Files
seaweedfs/weed/admin/dash/plugin_api.go
T
Chris Lu 9a7cfaf371 admin: stop holding the shell admin lock across whole scheduler batches (#10290)
* admin: let waiting shell clients win the admin lock

The admin lock manager re-acquired the cluster shell lock immediately
after releasing it, while a waiting weed shell client only polls the
master once a second, so an operator could lose the race indefinitely.
Leave the lock free for slightly more than one poll interval before
re-acquiring, and once overlapping reference-counted holds have kept
the lock continuously held for over a minute, make new admin-side
acquires wait for a full release before piggybacking.

* admin/plugin: take the shell lock per detection and per job, not per batch

The default-lane scheduler acquired the cluster shell lock once and
held it across the whole pass: every due job type's detection plus all
of its dispatched jobs, each drained to completion. A manual weed
shell operation sharing that lock could stall behind the batch for up
to the extended execution window.

Hold the lock only while it protects something: around each detection
scan and around each dispatched job. Manual shell operations now wait
for at most one in-flight job. The admin UI detect+execute path stops
wrapping dispatch in an outer hold, since a nested acquire would
deadlock against the lock manager's fairness window; detection and
per-job dispatch take the lock themselves.

* admin/plugin: stop extending the dispatch window to the largest job estimate

When any proposal's estimated runtime exceeded the remaining
JobTypeMaxRuntime, the whole dispatch context was replaced with a
fresh one capped at eight hours, so a single balance backlog could
hold the default lane (and with it erasure_coding and vacuum
detection) for that long.

Keep the dispatch window at JobTypeMaxRuntime and instead detach each
started attempt onto its own estimated-runtime deadline. Large jobs
still get their full time once started; jobs not yet started when the
window closes are canceled and re-proposed by a later detection, so
sibling job types get a turn every window.

* worker/balance: re-check each planned move against the master before executing

A balance plan is computed at detection time, but the admin lock is
released between detection and execution, so a manual shell operation
can rearrange the volume in the gap. The task's own guards catch a
vanished source, but a target that gained a replica in the meantime
would be silently overwritten by VolumeCopy and the source delete
would then reduce the volume to a single copy.

Before executing each move, ask the master for the volume's current
locations (uncached) and skip the move if the volume has left the
source or the target already holds a replica. Skipped moves fail with
a stale-move error and the next detection replans them. Without master
addresses in the cluster context the check is skipped, preserving the
old behavior with older admins.

* admin: block re-acquire while the final lock release is in flight

Release dropped the manager mutex before calling ReleaseLock, so a
concurrent Acquire could see hold count zero and call RequestLock
while the locker still considered itself locked. That request no-ops,
leaving a hold with no live master lease. Track the in-flight release
and make Acquire wait for it.

Also normalize a nil release function from lock manager
implementations, and make the yield and fairness windows per-instance
fields so tests stop mutating globals.
2026-07-09 20:43:55 -07:00

985 lines
30 KiB
Go

package dash
import (
"context"
"encoding/json"
"errors"
"fmt"
"io"
"net/http"
"sort"
"strconv"
"strings"
"time"
"github.com/gorilla/mux"
"github.com/seaweedfs/seaweedfs/weed/admin/plugin"
"github.com/seaweedfs/seaweedfs/weed/cluster"
"github.com/seaweedfs/seaweedfs/weed/glog"
"github.com/seaweedfs/seaweedfs/weed/pb/master_pb"
"github.com/seaweedfs/seaweedfs/weed/pb/plugin_pb"
"google.golang.org/protobuf/encoding/protojson"
"google.golang.org/protobuf/proto"
"google.golang.org/protobuf/types/known/timestamppb"
)
const (
defaultPluginDetectionTimeout = 45 * time.Second
defaultPluginExecutionTimeout = 90 * time.Second
maxPluginDetectionTimeout = 5 * time.Minute
maxPluginExecutionTimeout = 10 * time.Minute
defaultPluginRunTimeout = 5 * time.Minute
maxPluginRunTimeout = 30 * time.Minute
)
func matchesPluginLane(jobType, laneFilter string) bool {
laneFilter = strings.TrimSpace(laneFilter)
if laneFilter == "" {
return true
}
return plugin.JobTypeLane(jobType) == plugin.SchedulerLane(laneFilter)
}
func filterTrackedJobsByLane(jobs []plugin.TrackedJob, laneFilter string) []plugin.TrackedJob {
if strings.TrimSpace(laneFilter) == "" {
return jobs
}
filtered := make([]plugin.TrackedJob, 0, len(jobs))
for _, job := range jobs {
if matchesPluginLane(job.JobType, laneFilter) {
filtered = append(filtered, job)
}
}
return filtered
}
func filterActivitiesByLane(activities []plugin.JobActivity, laneFilter string) []plugin.JobActivity {
if strings.TrimSpace(laneFilter) == "" {
return activities
}
filtered := make([]plugin.JobActivity, 0, len(activities))
for _, activity := range activities {
if matchesPluginLane(activity.JobType, laneFilter) {
filtered = append(filtered, activity)
}
}
return filtered
}
// GetPluginStatusAPI returns plugin status.
func (s *AdminServer) GetPluginStatusAPI(w http.ResponseWriter, r *http.Request) {
plugin := s.GetPlugin()
if plugin == nil {
writeJSON(w, http.StatusOK, map[string]interface{}{
"enabled": false,
"worker_grpc_port": s.GetWorkerGrpcPort(),
})
return
}
writeJSON(w, http.StatusOK, map[string]interface{}{
"enabled": true,
"configured": plugin.IsConfigured(),
"base_dir": plugin.BaseDir(),
"worker_count": len(plugin.ListWorkers()),
"worker_grpc_port": s.GetWorkerGrpcPort(),
})
}
// GetPluginWorkersAPI returns currently connected plugin workers.
// Accepts an optional ?lane= query parameter to filter by scheduler lane.
func (s *AdminServer) GetPluginWorkersAPI(w http.ResponseWriter, r *http.Request) {
workers := s.GetPluginWorkers()
if workers == nil {
writeJSON(w, http.StatusOK, []interface{}{})
return
}
laneFilter := strings.TrimSpace(r.URL.Query().Get("lane"))
if laneFilter != "" {
lane := plugin.SchedulerLane(laneFilter)
filtered := make([]*plugin.WorkerSession, 0, len(workers))
for _, worker := range workers {
for jobType := range worker.Capabilities {
if plugin.JobTypeLane(jobType) == lane {
filtered = append(filtered, worker)
break
}
}
}
writeJSON(w, http.StatusOK, filtered)
return
}
writeJSON(w, http.StatusOK, workers)
}
// GetPluginJobTypesAPI returns known plugin job types from workers and persisted data.
// Accepts an optional ?lane= query parameter to filter by scheduler lane.
func (s *AdminServer) GetPluginJobTypesAPI(w http.ResponseWriter, r *http.Request) {
jobTypes, err := s.ListPluginJobTypes()
if err != nil {
writeJSONError(w, http.StatusInternalServerError, err.Error())
return
}
if jobTypes == nil {
writeJSON(w, http.StatusOK, []interface{}{})
return
}
laneFilter := strings.TrimSpace(r.URL.Query().Get("lane"))
if laneFilter != "" {
lane := plugin.SchedulerLane(laneFilter)
filtered := make([]plugin.JobTypeInfo, 0, len(jobTypes))
for _, jt := range jobTypes {
if plugin.JobTypeLane(jt.JobType) == lane {
filtered = append(filtered, jt)
}
}
writeJSON(w, http.StatusOK, filtered)
return
}
writeJSON(w, http.StatusOK, jobTypes)
}
// GetPluginJobsAPI returns tracked jobs for monitoring.
func (s *AdminServer) GetPluginJobsAPI(w http.ResponseWriter, r *http.Request) {
query := r.URL.Query()
jobType := strings.TrimSpace(query.Get("job_type"))
state := strings.TrimSpace(query.Get("state"))
laneFilter := strings.TrimSpace(query.Get("lane"))
limit := parsePositiveInt(query.Get("limit"), 200)
jobs := s.ListPluginJobs(jobType, state, limit)
if jobs == nil {
writeJSON(w, http.StatusOK, []interface{}{})
return
}
writeJSON(w, http.StatusOK, filterTrackedJobsByLane(jobs, laneFilter))
}
// GetPluginJobAPI returns one tracked job.
func (s *AdminServer) GetPluginJobAPI(w http.ResponseWriter, r *http.Request) {
jobID := strings.TrimSpace(mux.Vars(r)["jobId"])
if jobID == "" {
writeJSONError(w, http.StatusBadRequest, "jobId is required")
return
}
job, found := s.GetPluginJob(jobID)
if !found {
writeJSONError(w, http.StatusNotFound, "job not found")
return
}
writeJSON(w, http.StatusOK, job)
}
// GetPluginJobDetailAPI returns detailed information for one tracked plugin job.
func (s *AdminServer) GetPluginJobDetailAPI(w http.ResponseWriter, r *http.Request) {
jobID := strings.TrimSpace(mux.Vars(r)["jobId"])
if jobID == "" {
writeJSONError(w, http.StatusBadRequest, "jobId is required")
return
}
query := r.URL.Query()
activityLimit := parsePositiveInt(query.Get("activity_limit"), 500)
relatedLimit := parsePositiveInt(query.Get("related_limit"), 20)
detail, found, err := s.GetPluginJobDetail(jobID, activityLimit, relatedLimit)
if err != nil {
writeJSONError(w, http.StatusInternalServerError, err.Error())
return
}
if !found || detail == nil {
writeJSONError(w, http.StatusNotFound, "job detail not found")
return
}
writeJSON(w, http.StatusOK, detail)
}
// ExpirePluginJobAPI marks a job as failed so it no longer blocks scheduling.
func (s *AdminServer) ExpirePluginJobAPI(w http.ResponseWriter, r *http.Request) {
jobID := strings.TrimSpace(mux.Vars(r)["jobId"])
if jobID == "" {
writeJSONError(w, http.StatusBadRequest, "jobId is required")
return
}
var req struct {
Reason string `json:"reason"`
}
if err := decodeJSONBody(newJSONMaxReader(w, r), &req); err != nil && err != io.EOF {
writeJSONError(w, http.StatusBadRequest, "invalid request body: "+err.Error())
return
}
job, expired, err := s.ExpirePluginJob(jobID, req.Reason)
if err != nil {
if errors.Is(err, plugin.ErrJobNotFound) {
writeJSONError(w, http.StatusNotFound, err.Error())
return
}
writeJSONError(w, http.StatusInternalServerError, err.Error())
return
}
response := map[string]interface{}{
"job_id": jobID,
"expired": expired,
}
if job != nil {
response["job"] = job
}
if !expired {
response["message"] = "job is not active"
}
writeJSON(w, http.StatusOK, response)
}
// GetPluginActivitiesAPI returns recent plugin activities.
func (s *AdminServer) GetPluginActivitiesAPI(w http.ResponseWriter, r *http.Request) {
query := r.URL.Query()
jobType := strings.TrimSpace(query.Get("job_type"))
laneFilter := strings.TrimSpace(query.Get("lane"))
limit := parsePositiveInt(query.Get("limit"), 500)
activities := s.ListPluginActivities(jobType, limit)
if activities == nil {
writeJSON(w, http.StatusOK, []interface{}{})
return
}
writeJSON(w, http.StatusOK, filterActivitiesByLane(activities, laneFilter))
}
// GetPluginSchedulerStatesAPI returns per-job-type scheduler status for monitoring.
// Accepts optional ?job_type= and ?lane= query parameters.
func (s *AdminServer) GetPluginSchedulerStatesAPI(w http.ResponseWriter, r *http.Request) {
jobTypeFilter := strings.TrimSpace(r.URL.Query().Get("job_type"))
laneFilter := strings.TrimSpace(r.URL.Query().Get("lane"))
states, err := s.ListPluginSchedulerStates()
if err != nil {
writeJSONError(w, http.StatusInternalServerError, err.Error())
return
}
if jobTypeFilter != "" || laneFilter != "" {
filtered := make([]interface{}, 0, len(states))
for _, state := range states {
if jobTypeFilter != "" && state.JobType != jobTypeFilter {
continue
}
if laneFilter != "" && state.Lane != laneFilter {
continue
}
filtered = append(filtered, state)
}
writeJSON(w, http.StatusOK, filtered)
return
}
if states == nil {
writeJSON(w, http.StatusOK, []interface{}{})
return
}
writeJSON(w, http.StatusOK, states)
}
// GetPluginSchedulerStatusAPI returns scheduler status including in-process jobs and lock state.
// Accepts optional ?lane= query parameter to scope to a specific lane.
func (s *AdminServer) GetPluginSchedulerStatusAPI(w http.ResponseWriter, r *http.Request) {
pluginSvc := s.GetPlugin()
if pluginSvc == nil {
writeJSON(w, http.StatusOK, map[string]interface{}{
"enabled": false,
})
return
}
laneFilter := strings.TrimSpace(r.URL.Query().Get("lane"))
if laneFilter != "" {
lane := plugin.SchedulerLane(laneFilter)
response := map[string]interface{}{
"enabled": true,
"scheduler": pluginSvc.GetLaneSchedulerStatus(lane),
}
if s.pluginLock != nil {
response["lock"] = s.pluginLock.Status()
}
writeJSON(w, http.StatusOK, response)
return
}
response := map[string]interface{}{
"enabled": true,
"scheduler": pluginSvc.GetSchedulerStatus(),
}
if s.pluginLock != nil {
response["lock"] = s.pluginLock.Status()
}
writeJSON(w, http.StatusOK, response)
}
// GetPluginLanesAPI returns all scheduler lanes and their current status.
func (s *AdminServer) GetPluginLanesAPI(w http.ResponseWriter, r *http.Request) {
pluginSvc := s.GetPlugin()
if pluginSvc == nil {
writeJSON(w, http.StatusOK, []interface{}{})
return
}
lanes := plugin.AllLanes()
result := make([]map[string]interface{}, 0, len(lanes))
for _, lane := range lanes {
laneStatus := pluginSvc.GetLaneSchedulerStatus(lane)
result = append(result, map[string]interface{}{
"lane": string(lane),
"idle_sleep_sec": int(plugin.LaneIdleSleep(lane) / time.Second),
"job_types": plugin.LaneJobTypes(lane),
"status": laneStatus,
})
}
writeJSON(w, http.StatusOK, result)
}
// RequestPluginJobTypeSchemaAPI asks a worker for one job type schema.
func (s *AdminServer) RequestPluginJobTypeSchemaAPI(w http.ResponseWriter, r *http.Request) {
jobType := strings.TrimSpace(mux.Vars(r)["jobType"])
if jobType == "" {
writeJSONError(w, http.StatusBadRequest, "jobType is required")
return
}
forceRefresh := strings.EqualFold(r.URL.Query().Get("force_refresh"), "true")
ctx, cancel := context.WithTimeout(r.Context(), defaultPluginDetectionTimeout)
defer cancel()
descriptor, err := s.RequestPluginJobTypeDescriptor(ctx, jobType, forceRefresh)
if err != nil {
writeJSONError(w, http.StatusInternalServerError, err.Error())
return
}
renderProtoJSON(w, http.StatusOK, descriptor)
}
// GetPluginJobTypeDescriptorAPI returns persisted descriptor for a job type.
func (s *AdminServer) GetPluginJobTypeDescriptorAPI(w http.ResponseWriter, r *http.Request) {
jobType := strings.TrimSpace(mux.Vars(r)["jobType"])
if jobType == "" {
writeJSONError(w, http.StatusBadRequest, "jobType is required")
return
}
descriptor, err := s.LoadPluginJobTypeDescriptor(jobType)
if err != nil {
writeJSONError(w, http.StatusInternalServerError, err.Error())
return
}
if descriptor == nil {
writeJSONError(w, http.StatusNotFound, "descriptor not found")
return
}
renderProtoJSON(w, http.StatusOK, descriptor)
}
// GetPluginJobTypeConfigAPI loads persisted config for a job type.
func (s *AdminServer) GetPluginJobTypeConfigAPI(w http.ResponseWriter, r *http.Request) {
jobType := strings.TrimSpace(mux.Vars(r)["jobType"])
if jobType == "" {
writeJSONError(w, http.StatusBadRequest, "jobType is required")
return
}
config, err := s.LoadPluginJobTypeConfig(jobType)
if err != nil {
writeJSONError(w, http.StatusInternalServerError, err.Error())
return
}
if config == nil {
config = &plugin_pb.PersistedJobTypeConfig{
JobType: jobType,
AdminConfigValues: map[string]*plugin_pb.ConfigValue{},
WorkerConfigValues: map[string]*plugin_pb.ConfigValue{},
AdminRuntime: &plugin_pb.AdminRuntimeConfig{},
}
}
if descriptor, err := s.LoadPluginJobTypeDescriptor(jobType); err == nil && descriptor != nil {
applyDescriptorDefaultsToPersistedConfig(config, descriptor)
}
renderProtoJSON(w, http.StatusOK, config)
}
// UpdatePluginJobTypeConfigAPI stores persisted config for a job type.
func (s *AdminServer) UpdatePluginJobTypeConfigAPI(w http.ResponseWriter, r *http.Request) {
jobType := strings.TrimSpace(mux.Vars(r)["jobType"])
if jobType == "" {
writeJSONError(w, http.StatusBadRequest, "jobType is required")
return
}
config := &plugin_pb.PersistedJobTypeConfig{}
if err := parseProtoJSONBody(w, r, config); err != nil {
writeJSONError(w, http.StatusBadRequest, err.Error())
return
}
config.JobType = jobType
if config.UpdatedAt == nil {
config.UpdatedAt = timestamppb.Now()
}
if config.AdminRuntime == nil {
config.AdminRuntime = &plugin_pb.AdminRuntimeConfig{}
}
if config.AdminConfigValues == nil {
config.AdminConfigValues = map[string]*plugin_pb.ConfigValue{}
}
if config.WorkerConfigValues == nil {
config.WorkerConfigValues = map[string]*plugin_pb.ConfigValue{}
}
username := UsernameFromContext(r.Context())
if username == "" {
username = "admin"
}
config.UpdatedBy = username
// Reapply descriptor defaults so a save from an older form (or a UI
// that omits new fields) cannot silently clear a baseline like
// execution_timeout_seconds back to zero.
if descriptor, err := s.LoadPluginJobTypeDescriptor(jobType); err == nil && descriptor != nil {
applyDescriptorDefaultsToPersistedConfig(config, descriptor)
}
if err := s.SavePluginJobTypeConfig(config); err != nil {
writeJSONError(w, http.StatusInternalServerError, err.Error())
return
}
renderProtoJSON(w, http.StatusOK, config)
}
// GetPluginRunHistoryAPI returns bounded run history for a job type.
func (s *AdminServer) GetPluginRunHistoryAPI(w http.ResponseWriter, r *http.Request) {
jobType := strings.TrimSpace(mux.Vars(r)["jobType"])
if jobType == "" {
writeJSONError(w, http.StatusBadRequest, "jobType is required")
return
}
history, err := s.GetPluginRunHistory(jobType)
if err != nil {
writeJSONError(w, http.StatusInternalServerError, err.Error())
return
}
if history == nil {
writeJSON(w, http.StatusOK, map[string]interface{}{
"job_type": jobType,
"successful_runs": []interface{}{},
"error_runs": []interface{}{},
"last_updated_time": nil,
})
return
}
writeJSON(w, http.StatusOK, history)
}
// TriggerPluginDetectionAPI runs one detector for this job type and returns proposals.
func (s *AdminServer) TriggerPluginDetectionAPI(w http.ResponseWriter, r *http.Request) {
jobType := strings.TrimSpace(mux.Vars(r)["jobType"])
if jobType == "" {
writeJSONError(w, http.StatusBadRequest, "jobType is required")
return
}
var req struct {
ClusterContext json.RawMessage `json:"cluster_context"`
MaxResults int32 `json:"max_results"`
TimeoutSeconds int `json:"timeout_seconds"`
}
if err := decodeJSONBody(newJSONMaxReader(w, r), &req); err != nil && err != io.EOF {
writeJSONError(w, http.StatusBadRequest, "invalid request body: "+err.Error())
return
}
clusterContext, err := s.parseOrBuildClusterContext(req.ClusterContext)
if err != nil {
writeJSONError(w, http.StatusBadRequest, err.Error())
return
}
timeout := normalizeTimeout(req.TimeoutSeconds, defaultPluginDetectionTimeout, maxPluginDetectionTimeout)
ctx, cancel := context.WithTimeout(r.Context(), timeout)
defer cancel()
report, err := s.RunPluginDetectionWithReport(ctx, jobType, clusterContext, req.MaxResults)
proposals := make([]*plugin_pb.JobProposal, 0)
requestID := ""
detectorWorkerID := ""
totalProposals := int32(0)
if report != nil {
proposals = report.Proposals
requestID = report.RequestID
detectorWorkerID = report.WorkerID
if report.Complete != nil {
totalProposals = report.Complete.TotalProposals
}
}
proposalPayloads := make([]map[string]interface{}, 0, len(proposals))
for _, proposal := range proposals {
payload, marshalErr := protoMessageToMap(proposal)
if marshalErr != nil {
glog.Warningf("failed to marshal proposal for jobType=%s: %v", jobType, marshalErr)
continue
}
proposalPayloads = append(proposalPayloads, payload)
}
sort.Slice(proposalPayloads, func(i, j int) bool {
iPriorityStr, _ := proposalPayloads[i]["priority"].(string)
jPriorityStr, _ := proposalPayloads[j]["priority"].(string)
iPriority := plugin_pb.JobPriority_value[iPriorityStr]
jPriority := plugin_pb.JobPriority_value[jPriorityStr]
if iPriority != jPriority {
return iPriority > jPriority
}
iID, _ := proposalPayloads[i]["proposal_id"].(string)
jID, _ := proposalPayloads[j]["proposal_id"].(string)
return iID < jID
})
activities := s.ListPluginActivities(jobType, 500)
filteredActivities := make([]interface{}, 0, len(activities))
if requestID != "" {
for i := len(activities) - 1; i >= 0; i-- {
activity := activities[i]
if activity.RequestID != requestID {
continue
}
filteredActivities = append(filteredActivities, activity)
}
}
response := map[string]interface{}{
"job_type": jobType,
"request_id": requestID,
"detector_worker_id": detectorWorkerID,
"total_proposals": totalProposals,
"count": len(proposalPayloads),
"proposals": proposalPayloads,
"activities": filteredActivities,
}
if err != nil {
response["error"] = err.Error()
writeJSON(w, http.StatusInternalServerError, response)
return
}
writeJSON(w, http.StatusOK, response)
}
// RunPluginJobTypeAPI runs full workflow for one job type: detect then dispatch detected jobs.
// RunPluginDetection and the dispatch path take the cluster admin lock
// themselves (around detection and around each job), so no lock is held
// across the whole workflow here.
func (s *AdminServer) RunPluginJobTypeAPI(w http.ResponseWriter, r *http.Request) {
jobType := strings.TrimSpace(mux.Vars(r)["jobType"])
if jobType == "" {
writeJSONError(w, http.StatusBadRequest, "jobType is required")
return
}
var req struct {
ClusterContext json.RawMessage `json:"cluster_context"`
MaxResults int32 `json:"max_results"`
TimeoutSeconds int `json:"timeout_seconds"`
Attempt int32 `json:"attempt"`
}
if err := decodeJSONBody(newJSONMaxReader(w, r), &req); err != nil && err != io.EOF {
writeJSONError(w, http.StatusBadRequest, "invalid request body: "+err.Error())
return
}
if req.Attempt < 1 {
req.Attempt = 1
}
clusterContext, err := s.parseOrBuildClusterContext(req.ClusterContext)
if err != nil {
writeJSONError(w, http.StatusBadRequest, err.Error())
return
}
timeout := normalizeTimeout(req.TimeoutSeconds, defaultPluginRunTimeout, maxPluginRunTimeout)
ctx, cancel := context.WithTimeout(r.Context(), timeout)
defer cancel()
proposals, err := s.RunPluginDetection(ctx, jobType, clusterContext, req.MaxResults)
if err != nil {
writeJSONError(w, http.StatusInternalServerError, err.Error())
return
}
detectedCount := len(proposals)
filteredProposals, skippedActiveCount, err := s.FilterPluginProposalsWithActiveJobs(jobType, proposals)
if err != nil {
writeJSONError(w, http.StatusInternalServerError, err.Error())
return
}
successCount, errorCount, canceledCount, dispatchErr := s.DispatchPluginProposals(ctx, jobType, filteredProposals, clusterContext)
if dispatchErr != nil {
writeJSONError(w, http.StatusInternalServerError, dispatchErr.Error())
return
}
writeJSON(w, http.StatusOK, map[string]interface{}{
"job_type": jobType,
"detected_count": detectedCount,
"ready_to_execute_count": len(filteredProposals),
"skipped_active_count": skippedActiveCount,
"executed_count": successCount + errorCount + canceledCount,
"success_count": successCount,
"error_count": errorCount,
"canceled_count": canceledCount,
})
}
// ExecutePluginJobAPI executes one job on a capable worker and waits for completion.
func (s *AdminServer) ExecutePluginJobAPI(w http.ResponseWriter, r *http.Request) {
var req struct {
Job json.RawMessage `json:"job"`
ClusterContext json.RawMessage `json:"cluster_context"`
Attempt int32 `json:"attempt"`
TimeoutSeconds int `json:"timeout_seconds"`
}
if err := decodeJSONBody(newJSONMaxReader(w, r), &req); err != nil {
writeJSONError(w, http.StatusBadRequest, "invalid request body: "+err.Error())
return
}
if len(req.Job) == 0 {
writeJSONError(w, http.StatusBadRequest, "job is required")
return
}
job := &plugin_pb.JobSpec{}
if err := (protojson.UnmarshalOptions{DiscardUnknown: true}).Unmarshal(req.Job, job); err != nil {
writeJSONError(w, http.StatusBadRequest, "invalid job payload: "+err.Error())
return
}
clusterContext, err := s.parseOrBuildClusterContext(req.ClusterContext)
if err != nil {
writeJSONError(w, http.StatusBadRequest, err.Error())
return
}
if req.Attempt < 1 {
req.Attempt = 1
}
timeout := normalizeTimeout(req.TimeoutSeconds, defaultPluginExecutionTimeout, maxPluginExecutionTimeout)
ctx, cancel := context.WithTimeout(r.Context(), timeout)
defer cancel()
completed, err := s.ExecutePluginJob(ctx, job, clusterContext, req.Attempt)
if err != nil {
if completed != nil {
payload, marshalErr := protoMessageToMap(completed)
if marshalErr == nil {
writeJSON(w, http.StatusInternalServerError, map[string]interface{}{"error": err.Error(), "completion": payload})
return
}
}
writeJSONError(w, http.StatusInternalServerError, err.Error())
return
}
renderProtoJSON(w, http.StatusOK, completed)
}
func (s *AdminServer) parseOrBuildClusterContext(raw json.RawMessage) (*plugin_pb.ClusterContext, error) {
if len(raw) == 0 {
return s.buildDefaultPluginClusterContext(), nil
}
contextMessage := &plugin_pb.ClusterContext{}
if err := (protojson.UnmarshalOptions{DiscardUnknown: true}).Unmarshal(raw, contextMessage); err != nil {
return nil, fmt.Errorf("invalid cluster_context payload: %w", err)
}
fallback := s.buildDefaultPluginClusterContext()
if len(contextMessage.MasterGrpcAddresses) == 0 {
contextMessage.MasterGrpcAddresses = append(contextMessage.MasterGrpcAddresses, fallback.MasterGrpcAddresses...)
}
if len(contextMessage.FilerAddresses) == 0 {
contextMessage.FilerAddresses = append(contextMessage.FilerAddresses, fallback.FilerAddresses...)
}
if len(contextMessage.VolumeGrpcAddresses) == 0 {
contextMessage.VolumeGrpcAddresses = append(contextMessage.VolumeGrpcAddresses, fallback.VolumeGrpcAddresses...)
}
if len(contextMessage.S3GrpcAddresses) == 0 {
contextMessage.S3GrpcAddresses = append(contextMessage.S3GrpcAddresses, fallback.S3GrpcAddresses...)
}
if contextMessage.Metadata == nil {
contextMessage.Metadata = map[string]string{}
}
contextMessage.Metadata["source"] = "admin"
return contextMessage, nil
}
func (s *AdminServer) buildDefaultPluginClusterContext() *plugin_pb.ClusterContext {
clusterContext := &plugin_pb.ClusterContext{
MasterGrpcAddresses: make([]string, 0),
FilerAddresses: make([]string, 0),
VolumeGrpcAddresses: make([]string, 0),
S3GrpcAddresses: make([]string, 0),
Metadata: map[string]string{
"source": "admin",
},
}
masterAddress := string(s.masterClient.GetMaster(context.Background()))
if masterAddress != "" {
clusterContext.MasterGrpcAddresses = append(clusterContext.MasterGrpcAddresses, masterAddress)
}
// Master returns filers in pb.ServerAddress form (host:httpPort.grpcPort).
// Forward that verbatim; each worker converts to a gRPC or HTTP address as
// it needs (dialing wants gRPC, the admin shell wants the ServerAddress).
filerSeen := map[string]struct{}{}
for _, filer := range s.GetAllFilers() {
filer = strings.TrimSpace(filer)
if filer == "" {
continue
}
if _, exists := filerSeen[filer]; exists {
continue
}
filerSeen[filer] = struct{}{}
clusterContext.FilerAddresses = append(clusterContext.FilerAddresses, filer)
}
volumeSeen := map[string]struct{}{}
if volumeServers, err := s.GetClusterVolumeServers(); err == nil {
for _, server := range volumeServers.VolumeServers {
address := strings.TrimSpace(server.GetDisplayAddress())
if address == "" {
address = strings.TrimSpace(server.Address)
}
if address == "" {
continue
}
if _, exists := volumeSeen[address]; exists {
continue
}
volumeSeen[address] = struct{}{}
clusterContext.VolumeGrpcAddresses = append(clusterContext.VolumeGrpcAddresses, address)
}
} else {
glog.V(1).Infof("failed to build default plugin volume context: %v", err)
}
s3Seen := map[string]struct{}{}
if err := s.WithMasterClient(func(client master_pb.SeaweedClient) error {
resp, err := client.ListClusterNodes(context.Background(), s.listClusterNodesRequest(cluster.S3Type))
if err != nil {
return err
}
for _, node := range resp.GetClusterNodes() {
address := strings.TrimSpace(node.GetAddress())
if address == "" {
continue
}
if _, exists := s3Seen[address]; exists {
continue
}
s3Seen[address] = struct{}{}
clusterContext.S3GrpcAddresses = append(clusterContext.S3GrpcAddresses, address)
}
return nil
}); err != nil {
glog.V(1).Infof("failed to list cluster S3 nodes: %v", err)
}
sort.Strings(clusterContext.MasterGrpcAddresses)
sort.Strings(clusterContext.FilerAddresses)
sort.Strings(clusterContext.VolumeGrpcAddresses)
sort.Strings(clusterContext.S3GrpcAddresses)
return clusterContext
}
const parseProtoJSONBodyMaxBytes = 1 << 20 // 1 MB
func parseProtoJSONBody(w http.ResponseWriter, r *http.Request, message proto.Message) error {
limitedBody := http.MaxBytesReader(w, r.Body, parseProtoJSONBodyMaxBytes)
data, err := io.ReadAll(limitedBody)
if err != nil {
return fmt.Errorf("failed to read request body: %w", err)
}
if len(data) == 0 {
return fmt.Errorf("request body is empty")
}
if err := (protojson.UnmarshalOptions{DiscardUnknown: true}).Unmarshal(data, message); err != nil {
return fmt.Errorf("invalid protobuf json: %w", err)
}
return nil
}
func renderProtoJSON(w http.ResponseWriter, statusCode int, message proto.Message) {
payload, err := protojson.MarshalOptions{
UseProtoNames: true,
EmitUnpopulated: true,
}.Marshal(message)
if err != nil {
writeJSONError(w, http.StatusInternalServerError, "failed to encode response: "+err.Error())
return
}
w.Header().Set("Content-Type", "application/json")
w.WriteHeader(statusCode)
_, _ = w.Write(payload)
}
func protoMessageToMap(message proto.Message) (map[string]interface{}, error) {
payload, err := protojson.MarshalOptions{UseProtoNames: true}.Marshal(message)
if err != nil {
return nil, err
}
out := map[string]interface{}{}
if err := json.Unmarshal(payload, &out); err != nil {
return nil, err
}
return out, nil
}
func normalizeTimeout(timeoutSeconds int, defaultTimeout, maxTimeout time.Duration) time.Duration {
if timeoutSeconds <= 0 {
return defaultTimeout
}
timeout := time.Duration(timeoutSeconds) * time.Second
if timeout > maxTimeout {
return maxTimeout
}
return timeout
}
func applyDescriptorDefaultsToPersistedConfig(
config *plugin_pb.PersistedJobTypeConfig,
descriptor *plugin_pb.JobTypeDescriptor,
) {
if config == nil || descriptor == nil {
return
}
if config.AdminConfigValues == nil {
config.AdminConfigValues = map[string]*plugin_pb.ConfigValue{}
}
if config.WorkerConfigValues == nil {
config.WorkerConfigValues = map[string]*plugin_pb.ConfigValue{}
}
if config.AdminRuntime == nil {
config.AdminRuntime = &plugin_pb.AdminRuntimeConfig{}
}
if descriptor.AdminConfigForm != nil {
for key, value := range descriptor.AdminConfigForm.DefaultValues {
if value == nil {
continue
}
current := config.AdminConfigValues[key]
if current == nil {
config.AdminConfigValues[key] = proto.Clone(value).(*plugin_pb.ConfigValue)
continue
}
if strings.EqualFold(descriptor.JobType, "admin_script") &&
key == "script" &&
isBlankStringConfigValue(current) {
config.AdminConfigValues[key] = proto.Clone(value).(*plugin_pb.ConfigValue)
}
}
}
if descriptor.WorkerConfigForm != nil {
for key, value := range descriptor.WorkerConfigForm.DefaultValues {
if value == nil {
continue
}
if config.WorkerConfigValues[key] != nil {
continue
}
config.WorkerConfigValues[key] = proto.Clone(value).(*plugin_pb.ConfigValue)
}
}
if descriptor.AdminRuntimeDefaults != nil {
runtime := config.AdminRuntime
defaults := descriptor.AdminRuntimeDefaults
if runtime.DetectionIntervalMinutes <= 0 {
runtime.DetectionIntervalMinutes = defaults.DetectionIntervalMinutes
}
if runtime.DetectionTimeoutSeconds <= 0 {
runtime.DetectionTimeoutSeconds = defaults.DetectionTimeoutSeconds
}
if runtime.MaxJobsPerDetection <= 0 {
runtime.MaxJobsPerDetection = defaults.MaxJobsPerDetection
}
if runtime.GlobalExecutionConcurrency <= 0 {
runtime.GlobalExecutionConcurrency = defaults.GlobalExecutionConcurrency
}
if runtime.PerWorkerExecutionConcurrency <= 0 {
runtime.PerWorkerExecutionConcurrency = defaults.PerWorkerExecutionConcurrency
}
if runtime.JobTypeMaxRuntimeSeconds <= 0 {
runtime.JobTypeMaxRuntimeSeconds = defaults.JobTypeMaxRuntimeSeconds
}
if runtime.ExecutionTimeoutSeconds <= 0 {
runtime.ExecutionTimeoutSeconds = defaults.ExecutionTimeoutSeconds
}
if runtime.RetryBackoffSeconds <= 0 {
runtime.RetryBackoffSeconds = defaults.RetryBackoffSeconds
}
if runtime.RetryLimit < 0 {
runtime.RetryLimit = defaults.RetryLimit
}
}
}
func isBlankStringConfigValue(value *plugin_pb.ConfigValue) bool {
if value == nil {
return true
}
kind, ok := value.Kind.(*plugin_pb.ConfigValue_StringValue)
if !ok {
return false
}
return strings.TrimSpace(kind.StringValue) == ""
}
func parsePositiveInt(raw string, defaultValue int) int {
value, err := strconv.Atoi(strings.TrimSpace(raw))
if err != nil || value <= 0 {
return defaultValue
}
return value
}
// cloneConfigValueMap is now exported by the plugin package as CloneConfigValueMap