Files
seaweedfs/weed/admin/dash/collection_management.go
T
Chris LuandGitHub 52d74df4d1 clients: stream the volume listings that ask for everything (#10679)
* master: stream volume listings

A listing of 800k volumes is 36MB on the wire but 305MB as messages, and the
master built all of it, then held it while grpc encoded it. Two of those at
once is most of a small master's heap, and the maintenance scanner asks every
30 minutes.

The topology goes out first, listing nothing, then its volumes in batches, so
the master holds a batch rather than a cluster: 341MB of live heap for one
listing becomes 4.4MB. It allocates much the same either way -- what changes is
how much of it has to be live at once, which is what sets the heap ceiling.

Batches are built under their disk's lock and sent outside it, so a slow reader
stalls the stream rather than the topology. They therefore do not share one
instant, which a single listing did not either: it takes each disk's lock in
turn, so a volume moving during either can be seen twice or not at all.

The client helper hides which kind of master answered: one too old for the
stream is asked the old way and its reply cut into the same batches. Either way
the topology handed over lists no volumes, so a caller cannot come to depend on
finding them there.

* admin: stream the listing the maintenance scan reads

It asks for every volume in the cluster every 30 minutes. Reassembling it
client-side keeps the scan identical -- ActiveTopology splits disks by the
disk ids on the volumes, so it needs them in the topology -- while the master
no longer builds the whole reply to send it.

* topology: report a disk id that does not depend on map order

A topology disk that fronts several physical disks took its reported id from
whichever volume the map yielded first, so two listings of an unchanged disk
could disagree. Take the smallest instead.

* topology: test that a streamed listing rebuilds to the whole one

The callers that stream now rebuild the listing from a topology sent without
volumes plus the batches after it, so that has to come out the same as being
sent it whole, at every batch size and under a filter.

* clients: stream the volume listings that ask for everything

The dashboard's list and export pages, the collection and ec shard pages, the
topology view, the worker metrics and two shell commands each asked the master
to build all 800k volumes into one reply. They read the same listing as before,
rebuilt on their side, so the master no longer holds it.

The three that already ask for one volume or one collection stay as they are:
their replies are small, and streaming one costs a round trip to say so.
2026-08-10 09:51:08 -07:00

396 lines
11 KiB
Go

package dash
import (
"context"
"sort"
"time"
"github.com/seaweedfs/seaweedfs/weed/pb"
"github.com/seaweedfs/seaweedfs/weed/pb/master_pb"
)
// GetClusterCollections retrieves cluster collections data
func (s *AdminServer) GetClusterCollections() (*ClusterCollectionsData, error) {
var collections []CollectionInfo
var totalVolumes int
var totalEcVolumes int
var totalChunks int64
var totalSize int64
collectionMap := make(map[string]*CollectionInfo)
// Get actual collection information from volume data
err := s.WithMasterClient(func(client master_pb.SeaweedClient) error {
resp, err := pb.CollectVolumeList(context.Background(), client, &master_pb.VolumeListRequest{})
if err != nil {
return err
}
if resp.TopologyInfo != nil {
for _, dc := range resp.TopologyInfo.DataCenterInfos {
for _, rack := range dc.RackInfos {
for _, node := range rack.DataNodeInfos {
for _, diskInfo := range node.DiskInfos {
// Process regular volumes
for _, volInfo := range diskInfo.VolumeInfos {
// Extract collection name from volume info
collectionName := volInfo.Collection
if collectionName == "" {
collectionName = "default" // Default collection for volumes without explicit collection
}
// Get disk type from volume info, default to hdd if empty
diskType := volInfo.DiskType
if diskType == "" {
diskType = "hdd"
}
// Get or create collection info
if collection, exists := collectionMap[collectionName]; exists {
collection.VolumeCount++
collection.TotalSize += int64(volInfo.Size)
// Update data center if this collection spans multiple DCs
if collection.DataCenter != dc.Id && collection.DataCenter != "multi" {
collection.DataCenter = "multi"
}
// Add disk type if not already present
diskTypeExists := false
for _, existingDiskType := range collection.DiskTypes {
if existingDiskType == diskType {
diskTypeExists = true
break
}
}
if !diskTypeExists {
collection.DiskTypes = append(collection.DiskTypes, diskType)
}
totalVolumes++
totalSize += int64(volInfo.Size)
} else {
newCollection := CollectionInfo{
Name: collectionName,
DataCenter: dc.Id,
VolumeCount: 1,
EcVolumeCount: 0,
TotalSize: int64(volInfo.Size),
DiskTypes: []string{diskType},
}
collectionMap[collectionName] = &newCollection
totalVolumes++
totalSize += int64(volInfo.Size)
}
}
// Process EC volumes
ecVolumeMap := make(map[uint32]bool) // Track unique EC volumes to avoid double counting
for _, ecShardInfo := range diskInfo.EcShardInfos {
// Extract collection name from EC shard info
collectionName := ecShardInfo.Collection
if collectionName == "" {
collectionName = "default" // Default collection for EC volumes without explicit collection
}
// Only count each EC volume once (not per shard)
if !ecVolumeMap[ecShardInfo.Id] {
ecVolumeMap[ecShardInfo.Id] = true
// Get disk type from disk info, default to hdd if empty
diskType := diskInfo.Type
if diskType == "" {
diskType = "hdd"
}
// Get or create collection info
if collection, exists := collectionMap[collectionName]; exists {
collection.EcVolumeCount++
// Update data center if this collection spans multiple DCs
if collection.DataCenter != dc.Id && collection.DataCenter != "multi" {
collection.DataCenter = "multi"
}
// Add disk type if not already present
diskTypeExists := false
for _, existingDiskType := range collection.DiskTypes {
if existingDiskType == diskType {
diskTypeExists = true
break
}
}
if !diskTypeExists {
collection.DiskTypes = append(collection.DiskTypes, diskType)
}
totalEcVolumes++
} else {
newCollection := CollectionInfo{
Name: collectionName,
DataCenter: dc.Id,
VolumeCount: 0,
EcVolumeCount: 1,
TotalSize: 0,
DiskTypes: []string{diskType},
}
collectionMap[collectionName] = &newCollection
totalEcVolumes++
}
}
}
}
}
}
}
// Chunk counts come from the shared collection aggregation, which
// nets out tombstones and counts a chunk once no matter how many
// volume replicas or EC shard holders report it.
for collectionName, stats := range collectCollectionStats(resp.TopologyInfo) {
if collection, exists := collectionMap[collectionName]; exists {
collection.ChunkCount = stats.FileCount
totalChunks += stats.FileCount
}
}
}
return nil
})
if err != nil {
return nil, err
}
// Convert map to slice
for _, collection := range collectionMap {
collections = append(collections, *collection)
}
// Sort collections alphabetically by name
sort.Slice(collections, func(i, j int) bool {
return collections[i].Name < collections[j].Name
})
// If no collections found, show a message indicating no collections exist
if len(collections) == 0 {
// Return empty collections data instead of creating fake ones
return &ClusterCollectionsData{
Collections: []CollectionInfo{},
TotalCollections: 0,
TotalVolumes: 0,
TotalEcVolumes: 0,
TotalChunks: 0,
TotalSize: 0,
LastUpdated: time.Now(),
}, nil
}
return &ClusterCollectionsData{
Collections: collections,
TotalCollections: len(collections),
TotalVolumes: totalVolumes,
TotalEcVolumes: totalEcVolumes,
TotalChunks: totalChunks,
TotalSize: totalSize,
LastUpdated: time.Now(),
}, nil
}
// GetCollectionDetails retrieves detailed information for a specific collection including volumes and EC volumes
func (s *AdminServer) GetCollectionDetails(collectionName string, page int, pageSize int, sortBy string, sortOrder string) (*CollectionDetailsData, error) {
// Set defaults
if page < 1 {
page = 1
}
if pageSize < 1 || pageSize > 1000 {
pageSize = 25
}
if sortBy == "" {
sortBy = "volume_id"
}
if sortOrder == "" {
sortOrder = "asc"
}
var regularVolumes []VolumeWithTopology
var ecVolumes []EcVolumeWithShards
var totalSize int64
dataCenters := make(map[string]bool)
diskTypes := make(map[string]bool)
// Get regular volumes for this collection
regularVolumeData, err := s.GetClusterVolumes(1, 10000, "volume_id", "asc", collectionName) // Get all volumes
if err != nil {
return nil, err
}
regularVolumes = regularVolumeData.Volumes
totalSize = regularVolumeData.TotalSize
// Chunk counts come from the shared collection aggregation, which nets out
// tombstones and counts a chunk once no matter how many volume replicas or
// EC shard holders report it. Fail rather than render a zero that reads
// like an empty collection.
stats, err := s.getCollectionStats()
if err != nil {
return nil, err
}
totalChunks := stats[collectionName].FileCount
// Collect data centers and disk types from regular volumes
for _, vol := range regularVolumes {
dataCenters[vol.DataCenter] = true
diskTypes[vol.DiskType] = true
}
// Get EC volumes for this collection
ecVolumeData, err := s.GetClusterEcVolumes(1, 10000, "volume_id", "asc", collectionName) // Get all EC volumes
if err != nil {
return nil, err
}
ecVolumes = ecVolumeData.EcVolumes
// Collect data centers from EC volumes
for _, ecVol := range ecVolumes {
for _, dc := range ecVol.DataCenters {
dataCenters[dc] = true
}
}
// Combine all volumes for sorting and pagination
type VolumeForSorting struct {
Type string // "regular" or "ec"
RegularVolume *VolumeWithTopology
EcVolume *EcVolumeWithShards
}
var allVolumes []VolumeForSorting
for i := range regularVolumes {
allVolumes = append(allVolumes, VolumeForSorting{
Type: "regular",
RegularVolume: &regularVolumes[i],
})
}
for i := range ecVolumes {
allVolumes = append(allVolumes, VolumeForSorting{
Type: "ec",
EcVolume: &ecVolumes[i],
})
}
// Sort all volumes
sort.Slice(allVolumes, func(i, j int) bool {
var less bool
switch sortBy {
case "volume_id":
var idI, idJ uint32
if allVolumes[i].Type == "regular" {
idI = allVolumes[i].RegularVolume.Id
} else {
idI = allVolumes[i].EcVolume.VolumeID
}
if allVolumes[j].Type == "regular" {
idJ = allVolumes[j].RegularVolume.Id
} else {
idJ = allVolumes[j].EcVolume.VolumeID
}
less = idI < idJ
case "type":
// Sort by type first (regular before ec), then by volume ID
if allVolumes[i].Type == allVolumes[j].Type {
var idI, idJ uint32
if allVolumes[i].Type == "regular" {
idI = allVolumes[i].RegularVolume.Id
} else {
idI = allVolumes[i].EcVolume.VolumeID
}
if allVolumes[j].Type == "regular" {
idJ = allVolumes[j].RegularVolume.Id
} else {
idJ = allVolumes[j].EcVolume.VolumeID
}
less = idI < idJ
} else {
less = allVolumes[i].Type < allVolumes[j].Type // "ec" < "regular"
}
default:
// Default to volume ID sort
var idI, idJ uint32
if allVolumes[i].Type == "regular" {
idI = allVolumes[i].RegularVolume.Id
} else {
idI = allVolumes[i].EcVolume.VolumeID
}
if allVolumes[j].Type == "regular" {
idJ = allVolumes[j].RegularVolume.Id
} else {
idJ = allVolumes[j].EcVolume.VolumeID
}
less = idI < idJ
}
if sortOrder == "desc" {
return !less
}
return less
})
// Apply pagination
totalVolumesAndEc := len(allVolumes)
totalPages := (totalVolumesAndEc + pageSize - 1) / pageSize
startIndex := (page - 1) * pageSize
endIndex := startIndex + pageSize
if endIndex > totalVolumesAndEc {
endIndex = totalVolumesAndEc
}
if startIndex >= totalVolumesAndEc {
startIndex = 0
endIndex = 0
}
// Extract paginated results
var paginatedRegularVolumes []VolumeWithTopology
var paginatedEcVolumes []EcVolumeWithShards
for i := startIndex; i < endIndex; i++ {
if allVolumes[i].Type == "regular" {
paginatedRegularVolumes = append(paginatedRegularVolumes, *allVolumes[i].RegularVolume)
} else {
paginatedEcVolumes = append(paginatedEcVolumes, *allVolumes[i].EcVolume)
}
}
// Convert maps to slices
var dcList []string
for dc := range dataCenters {
dcList = append(dcList, dc)
}
sort.Strings(dcList)
var diskTypeList []string
for diskType := range diskTypes {
diskTypeList = append(diskTypeList, diskType)
}
sort.Strings(diskTypeList)
return &CollectionDetailsData{
CollectionName: collectionName,
RegularVolumes: paginatedRegularVolumes,
EcVolumes: paginatedEcVolumes,
TotalVolumes: len(regularVolumes),
TotalEcVolumes: len(ecVolumes),
TotalChunks: totalChunks,
TotalSize: totalSize,
DataCenters: dcList,
DiskTypes: diskTypeList,
LastUpdated: time.Now(),
Page: page,
PageSize: pageSize,
TotalPages: totalPages,
SortBy: sortBy,
SortOrder: sortOrder,
}, nil
}