Files
seaweedfs/weed/shell/commands.go
T
Chris LuandGitHub 902a12fd6f wdclient: bound the wait for a master leader by the caller's context (#11002)
* wdclient: bound the wait for a master leader by the caller's context

WithClient waited on GetMaster with context.Background(), so a caller that
arrived while no master leader was known parked in a 200ms poll loop until one
appeared, whatever deadline it had already set on the RPC. Each retry above it
then left another goroutine in the same wait.

Take the context in WithClient and WithClientCustomGetMaster and hand it to
GetMaster, and stop the retry loop once it is done. The dial keeps
context.Background(): fn brings its own RPC context, so a cancellation seen
here cannot be attributed to the shared connection.

Call sites pass whatever they hold: the request context in the filer's
CollectionList, DeleteCollection and Statistics handlers and in the credential
store's propagation, the operation context in the shell's s3.bucket.delete and
the kafka gateway's broker and filer discovery, and context.Background() where
there is none - the shell commands, the admin dashboard wrapper, and the
exclusive locker's initial lease. The locker's release keeps its own
uncancelled context so a slow unlock cannot turn into a ghost lock.

Claude-Session: https://claude.ai/code/session_01BjDWtZsCoZY6x4pdDmGWxU

* wdclient: test that WithClient gives up with the caller's context

Claude-Session: https://claude.ai/code/session_01BjDWtZsCoZY6x4pdDmGWxU

* wdclient: cut the master retry backoff short when the caller gives up

util.Retry sleeps unconditionally between attempts, so a transient error
arriving just before the caller's deadline still cost it a full backoff step.
Use the context-aware util.RetryWithBackoff, the same helper the volume lookup
in this file already uses.

Two call sites went with it: the shell's lock-holder lookup builds its three
second bound before WithClient so it also covers finding the leader, as its
comment already promised, and the filer's post-delete collection cleanup goes
back to an uncancelled context - the entry is already gone, so a caller that
hung up must not leave the collection behind.

Claude-Session: https://claude.ai/code/session_01BjDWtZsCoZY6x4pdDmGWxU

* wdclient: test that a cancel during backoff ends the retry

Claude-Session: https://claude.ai/code/session_01BjDWtZsCoZY6x4pdDmGWxU
2026-08-27 22:27:45 -07:00

224 lines
5.9 KiB
Go

package shell
import (
"context"
"fmt"
"io"
"strings"
"time"
"github.com/seaweedfs/seaweedfs/weed/cluster"
"github.com/seaweedfs/seaweedfs/weed/operation"
"github.com/seaweedfs/seaweedfs/weed/pb/master_pb"
"github.com/seaweedfs/seaweedfs/weed/pb/volume_server_pb"
"github.com/seaweedfs/seaweedfs/weed/storage/needle_map"
"google.golang.org/grpc"
"github.com/seaweedfs/seaweedfs/weed/pb"
"github.com/seaweedfs/seaweedfs/weed/pb/filer_pb"
"github.com/seaweedfs/seaweedfs/weed/util"
"github.com/seaweedfs/seaweedfs/weed/wdclient"
"github.com/seaweedfs/seaweedfs/weed/wdclient/exclusive_locks"
)
type ShellOptions struct {
Masters *string
GrpcDialOption grpc.DialOption
// shell transient context
FilerHost string
FilerPort int64
FilerGroup *string
FilerAddress pb.ServerAddress
Directory string
Debug bool
}
type CommandEnv struct {
env map[string]string
MasterClient *wdclient.MasterClient
option *ShellOptions
locker *exclusive_locks.ExclusiveLocker
noLock bool
forceNoLock bool
verbose bool
}
func NewCommandEnv(options *ShellOptions) *CommandEnv {
ce := &CommandEnv{
env: make(map[string]string),
MasterClient: wdclient.NewMasterClient(options.GrpcDialOption, *options.FilerGroup, pb.AdminShellClient, "", "", "", *pb.ServerAddresses(*options.Masters).ToServiceDiscovery()),
option: options,
noLock: false,
}
ce.locker = exclusive_locks.NewExclusiveLocker(ce.MasterClient, cluster.AdminShellLockName)
return ce
}
func (ce *CommandEnv) parseUrl(input string) (path string, err error) {
if strings.HasPrefix(input, "http") {
err = fmt.Errorf("http://<filer>:<port> prefix is not supported any more")
return
}
if !strings.HasPrefix(input, "/") {
input = util.Join(ce.option.Directory, input)
}
return input, err
}
func (ce *CommandEnv) isDirectory(path string) bool {
return ce.checkDirectory(path) == nil
}
func (ce *CommandEnv) confirmIsLocked(args []string) error {
if ce.noLock || ce.forceNoLock {
return nil
}
if ce.locker.IsLocked() {
return nil
}
ce.locker.SetMessage(fmt.Sprintf("%v", args))
return fmt.Errorf("need to run \"lock\" first to continue")
}
func (ce *CommandEnv) SetNoLock(noLock bool) {
if ce == nil {
return
}
ce.noLock = noLock
}
func (ce *CommandEnv) ForceNoLock() {
if ce == nil {
return
}
ce.forceNoLock = true
}
func (ce *CommandEnv) isLocked() bool {
if ce == nil {
return true
}
if ce.noLock || ce.forceNoLock {
return true
}
return ce.locker.IsLocked()
}
// shellLockHolder asks the master who currently holds the cluster-wide shell
// lock. Best effort: masters without GetAdminLockStatus report no holder, and
// the lookup is bounded so an unresponsive master cannot hang the shell.
func (ce *CommandEnv) shellLockHolder() (clientName string, message string, held bool) {
attemptCtx, cancel := context.WithTimeout(context.Background(), 3*time.Second)
defer cancel()
ce.MasterClient.WithClient(attemptCtx, false, func(client master_pb.SeaweedClient) error {
resp, err := client.GetAdminLockStatus(attemptCtx, &master_pb.GetAdminLockStatusRequest{
LockName: cluster.AdminShellLockName,
})
if err == nil && resp.IsLocked {
clientName, message, held = resp.ClientName, resp.Message, true
}
return err
})
return
}
func (ce *CommandEnv) checkDirectory(path string) error {
dir, name := util.FullPath(path).DirAndName()
exists, err := filer_pb.Exists(context.Background(), ce, dir, name, true)
if !exists {
return fmt.Errorf("%s is not a directory", path)
}
return err
}
var _ = filer_pb.FilerClient(&CommandEnv{})
func (ce *CommandEnv) WithFilerClient(streamingMode bool, fn func(filer_pb.SeaweedFilerClient) error) error {
return pb.WithGrpcFilerClient(streamingMode, 0, ce.option.FilerAddress, ce.option.GrpcDialOption, fn)
}
func (ce *CommandEnv) AdjustedUrl(location *filer_pb.Location) string {
return location.Url
}
func (ce *CommandEnv) GetDataCenter() string {
return ce.MasterClient.GetDataCenter()
}
func findInputDirectory(args []string) (input string) {
input = "."
if len(args) > 0 {
input = args[len(args)-1]
if strings.HasPrefix(input, "-") {
input = "."
}
}
return input
}
// isHelpRequest checks if the args contain a help flag (-h, --help, or -help)
// It also handles combined short flags like -lh or -hl
func isHelpRequest(args []string) bool {
for _, arg := range args {
// Check for exact matches
if arg == "-h" || arg == "--help" || arg == "-help" {
return true
}
// Check for combined short flags (e.g., -lh, -hl, -rfh)
// Limit to reasonable length (2-4 chars total) to avoid matching long options like -verbose
if strings.HasPrefix(arg, "-") && !strings.HasPrefix(arg, "--") && len(arg) > 1 && len(arg) <= 4 {
for _, char := range arg[1:] {
if char == 'h' {
return true
}
}
}
}
return false
}
// handleHelpRequest checks for help flags and prints the help message if requested.
// It returns true if the help message was printed, indicating the command should exit.
func handleHelpRequest(c command, args []string, writer io.Writer) bool {
if isHelpRequest(args) {
fmt.Fprintln(writer, c.Help())
return true
}
return false
}
func readNeedleStatus(grpcDialOption grpc.DialOption, sourceVolumeServer pb.ServerAddress, volumeId uint32, needleValue needle_map.NeedleValue) (resp *volume_server_pb.VolumeNeedleStatusResponse, err error) {
err = operation.WithVolumeServerClient(false, sourceVolumeServer, grpcDialOption,
func(client volume_server_pb.VolumeServerClient) error {
if resp, err = client.VolumeNeedleStatus(context.Background(), &volume_server_pb.VolumeNeedleStatusRequest{
VolumeId: volumeId,
NeedleId: uint64(needleValue.Key),
}); err != nil {
return err
}
return nil
},
)
return
}
func getCollectionName(commandEnv *CommandEnv, bucket string) string {
if *commandEnv.option.FilerGroup != "" {
return fmt.Sprintf("%s_%s", *commandEnv.option.FilerGroup, bucket)
}
return bucket
}