Files
seaweedfs/weed/pb/master_pb/master_helper.go
T
Chris Lu 41b6ad002b fix(volume.list): show one entry per physical disk on multi-disk nodes (#9541)
* fix(volume.list): show one entry per physical disk on multi-disk nodes

DataNodeInfo.DiskInfos is keyed by disk type, so several same-type
physical disks on one node collapse to a single map entry at the master.
volume.list iterated that map directly and reported one "Disk hdd ...
id:0" line per node, hiding the per-disk volume and shard layout. EC
operators on multi-disk volume servers had no way to verify which
physical disk a shard landed on.

Lift the per-physical-disk split into a DiskInfo.SplitByPhysicalDisk()
method on the proto type so consumers outside admin/topology can use
it. Apply it in writeDataNodeInfo so the verbose Disk block shows one
entry per physical disk, ordered by DiskId. Capacity counters are
split evenly across reconstructed disks since the wire format doesn't
carry per-disk capacity yet.

This is a display-only change. ActiveTopology already did the split on
its own and is now updated to call the shared helper.

* fix(volume.list): preserve totals, count active/remote exactly, dedupe header

Address review feedback on the per-physical-disk split:

- share() truncated remainders so reconstructed per-disk counters could
  sum to less than the original aggregate (10 / 3 = 3+3+3). Distribute
  the remainder to the lowest disk ids so MaxVolumeCount and
  FreeVolumeCount sum exactly back to the node totals.
- ActiveVolumeCount and RemoteVolumeCount are derivable per disk from
  the VolumeInfos already grouped by DiskId, so count them exactly
  (ReadOnly=false and RemoteStorageName!="" respectively) instead of
  approximating with an even split.
- writeDataNodeInfo's per-disk callback fired the DataNode header on
  every iteration after the split, so a node with 6 physical disks
  emitted 6 DataNode headers. Guard the callback with headerPrinted so
  the header still appears at most once per node.
- Sort split disks deterministically using explicit DiskId comparison
  to avoid int overflow risk on 32-bit systems.
- Tighten the volume.list test substring to "id:N\n" so unrelated
  tokens like "ec volume id:101" don't accidentally match the id:1
  needle, and assert the rack callback fires once.
2026-05-18 14:43:44 -07:00

112 lines
3.4 KiB
Go

package master_pb
import "sort"
func (v *VolumeLocation) IsEmptyUrl() bool {
return v.Url == "" || v.Url == ":0"
}
// SplitByPhysicalDisk returns one DiskInfo per physical disk_id observed in
// VolumeInfos / EcShardInfos. The wire format keys DataNodeInfo.DiskInfos by
// disk type, so multiple same-type physical disks on one DataNode collapse
// into a single DiskInfo entry. Per-volume and per-shard records carry the
// real physical DiskId; this helper rebuilds a per-physical-disk view from
// those records so consumers (topology indexes, shell output) can target
// individual disks instead of treating each node as one big disk.
//
// ActiveVolumeCount and RemoteVolumeCount are computed exactly from each
// disk's VolumeInfos (read-only and remote-backed are known per-volume).
// MaxVolumeCount and FreeVolumeCount are not derivable from per-volume
// records, so they are split across reconstructed disks with the remainder
// distributed to the lowest disk ids — the sums are preserved exactly.
func (d *DiskInfo) SplitByPhysicalDisk() []*DiskInfo {
if d == nil {
return nil
}
normalize := func(id uint32) uint32 {
if id == 0 && d.DiskId != 0 {
return d.DiskId
}
return id
}
diskIDs := make(map[uint32]struct{})
for _, vi := range d.VolumeInfos {
diskIDs[normalize(vi.DiskId)] = struct{}{}
}
for _, eci := range d.EcShardInfos {
diskIDs[normalize(eci.DiskId)] = struct{}{}
}
if len(diskIDs) == 0 {
diskIDs[d.DiskId] = struct{}{}
}
if len(diskIDs) == 1 {
for diskID := range diskIDs {
if diskID == d.DiskId {
return []*DiskInfo{d}
}
}
}
perDiskVolumes := make(map[uint32][]*VolumeInformationMessage)
for _, vi := range d.VolumeInfos {
id := normalize(vi.DiskId)
perDiskVolumes[id] = append(perDiskVolumes[id], vi)
}
perDiskShards := make(map[uint32][]*VolumeEcShardInformationMessage)
for _, eci := range d.EcShardInfos {
id := normalize(eci.DiskId)
perDiskShards[id] = append(perDiskShards[id], eci)
}
// Sort disk IDs so the remainder distribution is deterministic and the
// reconstructed slice is in DiskId order, which is what downstream
// renderers expect.
ids := make([]uint32, 0, len(diskIDs))
for id := range diskIDs {
ids = append(ids, id)
}
sort.Slice(ids, func(i, j int) bool { return ids[i] < ids[j] })
count := int64(len(ids))
// share returns total / count, plus one extra for the first
// (total % count) entries so the sum of shares equals total. Without
// the remainder distribution, splitting 10 across 3 disks would yield
// 3+3+3 = 9 and under-report aggregate capacity.
share := func(total int64, idx int) int64 {
base := total / count
if int64(idx) < total%count {
return base + 1
}
return base
}
result := make([]*DiskInfo, 0, len(ids))
for i, diskID := range ids {
var activeCount, remoteCount int64
for _, vi := range perDiskVolumes[diskID] {
if !vi.ReadOnly {
activeCount++
}
if vi.RemoteStorageName != "" {
remoteCount++
}
}
result = append(result, &DiskInfo{
Type: d.Type,
MaxVolumeCount: share(d.MaxVolumeCount, i),
VolumeCount: int64(len(perDiskVolumes[diskID])),
FreeVolumeCount: share(d.FreeVolumeCount, i),
ActiveVolumeCount: activeCount,
RemoteVolumeCount: remoteCount,
VolumeInfos: perDiskVolumes[diskID],
EcShardInfos: perDiskShards[diskID],
DiskId: diskID,
Tags: append([]string(nil), d.Tags...),
})
}
return result
}