Files
seaweedfs/weed/admin/handlers/cluster_handlers.go
T
Chris Lu 75ec5ec193 admin: allow setting volume read-only and read/write modes (#11217)
* admin: support setting volume read-only and read/write modes

* admin: address PR review on volume access-mode persistence

Reject trailing JSON values in the SetVolumeReadOnly handler so
requests like {"read_only":true}{} no longer pass validation, and add
a trailing-value case to the invalid-request test.

Propagate .vif persistence failures through the access-mode chain.
PersistReadOnly now returns the SaveVolumeInfo error and rolls back
the in-memory volumeInfo on failure; Store.MarkVolumeReadonly and
Store.MarkVolumeWritable propagate that error and roll back their
noWrite flags, so the API reports failure instead of success while
restart would revert the mode.

* admin: make .vif persistence atomic and preserve error chain

SaveVolumeInfo now writes to a .vif.tmp file, syncs it, renames it
over the target, and fsyncs the directory. A write/sync/close failure
leaves the existing .vif intact, so the PersistReadOnly in-memory
rollback matches the durable state instead of diverging from a
partially written file that restart would apply.

Switch the error wrappers in PersistReadOnly, MarkVolumeReadonly, and
MarkVolumeWritable from %v to %w so callers can use errors.Is and
errors.As to classify persistence failures.

* admin: treat post-rename dir fsync failure as a warning

After os.Rename commits the new .vif, the on-disk file already holds
the requested mode. A directory fsync failure only risks losing the
rename across a crash; returning an error here would make
PersistReadOnly roll back in-memory state while the durable file keeps
the new mode, splitting the replica. Log the failure as a warning
instead, matching the best-effort nature of FsyncDir (already skipped
on Windows).

* admin: distinguish post-rename durability failures and use unique temp files

SaveVolumeInfo now uses os.CreateTemp for the staging file, preventing
concurrent saves for the same volume from colliding on a shared .tmp
path.

A directory fsync failure after os.Rename returns a
NotCrashDurableError instead of being silently swallowed. The rename
already committed the new metadata to disk, so PersistReadOnly,
MarkVolumeReadonly, and MarkVolumeWritable skip the in-memory rollback
for this error type (keeping state aligned with the durable file) while
still propagating the failure to the API. Pre-commit failures continue
to roll back as before.

* admin: continue post-commit work after NotCrashDurableError

MarkVolumeWritable now clears the EIO quarantine and the gRPC handlers
(makeVolumeReadonly step 3, makeVolumeWritable master notification)
proceed with their post-commit work when SaveVolumeInfo returns a
NotCrashDurableError, instead of aborting and leaving the volume
unavailable or the master unaware of the mode change. The durability
warning is still propagated to the API caller. Pre-commit failures
continue to abort early as before.

* admin: handle NotCrashDurableError in tier and EC callers

VolumeTierMoveDatFromRemote and VolumeEcShardsGenerate now check for
NotCrashDurableError from SaveVolumeInfo. When the rename has already
committed the new .vif, they continue with their post-commit work
(backend switch, remote deletion, keeping generated EC shards) instead
of aborting and leaving the on-disk metadata inconsistent with the
file layout. The durability warning is logged for the operator.
2026-09-07 18:40:37 -07:00

508 lines
17 KiB
Go

package handlers
import (
"context"
"encoding/json"
"fmt"
"io"
"math"
"mime"
"net/http"
"strconv"
"time"
"github.com/gorilla/mux"
"github.com/seaweedfs/seaweedfs/weed/admin/dash"
"github.com/seaweedfs/seaweedfs/weed/admin/view/app"
"github.com/seaweedfs/seaweedfs/weed/admin/view/layout"
"github.com/seaweedfs/seaweedfs/weed/glog"
)
// ClusterHandlers contains all the HTTP handlers for cluster management
type ClusterHandlers struct {
adminServer *dash.AdminServer
}
// NewClusterHandlers creates a new instance of ClusterHandlers
func NewClusterHandlers(adminServer *dash.AdminServer) *ClusterHandlers {
return &ClusterHandlers{
adminServer: adminServer,
}
}
// ShowClusterVolumeServers renders the cluster volume servers page
func (h *ClusterHandlers) ShowClusterVolumeServers(w http.ResponseWriter, r *http.Request) {
// Get cluster volume servers data
volumeServersData, err := h.adminServer.GetClusterVolumeServers()
if err != nil {
writeJSONError(w, http.StatusInternalServerError, "Failed to get cluster volume servers: "+err.Error())
return
}
username := usernameOrDefault(r)
volumeServersData.Username = username
// Render HTML template
w.Header().Set("Content-Type", "text/html")
volumeServersComponent := app.ClusterVolumeServers(*volumeServersData)
viewCtx := layout.NewViewContext(r, username, dash.CSRFTokenFromContext(r.Context()))
layoutComponent := layout.Layout(viewCtx, volumeServersComponent)
if err := layoutComponent.Render(r.Context(), w); err != nil {
writeJSONError(w, http.StatusInternalServerError, "Failed to render template: "+err.Error())
return
}
}
// ShowMountClients renders the connected FUSE/VFS mount clients page
func (h *ClusterHandlers) ShowMountClients(w http.ResponseWriter, r *http.Request) {
data, err := h.adminServer.GetMountClients()
if err != nil {
writeJSONError(w, http.StatusInternalServerError, "Failed to get mount clients: "+err.Error())
return
}
username := usernameOrDefault(r)
data.Username = username
// Render HTML template
w.Header().Set("Content-Type", "text/html")
mountClientsComponent := app.MountClients(*data)
viewCtx := layout.NewViewContext(r, username, dash.CSRFTokenFromContext(r.Context()))
layoutComponent := layout.Layout(viewCtx, mountClientsComponent)
if err := layoutComponent.Render(r.Context(), w); err != nil {
writeJSONError(w, http.StatusInternalServerError, "Failed to render template: "+err.Error())
return
}
}
// ShowClusterVolumes renders the cluster volumes page
func (h *ClusterHandlers) ShowClusterVolumes(w http.ResponseWriter, r *http.Request) {
// Get pagination and sorting parameters from query string
page := 1
if p := r.URL.Query().Get("page"); p != "" {
if parsed, err := strconv.Atoi(p); err == nil && parsed > 0 {
page = parsed
}
}
pageSize := 100
if ps := r.URL.Query().Get("pageSize"); ps != "" {
if parsed, err := strconv.Atoi(ps); err == nil && parsed > 0 && parsed <= 1000 {
pageSize = parsed
}
}
sortBy := defaultQuery(r.URL.Query().Get("sortBy"), "id")
sortOrder := defaultQuery(r.URL.Query().Get("sortOrder"), "asc")
collection := r.URL.Query().Get("collection") // Optional collection filter
// Get cluster volumes data
volumesData, err := h.adminServer.GetClusterVolumes(page, pageSize, sortBy, sortOrder, collection)
if err != nil {
writeJSONError(w, http.StatusInternalServerError, "Failed to get cluster volumes: "+err.Error())
return
}
username := usernameOrDefault(r)
volumesData.Username = username
// Render HTML template
w.Header().Set("Content-Type", "text/html")
volumesComponent := app.ClusterVolumes(*volumesData)
viewCtx := layout.NewViewContext(r, username, dash.CSRFTokenFromContext(r.Context()))
layoutComponent := layout.Layout(viewCtx, volumesComponent)
if err := layoutComponent.Render(r.Context(), w); err != nil {
writeJSONError(w, http.StatusInternalServerError, "Failed to render template: "+err.Error())
return
}
}
// ExportClusterVolumes streams the full-cluster volume list as a downloadable
// JSON report: every volume and EC shard across the topology, with more fields
// than the paginated table (a superset of the volume.list shell command).
func (h *ClusterHandlers) ExportClusterVolumes(w http.ResponseWriter, r *http.Request) {
collection := r.URL.Query().Get("collection") // Optional collection filter
export, err := h.adminServer.ExportClusterVolumeList(r.Context(), collection, time.Now().UTC())
if err != nil {
writeJSONError(w, http.StatusInternalServerError, "Failed to export volume list: "+err.Error())
return
}
filename := fmt.Sprintf("seaweedfs-volumes-%s.json", export.GeneratedAt.Format("20060102-150405"))
w.Header().Set("Content-Type", "application/json")
w.Header().Set("Content-Disposition", mime.FormatMediaType("attachment", map[string]string{"filename": filename}))
enc := json.NewEncoder(w)
enc.SetIndent("", " ")
if err := enc.Encode(export); err != nil {
// The response is already streaming, so we can only log a late failure.
glog.Errorf("export volume list: encode failed: %v", err)
}
}
// ShowVolumeDetails renders the volume details page
func (h *ClusterHandlers) ShowVolumeDetails(w http.ResponseWriter, r *http.Request) {
vars := mux.Vars(r)
volumeIDStr := vars["id"]
server := vars["server"]
if volumeIDStr == "" {
writeJSONError(w, http.StatusBadRequest, "Volume ID is required")
return
}
if server == "" {
writeJSONError(w, http.StatusBadRequest, "Server is required")
return
}
volumeID, err := strconv.ParseUint(volumeIDStr, 10, 32)
if err != nil {
writeJSONError(w, http.StatusBadRequest, "Invalid volume ID")
return
}
// Get volume details
volumeDetails, err := h.adminServer.GetVolumeDetails(uint32(volumeID), server)
if err != nil {
writeJSONError(w, http.StatusInternalServerError, "Failed to get volume details: "+err.Error())
return
}
username := usernameOrDefault(r)
// Render HTML template
w.Header().Set("Content-Type", "text/html")
volumeDetailsComponent := app.VolumeDetails(*volumeDetails)
viewCtx := layout.NewViewContext(r, username, dash.CSRFTokenFromContext(r.Context()))
layoutComponent := layout.Layout(viewCtx, volumeDetailsComponent)
if err := layoutComponent.Render(r.Context(), w); err != nil {
writeJSONError(w, http.StatusInternalServerError, "Failed to render template: "+err.Error())
return
}
}
// ShowClusterCollections renders the cluster collections page
func (h *ClusterHandlers) ShowClusterCollections(w http.ResponseWriter, r *http.Request) {
// Get cluster collections data
collectionsData, err := h.adminServer.GetClusterCollections()
if err != nil {
writeJSONError(w, http.StatusInternalServerError, "Failed to get cluster collections: "+err.Error())
return
}
username := usernameOrDefault(r)
collectionsData.Username = username
// Render HTML template
w.Header().Set("Content-Type", "text/html")
collectionsComponent := app.ClusterCollections(*collectionsData)
viewCtx := layout.NewViewContext(r, username, dash.CSRFTokenFromContext(r.Context()))
layoutComponent := layout.Layout(viewCtx, collectionsComponent)
if err := layoutComponent.Render(r.Context(), w); err != nil {
writeJSONError(w, http.StatusInternalServerError, "Failed to render template: "+err.Error())
return
}
}
// ShowCollectionDetails renders the collection detail page
func (h *ClusterHandlers) ShowCollectionDetails(w http.ResponseWriter, r *http.Request) {
collectionName := mux.Vars(r)["name"]
if collectionName == "" {
writeJSONError(w, http.StatusBadRequest, "Collection name is required")
return
}
// Parse query parameters
query := r.URL.Query()
page, _ := strconv.Atoi(defaultQuery(query.Get("page"), "1"))
pageSize, _ := strconv.Atoi(defaultQuery(query.Get("page_size"), "25"))
sortBy := defaultQuery(query.Get("sort_by"), "volume_id")
sortOrder := defaultQuery(query.Get("sort_order"), "asc")
// Get collection details data (volumes and EC volumes)
collectionDetailsData, err := h.adminServer.GetCollectionDetails(collectionName, page, pageSize, sortBy, sortOrder)
if err != nil {
writeJSONError(w, http.StatusInternalServerError, "Failed to get collection details: "+err.Error())
return
}
username := usernameOrDefault(r)
collectionDetailsData.Username = username
// Render HTML template
w.Header().Set("Content-Type", "text/html")
collectionDetailsComponent := app.CollectionDetails(*collectionDetailsData)
viewCtx := layout.NewViewContext(r, username, dash.CSRFTokenFromContext(r.Context()))
layoutComponent := layout.Layout(viewCtx, collectionDetailsComponent)
if err := layoutComponent.Render(r.Context(), w); err != nil {
writeJSONError(w, http.StatusInternalServerError, "Failed to render template: "+err.Error())
return
}
}
// ShowClusterEcShards handles the cluster EC shards page (individual shards view)
func (h *ClusterHandlers) ShowClusterEcShards(w http.ResponseWriter, r *http.Request) {
// Parse query parameters
query := r.URL.Query()
page, _ := strconv.Atoi(defaultQuery(query.Get("page"), "1"))
pageSize, _ := strconv.Atoi(defaultQuery(query.Get("page_size"), "100"))
sortBy := defaultQuery(query.Get("sort_by"), "volume_id")
sortOrder := defaultQuery(query.Get("sort_order"), "asc")
collection := defaultQuery(query.Get("collection"), "")
// Get data from admin server
data, err := h.adminServer.GetClusterEcVolumes(page, pageSize, sortBy, sortOrder, collection)
if err != nil {
writeJSONError(w, http.StatusInternalServerError, err.Error())
return
}
username := usernameOrDefault(r)
data.Username = username
// Render template
w.Header().Set("Content-Type", "text/html")
ecVolumesComponent := app.ClusterEcVolumes(*data)
viewCtx := layout.NewViewContext(r, username, dash.CSRFTokenFromContext(r.Context()))
layoutComponent := layout.Layout(viewCtx, ecVolumesComponent)
if err := layoutComponent.Render(r.Context(), w); err != nil {
writeJSONError(w, http.StatusInternalServerError, err.Error())
return
}
}
// ShowEcVolumeDetails renders the EC volume details page
func (h *ClusterHandlers) ShowEcVolumeDetails(w http.ResponseWriter, r *http.Request) {
volumeIDStr := mux.Vars(r)["id"]
if volumeIDStr == "" {
writeJSONError(w, http.StatusBadRequest, "Volume ID is required")
return
}
volumeID, err := strconv.Atoi(volumeIDStr)
if err != nil {
writeJSONError(w, http.StatusBadRequest, "Invalid volume ID")
return
}
// Check that volumeID is within uint32 range
if volumeID < 0 || uint64(volumeID) > math.MaxUint32 {
writeJSONError(w, http.StatusBadRequest, "Volume ID out of range")
return
}
// Parse sorting parameters
query := r.URL.Query()
sortBy := defaultQuery(query.Get("sort_by"), "shard_id")
sortOrder := defaultQuery(query.Get("sort_order"), "asc")
// Get EC volume details
ecVolumeDetails, err := h.adminServer.GetEcVolumeDetails(uint32(volumeID), sortBy, sortOrder)
if err != nil {
writeJSONError(w, http.StatusInternalServerError, "Failed to get EC volume details: "+err.Error())
return
}
username := usernameOrDefault(r)
ecVolumeDetails.Username = username
// Render HTML template
w.Header().Set("Content-Type", "text/html")
ecVolumeDetailsComponent := app.EcVolumeDetails(*ecVolumeDetails)
viewCtx := layout.NewViewContext(r, username, dash.CSRFTokenFromContext(r.Context()))
layoutComponent := layout.Layout(viewCtx, ecVolumeDetailsComponent)
if err := layoutComponent.Render(r.Context(), w); err != nil {
writeJSONError(w, http.StatusInternalServerError, "Failed to render template: "+err.Error())
return
}
}
// ShowClusterMasters renders the cluster masters page
func (h *ClusterHandlers) ShowClusterMasters(w http.ResponseWriter, r *http.Request) {
// Get cluster masters data
mastersData, err := h.adminServer.GetClusterMasters()
if err != nil {
writeJSONError(w, http.StatusInternalServerError, "Failed to get cluster masters: "+err.Error())
return
}
username := usernameOrDefault(r)
mastersData.Username = username
// Render HTML template
w.Header().Set("Content-Type", "text/html")
mastersComponent := app.ClusterMasters(*mastersData)
viewCtx := layout.NewViewContext(r, username, dash.CSRFTokenFromContext(r.Context()))
layoutComponent := layout.Layout(viewCtx, mastersComponent)
if err := layoutComponent.Render(r.Context(), w); err != nil {
writeJSONError(w, http.StatusInternalServerError, "Failed to render template: "+err.Error())
return
}
}
// ShowClusterFilers renders the cluster filers page
func (h *ClusterHandlers) ShowClusterFilers(w http.ResponseWriter, r *http.Request) {
// Get cluster filers data
filersData, err := h.adminServer.GetClusterFilers()
if err != nil {
writeJSONError(w, http.StatusInternalServerError, "Failed to get cluster filers: "+err.Error())
return
}
username := usernameOrDefault(r)
filersData.Username = username
// Render HTML template
w.Header().Set("Content-Type", "text/html")
filersComponent := app.ClusterFilers(*filersData)
viewCtx := layout.NewViewContext(r, username, dash.CSRFTokenFromContext(r.Context()))
layoutComponent := layout.Layout(viewCtx, filersComponent)
if err := layoutComponent.Render(r.Context(), w); err != nil {
writeJSONError(w, http.StatusInternalServerError, "Failed to render template: "+err.Error())
return
}
}
// ShowClusterS3Servers renders the cluster S3 servers page
func (h *ClusterHandlers) ShowClusterS3Servers(w http.ResponseWriter, r *http.Request) {
// Get cluster S3 servers data
s3ServersData, err := h.adminServer.GetClusterS3Servers()
if err != nil {
writeJSONError(w, http.StatusInternalServerError, "Failed to get cluster S3 servers: "+err.Error())
return
}
username := usernameOrDefault(r)
s3ServersData.Username = username
// Render HTML template
w.Header().Set("Content-Type", "text/html")
s3ServersComponent := app.ClusterS3Servers(*s3ServersData)
viewCtx := layout.NewViewContext(r, username, dash.CSRFTokenFromContext(r.Context()))
layoutComponent := layout.Layout(viewCtx, s3ServersComponent)
if err := layoutComponent.Render(r.Context(), w); err != nil {
writeJSONError(w, http.StatusInternalServerError, "Failed to render template: "+err.Error())
return
}
}
// GetClusterTopology returns the cluster topology as JSON
func (h *ClusterHandlers) GetClusterTopology(w http.ResponseWriter, r *http.Request) {
topology, err := h.adminServer.GetClusterTopology()
if err != nil {
writeJSONError(w, http.StatusInternalServerError, err.Error())
return
}
writeJSON(w, http.StatusOK, topology)
}
// GetMasters returns master node information
func (h *ClusterHandlers) GetMasters(w http.ResponseWriter, r *http.Request) {
mastersData, err := h.adminServer.GetClusterMasters()
if err != nil {
writeJSONError(w, http.StatusInternalServerError, "Failed to get cluster masters: "+err.Error())
return
}
writeJSON(w, http.StatusOK, mastersData)
}
// GetVolumeServers returns volume server information
func (h *ClusterHandlers) GetVolumeServers(w http.ResponseWriter, r *http.Request) {
topology, err := h.adminServer.GetClusterTopology()
if err != nil {
writeJSONError(w, http.StatusInternalServerError, err.Error())
return
}
writeJSON(w, http.StatusOK, map[string]interface{}{"volume_servers": topology.VolumeServers})
}
// SetVolumeReadOnly handles access mode changes for a single volume replica.
func (h *ClusterHandlers) SetVolumeReadOnly(w http.ResponseWriter, r *http.Request) {
vars := mux.Vars(r)
volumeID, err := strconv.ParseUint(vars["id"], 10, 32)
if err != nil {
writeJSONError(w, http.StatusBadRequest, "Invalid volume ID")
return
}
server := vars["server"]
if server == "" {
writeJSONError(w, http.StatusBadRequest, "Server is required")
return
}
var request struct {
ReadOnly *bool `json:"read_only"`
}
decoder := json.NewDecoder(r.Body)
if err := decoder.Decode(&request); err != nil || request.ReadOnly == nil {
writeJSONError(w, http.StatusBadRequest, "read_only must be a boolean")
return
}
if err := decoder.Decode(&struct{}{}); err != io.EOF {
writeJSONError(w, http.StatusBadRequest, "read_only must be a boolean")
return
}
ctx, cancel := context.WithTimeout(r.Context(), 30*time.Second)
defer cancel()
if err := h.adminServer.SetVolumeReadOnly(ctx, uint32(volumeID), server, *request.ReadOnly); err != nil {
writeJSONError(w, http.StatusInternalServerError, "Failed to change volume access mode: "+err.Error())
return
}
mode := "read/write"
if *request.ReadOnly {
mode = "read-only"
}
writeJSON(w, http.StatusOK, map[string]interface{}{
"message": fmt.Sprintf("Volume %d on %s marked %s", volumeID, server, mode),
"volume_id": volumeID,
"server": server,
"read_only": *request.ReadOnly,
})
}
// VacuumVolume handles volume vacuum requests via API
func (h *ClusterHandlers) VacuumVolume(w http.ResponseWriter, r *http.Request) {
vars := mux.Vars(r)
volumeIDStr := vars["id"]
server := vars["server"]
if volumeIDStr == "" {
writeJSONError(w, http.StatusBadRequest, "Volume ID is required")
return
}
volumeID, err := strconv.Atoi(volumeIDStr)
if err != nil {
writeJSONError(w, http.StatusBadRequest, "Invalid volume ID")
return
}
if server == "" {
writeJSONError(w, http.StatusBadRequest, "Server is required")
return
}
// Perform vacuum operation
err = h.adminServer.VacuumVolume(volumeID, server)
if err != nil {
writeJSONError(w, http.StatusInternalServerError, "Failed to vacuum volume: "+err.Error())
return
}
writeJSON(w, http.StatusOK, map[string]interface{}{
"message": "Volume vacuum started successfully",
"volume_id": volumeID,
"server": server,
})
}
func usernameOrDefault(r *http.Request) string {
username := dash.UsernameFromContext(r.Context())
if username == "" {
return "admin"
}
return username
}