manages the mux-clients. Requests issued to the connection will be sent to the remote. * `func (c *Connection) Request(ctx context.Context, h HandlerID, req []byte) ([]byte, error)` performs a single request and returns the result. Any deadline provided on the request is forwarded to the server, and canceling the context will make the function return at once. * `func (c *Connection) NewStream(ctx context.Context, h HandlerID, payload []byte) (st *Stream, err error)` will initiate a remote call and send the initial payload. ```Go // A Stream is a two-way stream. // All responses *must* be read by the caller. // If the call is canceled through the context, //The appropriate error will be returned. type Stream struct { // Responses from the remote server. // Channel will be closed after an error or when the remote closes. // All responses *must* be read by the caller until either an error is returned or the channel is closed. // Canceling the context will cause the context cancellation error to be returned. Responses <-chan Response // Requests sent to the server. // If the handler is defined with 0 incoming capacity this will be nil. // Channel *must* be closed to signal the end of the stream. // If the request context is canceled, the stream will no longer process requests. Requests chan<- []byte } type Response struct { Msg []byte Err error } ``` There are generic versions of the server/client handlers that allow the use of type safe implementations for data types that support msgpack marshal/unmarshal.
2023-11-20 20:09:35 -05:00
"github.com/minio/minio/internal/grid"
xioutil "github.com/minio/minio/internal/ioutil"
"github.com/valyala/bytebufferpool"
)
perf: websocket grid connectivity for all internode communication (#18461) This PR adds a WebSocket grid feature that allows servers to communicate via a single two-way connection. There are two request types: * Single requests, which are `[]byte => ([]byte, error)`. This is for efficient small roundtrips with small payloads. * Streaming requests which are `[]byte, chan []byte => chan []byte (and error)`, which allows for different combinations of full two-way streams with an initial payload. Only a single stream is created between two machines - and there is, as such, no server/client relation since both sides can initiate and handle requests. Which server initiates the request is decided deterministically on the server names. Requests are made through a mux client and server, which handles message passing, congestion, cancelation, timeouts, etc. If a connection is lost, all requests are canceled, and the calling server will try to reconnect. Registered handlers can operate directly on byte slices or use a higher-level generics abstraction. There is no versioning of handlers/clients, and incompatible changes should be handled by adding new handlers. The request path can be changed to a new one for any protocol changes. First, all servers create a "Manager." The manager must know its address as well as all remote addresses. This will manage all connections. To get a connection to any remote, ask the manager to provide it given the remote address using. ``` func (m *Manager) Connection(host string) *Connection ``` All serverside handlers must also be registered on the manager. This will make sure that all incoming requests are served. The number of in-flight requests and responses must also be given for streaming requests. The "Connection" returned manages the mux-clients. Requests issued to the connection will be sent to the remote. * `func (c *Connection) Request(ctx context.Context, h HandlerID, req []byte) ([]byte, error)` performs a single request and returns the result. Any deadline provided on the request is forwarded to the server, and canceling the context will make the function return at once. * `func (c *Connection) NewStream(ctx context.Context, h HandlerID, payload []byte) (st *Stream, err error)` will initiate a remote call and send the initial payload. ```Go // A Stream is a two-way stream. // All responses *must* be read by the caller. // If the call is canceled through the context, //The appropriate error will be returned. type Stream struct { // Responses from the remote server. // Channel will be closed after an error or when the remote closes. // All responses *must* be read by the caller until either an error is returned or the channel is closed. // Canceling the context will cause the context cancellation error to be returned. Responses <-chan Response // Requests sent to the server. // If the handler is defined with 0 incoming capacity this will be nil. // Channel *must* be closed to signal the end of the stream. // If the request context is canceled, the stream will no longer process requests. Requests chan<- []byte } type Response struct { Msg []byte Err error } ``` There are generic versions of the server/client handlers that allow the use of type safe implementations for data types that support msgpack marshal/unmarshal.
2023-11-20 20:09:35 -05:00
//go:generate msgp -file $GOFILE
// WalkDirOptions provides options for WalkDir operations.
type WalkDirOptions struct {
// Bucket to scanner
Bucket string
// Directory inside the bucket.
BaseDir string
// Do a full recursive scan.
Recursive bool
// ReportNotFound will return errFileNotFound if all disks reports the BaseDir cannot be found.
ReportNotFound bool
// FilterPrefix will only return results with given prefix within folder.
// Should never contain a slash.
FilterPrefix string
// ForwardTo will forward to the given object path.
ForwardTo string
// Limit the number of returned objects if > 0.
Limit int
perf: websocket grid connectivity for all internode communication (#18461) This PR adds a WebSocket grid feature that allows servers to communicate via a single two-way connection. There are two request types: * Single requests, which are `[]byte => ([]byte, error)`. This is for efficient small roundtrips with small payloads. * Streaming requests which are `[]byte, chan []byte => chan []byte (and error)`, which allows for different combinations of full two-way streams with an initial payload. Only a single stream is created between two machines - and there is, as such, no server/client relation since both sides can initiate and handle requests. Which server initiates the request is decided deterministically on the server names. Requests are made through a mux client and server, which handles message passing, congestion, cancelation, timeouts, etc. If a connection is lost, all requests are canceled, and the calling server will try to reconnect. Registered handlers can operate directly on byte slices or use a higher-level generics abstraction. There is no versioning of handlers/clients, and incompatible changes should be handled by adding new handlers. The request path can be changed to a new one for any protocol changes. First, all servers create a "Manager." The manager must know its address as well as all remote addresses. This will manage all connections. To get a connection to any remote, ask the manager to provide it given the remote address using. ``` func (m *Manager) Connection(host string) *Connection ``` All serverside handlers must also be registered on the manager. This will make sure that all incoming requests are served. The number of in-flight requests and responses must also be given for streaming requests. The "Connection" returned manages the mux-clients. Requests issued to the connection will be sent to the remote. * `func (c *Connection) Request(ctx context.Context, h HandlerID, req []byte) ([]byte, error)` performs a single request and returns the result. Any deadline provided on the request is forwarded to the server, and canceling the context will make the function return at once. * `func (c *Connection) NewStream(ctx context.Context, h HandlerID, payload []byte) (st *Stream, err error)` will initiate a remote call and send the initial payload. ```Go // A Stream is a two-way stream. // All responses *must* be read by the caller. // If the call is canceled through the context, //The appropriate error will be returned. type Stream struct { // Responses from the remote server. // Channel will be closed after an error or when the remote closes. // All responses *must* be read by the caller until either an error is returned or the channel is closed. // Canceling the context will cause the context cancellation error to be returned. Responses <-chan Response // Requests sent to the server. // If the handler is defined with 0 incoming capacity this will be nil. // Channel *must* be closed to signal the end of the stream. // If the request context is canceled, the stream will no longer process requests. Requests chan<- []byte } type Response struct { Msg []byte Err error } ``` There are generic versions of the server/client handlers that allow the use of type safe implementations for data types that support msgpack marshal/unmarshal.
2023-11-20 20:09:35 -05:00
// DiskID contains the disk ID of the disk.
// Leave empty to not check disk ID.
DiskID string
}
// supported FS for Nlink optimization in readdir.
const (
xfs = "XFS"
ext4 = "EXT4"
)
// WalkDir will traverse a directory and return all entries found.
// On success a sorted meta cache stream will be returned.
// Metadata has data stripped, if any.
func (s *xlStorage) WalkDir(ctx context.Context, opts WalkDirOptions, wr io.Writer) (err error) {
legacyFS := !(s.fsType == xfs || s.fsType == ext4)
s.RLock()
legacy := s.formatLegacy
s.RUnlock()
// Verify if volume is valid and it exists.
volumeDir, err := s.getVolDir(opts.Bucket)
if err != nil {
return err
}
if !skipAccessChecks(opts.Bucket) {
// Stat a volume entry.
if err = Access(volumeDir); err != nil {
return convertAccessError(err, errVolumeAccessDenied)
}
}
// Use a small block size to start sending quickly
w := newMetacacheWriter(wr, 16<<10)
w.reuseBlocks = true // We are not sharing results, so reuse buffers.
defer w.Close()
out, err := w.stream()
if err != nil {
return err
}
defer xioutil.SafeClose(out)
var objsReturned int
objReturned := func(metadata []byte) {
if opts.Limit <= 0 {
return
}
if m, _, _ := isIndexedMetaV2(metadata); m != nil && !m.AllHidden(true) {
objsReturned++
}
}
send := func(entry metaCacheEntry) error {
objReturned(entry.metadata)
select {
case <-ctx.Done():
return ctx.Err()
case out <- entry:
}
return nil
}
// Fast exit track to check if we are listing an object with
// a trailing slash, this will avoid to list the object content.
if HasSuffix(opts.BaseDir, SlashSeparator) {
metadata, err := s.readMetadata(ctx, pathJoin(volumeDir,
opts.BaseDir[:len(opts.BaseDir)-1]+globalDirSuffix,
xlStorageFormatFile))
diskHealthCheckOK(ctx, err)
if err == nil {
// if baseDir is already a directory object, consider it
// as part of the list call, this is AWS S3 specific
// behavior.
if err := send(metaCacheEntry{
name: opts.BaseDir,
metadata: metadata,
}); err != nil {
return err
}
} else {
st, sterr := Lstat(pathJoin(volumeDir, opts.BaseDir, xlStorageFormatFile))
if sterr == nil && st.Mode().IsRegular() {
return errFileNotFound
}
}
}
prefix := opts.FilterPrefix
var scanDir func(path string) error
scanDir = func(current string) error {
// Skip forward, if requested...
sb := bytebufferpool.Get()
defer func() {
sb.Reset()
bytebufferpool.Put(sb)
}()
forward := ""
if len(opts.ForwardTo) > 0 && strings.HasPrefix(opts.ForwardTo, current) {
forward = strings.TrimPrefix(opts.ForwardTo, current)
// Trim further directories and trailing slash.
if idx := strings.IndexByte(forward, '/'); idx > 0 {
forward = forward[:idx]
}
}
if contextCanceled(ctx) {
return ctx.Err()
}
if opts.Limit > 0 && objsReturned >= opts.Limit {
return nil
}
if s.walkMu != nil {
s.walkMu.Lock()
}
entries, err := s.ListDir(ctx, "", opts.Bucket, current, -1)
if s.walkMu != nil {
s.walkMu.Unlock()
}
if err != nil {
// Folder could have gone away in-between
if err != errVolumeNotFound && err != errFileNotFound {
internalLogOnceIf(ctx, err, "metacache-walk-scan-dir")
}
if opts.ReportNotFound && err == errFileNotFound && current == opts.BaseDir {
err = errFileNotFound
} else {
err = nil
}
diskHealthCheckOK(ctx, err)
return err
}
Add local disk health checks (#14447) The main goal of this PR is to solve the situation where disks stop responding to operations. This generally causes an FD build-up and eventually will crash the server. This adds detection of hung disks, where calls on disk get stuck. We add functionality to `xlStorageDiskIDCheck` where it keeps track of the number of concurrent requests on a given disk. A total number of 100 operations are allowed. If this limit is reached we will block (but not reject) new requests, but we will monitor the state of the disk. If no requests have been completed or updated within a 15-second window, we mark the disk as offline. Requests that are blocked will be unblocked and return an error as "faulty disk". New requests will be rejected until the disk is marked OK again. Once a disk has been marked faulty, a check will run every 5 seconds that will attempt to write and read back a file. As long as this fails the disk will remain faulty. To prevent lots of long-running requests to mark the disk faulty we implement a callback feature that allows updating the status as parts of these operations are running. We add a reader and writer wrapper that will update the status of each successful read/write operation. This should allow fine enough granularity that a slow, but still operational disk will not reach 15 seconds where 50 operations have not progressed. Note that errors themselves are not enough to mark a disk faulty. A nil (or io.EOF) error will mark a disk as "good". * Make concurrent disk setting configurable via `_MINIO_DISK_MAX_CONCURRENT`. * de-couple IsOnline() from disk health tracker The purpose of IsOnline() is to ensure that we reconnect the drive only when the "drive" was - disconnected from network we need to validate if the drive is "correct" and is the same drive which belongs to this server. - drive was replaced we have to format it - we support hot swapping of the drives. IsOnline() is not meant for taking the drive offline when it is hung, it is not useful we can let the drive be online instead "return" errors for relevant calls. * return errFaultyDisk for DiskInfo() call Co-authored-by: Harshavardhana <harsha@minio.io> Possible future Improvements: * Unify the REST server and local xlStorageDiskIDCheck. This would also improve stats significantly. * Allow reads/writes to be aborted by the context. * Add usage stats, concurrent count, blocked operations, etc.
2022-03-09 14:38:54 -05:00
diskHealthCheckOK(ctx, err)
if len(entries) == 0 {
return nil
}
dirObjects := make(map[string]struct{})
2023-04-15 13:25:25 -04:00
// Avoid a bunch of cleanup when joining.
current = strings.Trim(current, SlashSeparator)
for i, entry := range entries {
if opts.Limit > 0 && objsReturned >= opts.Limit {
return nil
}
if len(prefix) > 0 && !strings.HasPrefix(entry, prefix) {
// Do not retain the file, since it doesn't
// match the prefix.
entries[i] = ""
continue
}
if len(forward) > 0 && entry < forward {
// Do not retain the file, since its
// lexially smaller than 'forward'
entries[i] = ""
continue
}
2023-04-15 13:25:25 -04:00
if hasSuffixByte(entry, SlashSeparatorChar) {
if strings.HasSuffix(entry, globalDirSuffixWithSlash) {
// Add without extension so it is sorted correctly.
entry = strings.TrimSuffix(entry, globalDirSuffixWithSlash) + slashSeparator
dirObjects[entry] = struct{}{}
entries[i] = entry
continue
}
// Trim slash, since we don't know if this is folder or object.
entries[i] = entries[i][:len(entry)-1]
continue
}
// Do not retain the file.
entries[i] = ""
if contextCanceled(ctx) {
return ctx.Err()
}
// If root was an object return it as such.
if HasSuffix(entry, xlStorageFormatFile) {
var meta metaCacheEntry
if s.walkReadMu != nil {
s.walkReadMu.Lock()
}
meta.metadata, err = s.readMetadata(ctx, pathJoinBuf(sb, volumeDir, current, entry))
if s.walkReadMu != nil {
s.walkReadMu.Unlock()
}
Add local disk health checks (#14447) The main goal of this PR is to solve the situation where disks stop responding to operations. This generally causes an FD build-up and eventually will crash the server. This adds detection of hung disks, where calls on disk get stuck. We add functionality to `xlStorageDiskIDCheck` where it keeps track of the number of concurrent requests on a given disk. A total number of 100 operations are allowed. If this limit is reached we will block (but not reject) new requests, but we will monitor the state of the disk. If no requests have been completed or updated within a 15-second window, we mark the disk as offline. Requests that are blocked will be unblocked and return an error as "faulty disk". New requests will be rejected until the disk is marked OK again. Once a disk has been marked faulty, a check will run every 5 seconds that will attempt to write and read back a file. As long as this fails the disk will remain faulty. To prevent lots of long-running requests to mark the disk faulty we implement a callback feature that allows updating the status as parts of these operations are running. We add a reader and writer wrapper that will update the status of each successful read/write operation. This should allow fine enough granularity that a slow, but still operational disk will not reach 15 seconds where 50 operations have not progressed. Note that errors themselves are not enough to mark a disk faulty. A nil (or io.EOF) error will mark a disk as "good". * Make concurrent disk setting configurable via `_MINIO_DISK_MAX_CONCURRENT`. * de-couple IsOnline() from disk health tracker The purpose of IsOnline() is to ensure that we reconnect the drive only when the "drive" was - disconnected from network we need to validate if the drive is "correct" and is the same drive which belongs to this server. - drive was replaced we have to format it - we support hot swapping of the drives. IsOnline() is not meant for taking the drive offline when it is hung, it is not useful we can let the drive be online instead "return" errors for relevant calls. * return errFaultyDisk for DiskInfo() call Co-authored-by: Harshavardhana <harsha@minio.io> Possible future Improvements: * Unify the REST server and local xlStorageDiskIDCheck. This would also improve stats significantly. * Allow reads/writes to be aborted by the context. * Add usage stats, concurrent count, blocked operations, etc.
2022-03-09 14:38:54 -05:00
diskHealthCheckOK(ctx, err)
if err != nil {
// It is totally possible that xl.meta was overwritten
// while being concurrently listed at the same time in
// such scenarios the 'xl.meta' might get truncated
if !IsErrIgnored(err, io.EOF, io.ErrUnexpectedEOF) {
internalLogOnceIf(ctx, err, "metacache-walk-read-metadata")
}
continue
}
meta.name = strings.TrimSuffix(entry, xlStorageFormatFile)
meta.name = strings.TrimSuffix(meta.name, SlashSeparator)
meta.name = pathJoinBuf(sb, current, meta.name)
meta.name = decodeDirObject(meta.name)
return send(meta)
}
// Check legacy.
if HasSuffix(entry, xlStorageFormatFileV1) && legacy {
var meta metaCacheEntry
meta.metadata, err = xioutil.ReadFile(pathJoinBuf(sb, volumeDir, current, entry))
Add local disk health checks (#14447) The main goal of this PR is to solve the situation where disks stop responding to operations. This generally causes an FD build-up and eventually will crash the server. This adds detection of hung disks, where calls on disk get stuck. We add functionality to `xlStorageDiskIDCheck` where it keeps track of the number of concurrent requests on a given disk. A total number of 100 operations are allowed. If this limit is reached we will block (but not reject) new requests, but we will monitor the state of the disk. If no requests have been completed or updated within a 15-second window, we mark the disk as offline. Requests that are blocked will be unblocked and return an error as "faulty disk". New requests will be rejected until the disk is marked OK again. Once a disk has been marked faulty, a check will run every 5 seconds that will attempt to write and read back a file. As long as this fails the disk will remain faulty. To prevent lots of long-running requests to mark the disk faulty we implement a callback feature that allows updating the status as parts of these operations are running. We add a reader and writer wrapper that will update the status of each successful read/write operation. This should allow fine enough granularity that a slow, but still operational disk will not reach 15 seconds where 50 operations have not progressed. Note that errors themselves are not enough to mark a disk faulty. A nil (or io.EOF) error will mark a disk as "good". * Make concurrent disk setting configurable via `_MINIO_DISK_MAX_CONCURRENT`. * de-couple IsOnline() from disk health tracker The purpose of IsOnline() is to ensure that we reconnect the drive only when the "drive" was - disconnected from network we need to validate if the drive is "correct" and is the same drive which belongs to this server. - drive was replaced we have to format it - we support hot swapping of the drives. IsOnline() is not meant for taking the drive offline when it is hung, it is not useful we can let the drive be online instead "return" errors for relevant calls. * return errFaultyDisk for DiskInfo() call Co-authored-by: Harshavardhana <harsha@minio.io> Possible future Improvements: * Unify the REST server and local xlStorageDiskIDCheck. This would also improve stats significantly. * Allow reads/writes to be aborted by the context. * Add usage stats, concurrent count, blocked operations, etc.
2022-03-09 14:38:54 -05:00
diskHealthCheckOK(ctx, err)
if err != nil {
if !IsErrIgnored(err, io.EOF, io.ErrUnexpectedEOF) {
internalLogIf(ctx, err)
}
continue
}
meta.name = strings.TrimSuffix(entry, xlStorageFormatFileV1)
meta.name = strings.TrimSuffix(meta.name, SlashSeparator)
meta.name = pathJoinBuf(sb, current, meta.name)
return send(meta)
}
// Skip all other files.
}
// Process in sort order.
sort.Strings(entries)
dirStack := make([]string, 0, 5)
prefix = "" // Remove prefix after first level as we have already filtered the list.
if len(forward) > 0 {
// Conservative forwarding. Entries may be either objects or prefixes.
for i, entry := range entries {
if entry >= forward || strings.HasPrefix(forward, entry) {
entries = entries[i:]
break
}
}
}
for _, entry := range entries {
if opts.Limit > 0 && objsReturned >= opts.Limit {
return nil
}
if entry == "" {
continue
}
if contextCanceled(ctx) {
return ctx.Err()
}
meta := metaCacheEntry{name: pathJoinBuf(sb, current, entry)}
// If directory entry on stack before this, pop it now.
for len(dirStack) > 0 && dirStack[len(dirStack)-1] < meta.name {
pop := dirStack[len(dirStack)-1]
select {
case <-ctx.Done():
return ctx.Err()
case out <- metaCacheEntry{name: pop}:
}
if opts.Recursive {
// Scan folder we found. Should be in correct sort order where we are.
err := scanDir(pop)
if err != nil && !IsErrIgnored(err, context.Canceled) {
internalLogIf(ctx, err)
}
}
dirStack = dirStack[:len(dirStack)-1]
}
// All objects will be returned as directories, there has been no object check yet.
// Check it by attempting to read metadata.
_, isDirObj := dirObjects[entry]
if isDirObj {
meta.name = meta.name[:len(meta.name)-1] + globalDirSuffixWithSlash
}
if s.walkReadMu != nil {
s.walkReadMu.Lock()
}
meta.metadata, err = s.readMetadata(ctx, pathJoinBuf(sb, volumeDir, meta.name, xlStorageFormatFile))
if s.walkReadMu != nil {
s.walkReadMu.Unlock()
}
Add local disk health checks (#14447) The main goal of this PR is to solve the situation where disks stop responding to operations. This generally causes an FD build-up and eventually will crash the server. This adds detection of hung disks, where calls on disk get stuck. We add functionality to `xlStorageDiskIDCheck` where it keeps track of the number of concurrent requests on a given disk. A total number of 100 operations are allowed. If this limit is reached we will block (but not reject) new requests, but we will monitor the state of the disk. If no requests have been completed or updated within a 15-second window, we mark the disk as offline. Requests that are blocked will be unblocked and return an error as "faulty disk". New requests will be rejected until the disk is marked OK again. Once a disk has been marked faulty, a check will run every 5 seconds that will attempt to write and read back a file. As long as this fails the disk will remain faulty. To prevent lots of long-running requests to mark the disk faulty we implement a callback feature that allows updating the status as parts of these operations are running. We add a reader and writer wrapper that will update the status of each successful read/write operation. This should allow fine enough granularity that a slow, but still operational disk will not reach 15 seconds where 50 operations have not progressed. Note that errors themselves are not enough to mark a disk faulty. A nil (or io.EOF) error will mark a disk as "good". * Make concurrent disk setting configurable via `_MINIO_DISK_MAX_CONCURRENT`. * de-couple IsOnline() from disk health tracker The purpose of IsOnline() is to ensure that we reconnect the drive only when the "drive" was - disconnected from network we need to validate if the drive is "correct" and is the same drive which belongs to this server. - drive was replaced we have to format it - we support hot swapping of the drives. IsOnline() is not meant for taking the drive offline when it is hung, it is not useful we can let the drive be online instead "return" errors for relevant calls. * return errFaultyDisk for DiskInfo() call Co-authored-by: Harshavardhana <harsha@minio.io> Possible future Improvements: * Unify the REST server and local xlStorageDiskIDCheck. This would also improve stats significantly. * Allow reads/writes to be aborted by the context. * Add usage stats, concurrent count, blocked operations, etc.
2022-03-09 14:38:54 -05:00
diskHealthCheckOK(ctx, err)
switch {
case err == nil:
// It was an object
if isDirObj {
meta.name = strings.TrimSuffix(meta.name, globalDirSuffixWithSlash) + slashSeparator
}
if err := send(meta); err != nil {
return err
}
case osIsNotExist(err), isSysErrIsDir(err):
if legacy {
meta.metadata, err = xioutil.ReadFile(pathJoinBuf(sb, volumeDir, meta.name, xlStorageFormatFileV1))
diskHealthCheckOK(ctx, err)
if err == nil {
// It was an object
if err := send(meta); err != nil {
return err
}
continue
}
}
// NOT an object, append to stack (with slash)
// If dirObject, but no metadata (which is unexpected) we skip it.
if !isDirObj {
if !isDirEmpty(pathJoinBuf(sb, volumeDir, meta.name), legacyFS) {
dirStack = append(dirStack, meta.name+slashSeparator)
}
}
case isSysErrNotDir(err):
// skip
}
}
// If directory entry left on stack, pop it now.
for len(dirStack) > 0 {
if opts.Limit > 0 && objsReturned >= opts.Limit {
return nil
}
if contextCanceled(ctx) {
return ctx.Err()
}
pop := dirStack[len(dirStack)-1]
select {
case <-ctx.Done():
return ctx.Err()
case out <- metaCacheEntry{name: pop}:
}
if opts.Recursive {
// Scan folder we found. Should be in correct sort order where we are.
internalLogIf(ctx, scanDir(pop))
}
dirStack = dirStack[:len(dirStack)-1]
}
return nil
}
// Stream output.
return scanDir(opts.BaseDir)
}
Add local disk health checks (#14447) The main goal of this PR is to solve the situation where disks stop responding to operations. This generally causes an FD build-up and eventually will crash the server. This adds detection of hung disks, where calls on disk get stuck. We add functionality to `xlStorageDiskIDCheck` where it keeps track of the number of concurrent requests on a given disk. A total number of 100 operations are allowed. If this limit is reached we will block (but not reject) new requests, but we will monitor the state of the disk. If no requests have been completed or updated within a 15-second window, we mark the disk as offline. Requests that are blocked will be unblocked and return an error as "faulty disk". New requests will be rejected until the disk is marked OK again. Once a disk has been marked faulty, a check will run every 5 seconds that will attempt to write and read back a file. As long as this fails the disk will remain faulty. To prevent lots of long-running requests to mark the disk faulty we implement a callback feature that allows updating the status as parts of these operations are running. We add a reader and writer wrapper that will update the status of each successful read/write operation. This should allow fine enough granularity that a slow, but still operational disk will not reach 15 seconds where 50 operations have not progressed. Note that errors themselves are not enough to mark a disk faulty. A nil (or io.EOF) error will mark a disk as "good". * Make concurrent disk setting configurable via `_MINIO_DISK_MAX_CONCURRENT`. * de-couple IsOnline() from disk health tracker The purpose of IsOnline() is to ensure that we reconnect the drive only when the "drive" was - disconnected from network we need to validate if the drive is "correct" and is the same drive which belongs to this server. - drive was replaced we have to format it - we support hot swapping of the drives. IsOnline() is not meant for taking the drive offline when it is hung, it is not useful we can let the drive be online instead "return" errors for relevant calls. * return errFaultyDisk for DiskInfo() call Co-authored-by: Harshavardhana <harsha@minio.io> Possible future Improvements: * Unify the REST server and local xlStorageDiskIDCheck. This would also improve stats significantly. * Allow reads/writes to be aborted by the context. * Add usage stats, concurrent count, blocked operations, etc.
2022-03-09 14:38:54 -05:00
func (p *xlStorageDiskIDCheck) WalkDir(ctx context.Context, opts WalkDirOptions, wr io.Writer) (err error) {
perf: websocket grid connectivity for all internode communication (#18461) This PR adds a WebSocket grid feature that allows servers to communicate via a single two-way connection. There are two request types: * Single requests, which are `[]byte => ([]byte, error)`. This is for efficient small roundtrips with small payloads. * Streaming requests which are `[]byte, chan []byte => chan []byte (and error)`, which allows for different combinations of full two-way streams with an initial payload. Only a single stream is created between two machines - and there is, as such, no server/client relation since both sides can initiate and handle requests. Which server initiates the request is decided deterministically on the server names. Requests are made through a mux client and server, which handles message passing, congestion, cancelation, timeouts, etc. If a connection is lost, all requests are canceled, and the calling server will try to reconnect. Registered handlers can operate directly on byte slices or use a higher-level generics abstraction. There is no versioning of handlers/clients, and incompatible changes should be handled by adding new handlers. The request path can be changed to a new one for any protocol changes. First, all servers create a "Manager." The manager must know its address as well as all remote addresses. This will manage all connections. To get a connection to any remote, ask the manager to provide it given the remote address using. ``` func (m *Manager) Connection(host string) *Connection ``` All serverside handlers must also be registered on the manager. This will make sure that all incoming requests are served. The number of in-flight requests and responses must also be given for streaming requests. The "Connection" returned manages the mux-clients. Requests issued to the connection will be sent to the remote. * `func (c *Connection) Request(ctx context.Context, h HandlerID, req []byte) ([]byte, error)` performs a single request and returns the result. Any deadline provided on the request is forwarded to the server, and canceling the context will make the function return at once. * `func (c *Connection) NewStream(ctx context.Context, h HandlerID, payload []byte) (st *Stream, err error)` will initiate a remote call and send the initial payload. ```Go // A Stream is a two-way stream. // All responses *must* be read by the caller. // If the call is canceled through the context, //The appropriate error will be returned. type Stream struct { // Responses from the remote server. // Channel will be closed after an error or when the remote closes. // All responses *must* be read by the caller until either an error is returned or the channel is closed. // Canceling the context will cause the context cancellation error to be returned. Responses <-chan Response // Requests sent to the server. // If the handler is defined with 0 incoming capacity this will be nil. // Channel *must* be closed to signal the end of the stream. // If the request context is canceled, the stream will no longer process requests. Requests chan<- []byte } type Response struct { Msg []byte Err error } ``` There are generic versions of the server/client handlers that allow the use of type safe implementations for data types that support msgpack marshal/unmarshal.
2023-11-20 20:09:35 -05:00
if err := p.checkID(opts.DiskID); err != nil {
return err
}
Add local disk health checks (#14447) The main goal of this PR is to solve the situation where disks stop responding to operations. This generally causes an FD build-up and eventually will crash the server. This adds detection of hung disks, where calls on disk get stuck. We add functionality to `xlStorageDiskIDCheck` where it keeps track of the number of concurrent requests on a given disk. A total number of 100 operations are allowed. If this limit is reached we will block (but not reject) new requests, but we will monitor the state of the disk. If no requests have been completed or updated within a 15-second window, we mark the disk as offline. Requests that are blocked will be unblocked and return an error as "faulty disk". New requests will be rejected until the disk is marked OK again. Once a disk has been marked faulty, a check will run every 5 seconds that will attempt to write and read back a file. As long as this fails the disk will remain faulty. To prevent lots of long-running requests to mark the disk faulty we implement a callback feature that allows updating the status as parts of these operations are running. We add a reader and writer wrapper that will update the status of each successful read/write operation. This should allow fine enough granularity that a slow, but still operational disk will not reach 15 seconds where 50 operations have not progressed. Note that errors themselves are not enough to mark a disk faulty. A nil (or io.EOF) error will mark a disk as "good". * Make concurrent disk setting configurable via `_MINIO_DISK_MAX_CONCURRENT`. * de-couple IsOnline() from disk health tracker The purpose of IsOnline() is to ensure that we reconnect the drive only when the "drive" was - disconnected from network we need to validate if the drive is "correct" and is the same drive which belongs to this server. - drive was replaced we have to format it - we support hot swapping of the drives. IsOnline() is not meant for taking the drive offline when it is hung, it is not useful we can let the drive be online instead "return" errors for relevant calls. * return errFaultyDisk for DiskInfo() call Co-authored-by: Harshavardhana <harsha@minio.io> Possible future Improvements: * Unify the REST server and local xlStorageDiskIDCheck. This would also improve stats significantly. * Allow reads/writes to be aborted by the context. * Add usage stats, concurrent count, blocked operations, etc.
2022-03-09 14:38:54 -05:00
ctx, done, err := p.TrackDiskHealth(ctx, storageMetricWalkDir, opts.Bucket, opts.BaseDir)
if err != nil {
return err
}
Add local disk health checks (#14447) The main goal of this PR is to solve the situation where disks stop responding to operations. This generally causes an FD build-up and eventually will crash the server. This adds detection of hung disks, where calls on disk get stuck. We add functionality to `xlStorageDiskIDCheck` where it keeps track of the number of concurrent requests on a given disk. A total number of 100 operations are allowed. If this limit is reached we will block (but not reject) new requests, but we will monitor the state of the disk. If no requests have been completed or updated within a 15-second window, we mark the disk as offline. Requests that are blocked will be unblocked and return an error as "faulty disk". New requests will be rejected until the disk is marked OK again. Once a disk has been marked faulty, a check will run every 5 seconds that will attempt to write and read back a file. As long as this fails the disk will remain faulty. To prevent lots of long-running requests to mark the disk faulty we implement a callback feature that allows updating the status as parts of these operations are running. We add a reader and writer wrapper that will update the status of each successful read/write operation. This should allow fine enough granularity that a slow, but still operational disk will not reach 15 seconds where 50 operations have not progressed. Note that errors themselves are not enough to mark a disk faulty. A nil (or io.EOF) error will mark a disk as "good". * Make concurrent disk setting configurable via `_MINIO_DISK_MAX_CONCURRENT`. * de-couple IsOnline() from disk health tracker The purpose of IsOnline() is to ensure that we reconnect the drive only when the "drive" was - disconnected from network we need to validate if the drive is "correct" and is the same drive which belongs to this server. - drive was replaced we have to format it - we support hot swapping of the drives. IsOnline() is not meant for taking the drive offline when it is hung, it is not useful we can let the drive be online instead "return" errors for relevant calls. * return errFaultyDisk for DiskInfo() call Co-authored-by: Harshavardhana <harsha@minio.io> Possible future Improvements: * Unify the REST server and local xlStorageDiskIDCheck. This would also improve stats significantly. * Allow reads/writes to be aborted by the context. * Add usage stats, concurrent count, blocked operations, etc.
2022-03-09 14:38:54 -05:00
defer done(&err)
return p.storage.WalkDir(ctx, opts, wr)
}
// WalkDir will traverse a directory and return all entries found.
// On success a meta cache stream will be returned, that should be closed when done.
func (client *storageRESTClient) WalkDir(ctx context.Context, opts WalkDirOptions, wr io.Writer) error {
perf: websocket grid connectivity for all internode communication (#18461) This PR adds a WebSocket grid feature that allows servers to communicate via a single two-way connection. There are two request types: * Single requests, which are `[]byte => ([]byte, error)`. This is for efficient small roundtrips with small payloads. * Streaming requests which are `[]byte, chan []byte => chan []byte (and error)`, which allows for different combinations of full two-way streams with an initial payload. Only a single stream is created between two machines - and there is, as such, no server/client relation since both sides can initiate and handle requests. Which server initiates the request is decided deterministically on the server names. Requests are made through a mux client and server, which handles message passing, congestion, cancelation, timeouts, etc. If a connection is lost, all requests are canceled, and the calling server will try to reconnect. Registered handlers can operate directly on byte slices or use a higher-level generics abstraction. There is no versioning of handlers/clients, and incompatible changes should be handled by adding new handlers. The request path can be changed to a new one for any protocol changes. First, all servers create a "Manager." The manager must know its address as well as all remote addresses. This will manage all connections. To get a connection to any remote, ask the manager to provide it given the remote address using. ``` func (m *Manager) Connection(host string) *Connection ``` All serverside handlers must also be registered on the manager. This will make sure that all incoming requests are served. The number of in-flight requests and responses must also be given for streaming requests. The "Connection" returned manages the mux-clients. Requests issued to the connection will be sent to the remote. * `func (c *Connection) Request(ctx context.Context, h HandlerID, req []byte) ([]byte, error)` performs a single request and returns the result. Any deadline provided on the request is forwarded to the server, and canceling the context will make the function return at once. * `func (c *Connection) NewStream(ctx context.Context, h HandlerID, payload []byte) (st *Stream, err error)` will initiate a remote call and send the initial payload. ```Go // A Stream is a two-way stream. // All responses *must* be read by the caller. // If the call is canceled through the context, //The appropriate error will be returned. type Stream struct { // Responses from the remote server. // Channel will be closed after an error or when the remote closes. // All responses *must* be read by the caller until either an error is returned or the channel is closed. // Canceling the context will cause the context cancellation error to be returned. Responses <-chan Response // Requests sent to the server. // If the handler is defined with 0 incoming capacity this will be nil. // Channel *must* be closed to signal the end of the stream. // If the request context is canceled, the stream will no longer process requests. Requests chan<- []byte } type Response struct { Msg []byte Err error } ``` There are generic versions of the server/client handlers that allow the use of type safe implementations for data types that support msgpack marshal/unmarshal.
2023-11-20 20:09:35 -05:00
// Ensure remote has the same disk ID.
opts.DiskID = client.diskID
b, err := opts.MarshalMsg(grid.GetByteBuffer()[:0])
if err != nil {
return toStorageErr(err)
}
perf: websocket grid connectivity for all internode communication (#18461) This PR adds a WebSocket grid feature that allows servers to communicate via a single two-way connection. There are two request types: * Single requests, which are `[]byte => ([]byte, error)`. This is for efficient small roundtrips with small payloads. * Streaming requests which are `[]byte, chan []byte => chan []byte (and error)`, which allows for different combinations of full two-way streams with an initial payload. Only a single stream is created between two machines - and there is, as such, no server/client relation since both sides can initiate and handle requests. Which server initiates the request is decided deterministically on the server names. Requests are made through a mux client and server, which handles message passing, congestion, cancelation, timeouts, etc. If a connection is lost, all requests are canceled, and the calling server will try to reconnect. Registered handlers can operate directly on byte slices or use a higher-level generics abstraction. There is no versioning of handlers/clients, and incompatible changes should be handled by adding new handlers. The request path can be changed to a new one for any protocol changes. First, all servers create a "Manager." The manager must know its address as well as all remote addresses. This will manage all connections. To get a connection to any remote, ask the manager to provide it given the remote address using. ``` func (m *Manager) Connection(host string) *Connection ``` All serverside handlers must also be registered on the manager. This will make sure that all incoming requests are served. The number of in-flight requests and responses must also be given for streaming requests. The "Connection" returned manages the mux-clients. Requests issued to the connection will be sent to the remote. * `func (c *Connection) Request(ctx context.Context, h HandlerID, req []byte) ([]byte, error)` performs a single request and returns the result. Any deadline provided on the request is forwarded to the server, and canceling the context will make the function return at once. * `func (c *Connection) NewStream(ctx context.Context, h HandlerID, payload []byte) (st *Stream, err error)` will initiate a remote call and send the initial payload. ```Go // A Stream is a two-way stream. // All responses *must* be read by the caller. // If the call is canceled through the context, //The appropriate error will be returned. type Stream struct { // Responses from the remote server. // Channel will be closed after an error or when the remote closes. // All responses *must* be read by the caller until either an error is returned or the channel is closed. // Canceling the context will cause the context cancellation error to be returned. Responses <-chan Response // Requests sent to the server. // If the handler is defined with 0 incoming capacity this will be nil. // Channel *must* be closed to signal the end of the stream. // If the request context is canceled, the stream will no longer process requests. Requests chan<- []byte } type Response struct { Msg []byte Err error } ``` There are generic versions of the server/client handlers that allow the use of type safe implementations for data types that support msgpack marshal/unmarshal.
2023-11-20 20:09:35 -05:00
st, err := client.gridConn.NewStream(ctx, grid.HandlerWalkDir, b)
if err != nil {
return toStorageErr(err)
}
perf: websocket grid connectivity for all internode communication (#18461) This PR adds a WebSocket grid feature that allows servers to communicate via a single two-way connection. There are two request types: * Single requests, which are `[]byte => ([]byte, error)`. This is for efficient small roundtrips with small payloads. * Streaming requests which are `[]byte, chan []byte => chan []byte (and error)`, which allows for different combinations of full two-way streams with an initial payload. Only a single stream is created between two machines - and there is, as such, no server/client relation since both sides can initiate and handle requests. Which server initiates the request is decided deterministically on the server names. Requests are made through a mux client and server, which handles message passing, congestion, cancelation, timeouts, etc. If a connection is lost, all requests are canceled, and the calling server will try to reconnect. Registered handlers can operate directly on byte slices or use a higher-level generics abstraction. There is no versioning of handlers/clients, and incompatible changes should be handled by adding new handlers. The request path can be changed to a new one for any protocol changes. First, all servers create a "Manager." The manager must know its address as well as all remote addresses. This will manage all connections. To get a connection to any remote, ask the manager to provide it given the remote address using. ``` func (m *Manager) Connection(host string) *Connection ``` All serverside handlers must also be registered on the manager. This will make sure that all incoming requests are served. The number of in-flight requests and responses must also be given for streaming requests. The "Connection" returned manages the mux-clients. Requests issued to the connection will be sent to the remote. * `func (c *Connection) Request(ctx context.Context, h HandlerID, req []byte) ([]byte, error)` performs a single request and returns the result. Any deadline provided on the request is forwarded to the server, and canceling the context will make the function return at once. * `func (c *Connection) NewStream(ctx context.Context, h HandlerID, payload []byte) (st *Stream, err error)` will initiate a remote call and send the initial payload. ```Go // A Stream is a two-way stream. // All responses *must* be read by the caller. // If the call is canceled through the context, //The appropriate error will be returned. type Stream struct { // Responses from the remote server. // Channel will be closed after an error or when the remote closes. // All responses *must* be read by the caller until either an error is returned or the channel is closed. // Canceling the context will cause the context cancellation error to be returned. Responses <-chan Response // Requests sent to the server. // If the handler is defined with 0 incoming capacity this will be nil. // Channel *must* be closed to signal the end of the stream. // If the request context is canceled, the stream will no longer process requests. Requests chan<- []byte } type Response struct { Msg []byte Err error } ``` There are generic versions of the server/client handlers that allow the use of type safe implementations for data types that support msgpack marshal/unmarshal.
2023-11-20 20:09:35 -05:00
return toStorageErr(st.Results(func(in []byte) error {
_, err := wr.Write(in)
return err
}))
}
perf: websocket grid connectivity for all internode communication (#18461) This PR adds a WebSocket grid feature that allows servers to communicate via a single two-way connection. There are two request types: * Single requests, which are `[]byte => ([]byte, error)`. This is for efficient small roundtrips with small payloads. * Streaming requests which are `[]byte, chan []byte => chan []byte (and error)`, which allows for different combinations of full two-way streams with an initial payload. Only a single stream is created between two machines - and there is, as such, no server/client relation since both sides can initiate and handle requests. Which server initiates the request is decided deterministically on the server names. Requests are made through a mux client and server, which handles message passing, congestion, cancelation, timeouts, etc. If a connection is lost, all requests are canceled, and the calling server will try to reconnect. Registered handlers can operate directly on byte slices or use a higher-level generics abstraction. There is no versioning of handlers/clients, and incompatible changes should be handled by adding new handlers. The request path can be changed to a new one for any protocol changes. First, all servers create a "Manager." The manager must know its address as well as all remote addresses. This will manage all connections. To get a connection to any remote, ask the manager to provide it given the remote address using. ``` func (m *Manager) Connection(host string) *Connection ``` All serverside handlers must also be registered on the manager. This will make sure that all incoming requests are served. The number of in-flight requests and responses must also be given for streaming requests. The "Connection" returned manages the mux-clients. Requests issued to the connection will be sent to the remote. * `func (c *Connection) Request(ctx context.Context, h HandlerID, req []byte) ([]byte, error)` performs a single request and returns the result. Any deadline provided on the request is forwarded to the server, and canceling the context will make the function return at once. * `func (c *Connection) NewStream(ctx context.Context, h HandlerID, payload []byte) (st *Stream, err error)` will initiate a remote call and send the initial payload. ```Go // A Stream is a two-way stream. // All responses *must* be read by the caller. // If the call is canceled through the context, //The appropriate error will be returned. type Stream struct { // Responses from the remote server. // Channel will be closed after an error or when the remote closes. // All responses *must* be read by the caller until either an error is returned or the channel is closed. // Canceling the context will cause the context cancellation error to be returned. Responses <-chan Response // Requests sent to the server. // If the handler is defined with 0 incoming capacity this will be nil. // Channel *must* be closed to signal the end of the stream. // If the request context is canceled, the stream will no longer process requests. Requests chan<- []byte } type Response struct { Msg []byte Err error } ``` There are generic versions of the server/client handlers that allow the use of type safe implementations for data types that support msgpack marshal/unmarshal.
2023-11-20 20:09:35 -05:00
// WalkDirHandler - remote caller to list files and folders in a requested directory path.
func (s *storageRESTServer) WalkDirHandler(ctx context.Context, payload []byte, _ <-chan []byte, out chan<- []byte) (gerr *grid.RemoteErr) {
var opts WalkDirOptions
_, err := opts.UnmarshalMsg(payload)
if err != nil {
return grid.NewRemoteErr(err)
}
if !s.checkID(opts.DiskID) {
return grid.NewRemoteErr(errDiskNotFound)
}
perf: websocket grid connectivity for all internode communication (#18461) This PR adds a WebSocket grid feature that allows servers to communicate via a single two-way connection. There are two request types: * Single requests, which are `[]byte => ([]byte, error)`. This is for efficient small roundtrips with small payloads. * Streaming requests which are `[]byte, chan []byte => chan []byte (and error)`, which allows for different combinations of full two-way streams with an initial payload. Only a single stream is created between two machines - and there is, as such, no server/client relation since both sides can initiate and handle requests. Which server initiates the request is decided deterministically on the server names. Requests are made through a mux client and server, which handles message passing, congestion, cancelation, timeouts, etc. If a connection is lost, all requests are canceled, and the calling server will try to reconnect. Registered handlers can operate directly on byte slices or use a higher-level generics abstraction. There is no versioning of handlers/clients, and incompatible changes should be handled by adding new handlers. The request path can be changed to a new one for any protocol changes. First, all servers create a "Manager." The manager must know its address as well as all remote addresses. This will manage all connections. To get a connection to any remote, ask the manager to provide it given the remote address using. ``` func (m *Manager) Connection(host string) *Connection ``` All serverside handlers must also be registered on the manager. This will make sure that all incoming requests are served. The number of in-flight requests and responses must also be given for streaming requests. The "Connection" returned manages the mux-clients. Requests issued to the connection will be sent to the remote. * `func (c *Connection) Request(ctx context.Context, h HandlerID, req []byte) ([]byte, error)` performs a single request and returns the result. Any deadline provided on the request is forwarded to the server, and canceling the context will make the function return at once. * `func (c *Connection) NewStream(ctx context.Context, h HandlerID, payload []byte) (st *Stream, err error)` will initiate a remote call and send the initial payload. ```Go // A Stream is a two-way stream. // All responses *must* be read by the caller. // If the call is canceled through the context, //The appropriate error will be returned. type Stream struct { // Responses from the remote server. // Channel will be closed after an error or when the remote closes. // All responses *must* be read by the caller until either an error is returned or the channel is closed. // Canceling the context will cause the context cancellation error to be returned. Responses <-chan Response // Requests sent to the server. // If the handler is defined with 0 incoming capacity this will be nil. // Channel *must* be closed to signal the end of the stream. // If the request context is canceled, the stream will no longer process requests. Requests chan<- []byte } type Response struct { Msg []byte Err error } ``` There are generic versions of the server/client handlers that allow the use of type safe implementations for data types that support msgpack marshal/unmarshal.
2023-11-20 20:09:35 -05:00
ctx, cancel := context.WithCancel(ctx)
defer cancel()
return grid.NewRemoteErr(s.getStorage().WalkDir(ctx, opts, grid.WriterToChannel(ctx, out)))
}