Files
seaweedFS/weed/worker/tasks/vacuum/vacuum.go
Chris Lu 891a2fb6eb Admin: misc improvements on admin server and workers. EC now works. (#7055)
* initial design

* added simulation as tests

* reorganized the codebase to move the simulation framework and tests into their own dedicated package

* integration test. ec worker task

* remove "enhanced" reference

* start master, volume servers, filer

Current Status
 Master: Healthy and running (port 9333)
 Filer: Healthy and running (port 8888)
 Volume Servers: All 6 servers running (ports 8080-8085)
🔄 Admin/Workers: Will start when dependencies are ready

* generate write load

* tasks are assigned

* admin start wtih grpc port. worker has its own working directory

* Update .gitignore

* working worker and admin. Task detection is not working yet.

* compiles, detection uses volumeSizeLimitMB from master

* compiles

* worker retries connecting to admin

* build and restart

* rendering pending tasks

* skip task ID column

* sticky worker id

* test canScheduleTaskNow

* worker reconnect to admin

* clean up logs

* worker register itself first

* worker can run ec work and report status

but:
1. one volume should not be repeatedly worked on.
2. ec shards needs to be distributed and source data should be deleted.

* move ec task logic

* listing ec shards

* local copy, ec. Need to distribute.

* ec is mostly working now

* distribution of ec shards needs improvement
* need configuration to enable ec

* show ec volumes

* interval field UI component

* rename

* integration test with vauuming

* garbage percentage threshold

* fix warning

* display ec shard sizes

* fix ec volumes list

* Update ui.go

* show default values

* ensure correct default value

* MaintenanceConfig use ConfigField

* use schema defined defaults

* config

* reduce duplication

* refactor to use BaseUIProvider

* each task register its schema

* checkECEncodingCandidate use ecDetector

* use vacuumDetector

* use volumeSizeLimitMB

* remove

remove

* remove unused

* refactor

* use new framework

* remove v2 reference

* refactor

* left menu can scroll now

* The maintenance manager was not being initialized when no data directory was configured for persistent storage.

* saving config

* Update task_config_schema_templ.go

* enable/disable tasks

* protobuf encoded task configurations

* fix system settings

* use ui component

* remove logs

* interface{} Reduction

* reduce interface{}

* reduce interface{}

* avoid from/to map

* reduce interface{}

* refactor

* keep it DRY

* added logging

* debug messages

* debug level

* debug

* show the log caller line

* use configured task policy

* log level

* handle admin heartbeat response

* Update worker.go

* fix EC rack and dc count

* Report task status to admin server

* fix task logging, simplify interface checking, use erasure_coding constants

* factor in empty volume server during task planning

* volume.list adds disk id

* track disk id also

* fix locking scheduled and manual scanning

* add active topology

* simplify task detector

* ec task completed, but shards are not showing up

* implement ec in ec_typed.go

* adjust log level

* dedup

* implementing ec copying shards and only ecx files

* use disk id when distributing ec shards

🎯 Planning: ActiveTopology creates DestinationPlan with specific TargetDisk
📦 Task Creation: maintenance_integration.go creates ECDestination with DiskId
🚀 Task Execution: EC task passes DiskId in VolumeEcShardsCopyRequest
💾 Volume Server: Receives disk_id and stores shards on specific disk (vs.store.Locations[req.DiskId])
📂 File System: EC shards and metadata land in the exact disk directory planned

* Delete original volume from all locations

* clean up existing shard locations

* local encoding and distributing

* Update docker/admin_integration/EC-TESTING-README.md

Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>

* check volume id range

* simplify

* fix tests

* fix types

* clean up logs and tests

---------

Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
2025-07-30 12:38:03 -07:00

215 lines
6.6 KiB
Go

package vacuum
import (
"context"
"fmt"
"io"
"time"
"github.com/seaweedfs/seaweedfs/weed/pb"
"github.com/seaweedfs/seaweedfs/weed/pb/volume_server_pb"
"github.com/seaweedfs/seaweedfs/weed/worker/tasks"
"github.com/seaweedfs/seaweedfs/weed/worker/types"
"google.golang.org/grpc"
"google.golang.org/grpc/credentials/insecure"
)
// Task implements vacuum operation to reclaim disk space
type Task struct {
*tasks.BaseTask
server string
volumeID uint32
garbageThreshold float64
}
// NewTask creates a new vacuum task instance
func NewTask(server string, volumeID uint32) *Task {
task := &Task{
BaseTask: tasks.NewBaseTask(types.TaskTypeVacuum),
server: server,
volumeID: volumeID,
garbageThreshold: 0.3, // Default 30% threshold
}
return task
}
// Execute performs the vacuum operation
func (t *Task) Execute(params types.TaskParams) error {
// Use BaseTask.ExecuteTask to handle logging initialization
return t.ExecuteTask(context.Background(), params, t.executeImpl)
}
// executeImpl is the actual vacuum implementation
func (t *Task) executeImpl(ctx context.Context, params types.TaskParams) error {
t.LogInfo("Starting vacuum for volume %d on server %s", t.volumeID, t.server)
// Parse garbage threshold from typed parameters
if params.TypedParams != nil {
if vacuumParams := params.TypedParams.GetVacuumParams(); vacuumParams != nil {
t.garbageThreshold = vacuumParams.GarbageThreshold
t.LogWithFields("INFO", "Using garbage threshold from parameters", map[string]interface{}{
"threshold": t.garbageThreshold,
})
}
}
// Convert server address to gRPC address and use proper dial option
grpcAddress := pb.ServerToGrpcAddress(t.server)
var dialOpt grpc.DialOption = grpc.WithTransportCredentials(insecure.NewCredentials())
if params.GrpcDialOption != nil {
dialOpt = params.GrpcDialOption
}
conn, err := grpc.NewClient(grpcAddress, dialOpt)
if err != nil {
t.LogError("Failed to connect to volume server %s: %v", t.server, err)
return fmt.Errorf("failed to connect to volume server %s: %v", t.server, err)
}
defer conn.Close()
client := volume_server_pb.NewVolumeServerClient(conn)
// Step 1: Check vacuum eligibility
t.SetProgress(10.0)
t.LogDebug("Checking vacuum eligibility for volume %d", t.volumeID)
checkResp, err := client.VacuumVolumeCheck(ctx, &volume_server_pb.VacuumVolumeCheckRequest{
VolumeId: t.volumeID,
})
if err != nil {
t.LogError("Vacuum check failed for volume %d: %v", t.volumeID, err)
return fmt.Errorf("vacuum check failed for volume %d: %v", t.volumeID, err)
}
// Check if garbage ratio meets threshold
if checkResp.GarbageRatio < t.garbageThreshold {
t.LogWarning("Volume %d garbage ratio %.2f%% is below threshold %.2f%%, skipping vacuum",
t.volumeID, checkResp.GarbageRatio*100, t.garbageThreshold*100)
return fmt.Errorf("volume %d garbage ratio %.2f%% is below threshold %.2f%%, skipping vacuum",
t.volumeID, checkResp.GarbageRatio*100, t.garbageThreshold*100)
}
t.LogWithFields("INFO", "Volume eligible for vacuum", map[string]interface{}{
"volume_id": t.volumeID,
"garbage_ratio": checkResp.GarbageRatio,
"threshold": t.garbageThreshold,
"garbage_percent": checkResp.GarbageRatio * 100,
})
// Step 2: Compact volume
t.SetProgress(30.0)
t.LogInfo("Starting compact for volume %d", t.volumeID)
compactStream, err := client.VacuumVolumeCompact(ctx, &volume_server_pb.VacuumVolumeCompactRequest{
VolumeId: t.volumeID,
})
if err != nil {
t.LogError("Vacuum compact failed for volume %d: %v", t.volumeID, err)
return fmt.Errorf("vacuum compact failed for volume %d: %v", t.volumeID, err)
}
// Process compact stream and track progress
var processedBytes int64
var totalBytes int64
for {
resp, err := compactStream.Recv()
if err != nil {
if err == io.EOF {
break
}
t.LogError("Vacuum compact stream error for volume %d: %v", t.volumeID, err)
return fmt.Errorf("vacuum compact stream error for volume %d: %v", t.volumeID, err)
}
processedBytes = resp.ProcessedBytes
if resp.LoadAvg_1M > 0 {
totalBytes = int64(resp.LoadAvg_1M) // This is a rough approximation
}
// Update progress based on processed bytes (30% to 70% of total progress)
if totalBytes > 0 {
compactProgress := float64(processedBytes) / float64(totalBytes)
if compactProgress > 1.0 {
compactProgress = 1.0
}
progress := 30.0 + (compactProgress * 40.0) // 30% to 70%
t.SetProgress(progress)
}
t.LogWithFields("DEBUG", "Volume compact progress", map[string]interface{}{
"volume_id": t.volumeID,
"processed_bytes": processedBytes,
"total_bytes": totalBytes,
"compact_progress": fmt.Sprintf("%.1f%%", (float64(processedBytes)/float64(totalBytes))*100),
})
}
// Step 3: Commit vacuum changes
t.SetProgress(80.0)
t.LogInfo("Committing vacuum for volume %d", t.volumeID)
commitResp, err := client.VacuumVolumeCommit(ctx, &volume_server_pb.VacuumVolumeCommitRequest{
VolumeId: t.volumeID,
})
if err != nil {
t.LogError("Vacuum commit failed for volume %d: %v", t.volumeID, err)
return fmt.Errorf("vacuum commit failed for volume %d: %v", t.volumeID, err)
}
// Step 4: Cleanup temporary files
t.SetProgress(90.0)
t.LogInfo("Cleaning up vacuum files for volume %d", t.volumeID)
_, err = client.VacuumVolumeCleanup(ctx, &volume_server_pb.VacuumVolumeCleanupRequest{
VolumeId: t.volumeID,
})
if err != nil {
// Log warning but don't fail the task
t.LogWarning("Vacuum cleanup warning for volume %d: %v", t.volumeID, err)
}
t.SetProgress(100.0)
newVolumeSize := commitResp.VolumeSize
t.LogWithFields("INFO", "Successfully completed vacuum", map[string]interface{}{
"volume_id": t.volumeID,
"server": t.server,
"new_volume_size": newVolumeSize,
"garbage_reclaimed": true,
})
return nil
}
// Validate validates the task parameters
func (t *Task) Validate(params types.TaskParams) error {
if params.VolumeID == 0 {
return fmt.Errorf("volume_id is required")
}
if params.Server == "" {
return fmt.Errorf("server is required")
}
return nil
}
// EstimateTime estimates the time needed for the task
func (t *Task) EstimateTime(params types.TaskParams) time.Duration {
// Base time for vacuum operations - varies by volume size and garbage ratio
// Typically vacuum is faster than EC encoding
baseTime := 5 * time.Minute
// Use default estimation since volume size is not available in typed params
return baseTime
}
// GetProgress returns the current progress
func (t *Task) GetProgress() float64 {
return t.BaseTask.GetProgress()
}
// Cancel cancels the task
func (t *Task) Cancel() error {
return t.BaseTask.Cancel()
}