gsh-digital-services/seaweedFS
5
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Rework parameters passing for functions within volume.check.disk. (#7448)

Browse Source 
* Rework parameters passing for functions within `volume.check.disk`.

We'll need to rework this logic to account for read-only volumes, and there're already way too many parameters shuffled around.

Grouping these into a single struct simplifies the overall codebase.

* similar fix

* Improved Error Handling in Tests

* propagate the errors

* edge cases

* edge case on modified time

* clean up

---------

Co-authored-by: chrislu <chris.lu@gmail.com>
This commit is contained in:
Lisandro Pin
2025-11-11 01:03:38 +01:00
committed by GitHub
parent d8dd3e4c53
commit 9744382a18
3 changed files with 437 additions and 147 deletions
Download Patch File Download Diff File Expand all files Collapse all files

23
weed/shell/command_volume_fix_replication.go
View File

@@ -16,7 +16,6 @@ import (
"github.com/seaweedfs/seaweedfs/weed/storage/needle_map"
"github.com/seaweedfs/seaweedfs/weed/storage/types"
"github.com/seaweedfs/seaweedfs/weed/util"
"google.golang.org/grpc"
"github.com/seaweedfs/seaweedfs/weed/operation"
"github.com/seaweedfs/seaweedfs/weed/pb/master_pb"
@@ -204,22 +203,32 @@ func collectVolumeReplicaLocations(topologyInfo *master_pb.TopologyInfo) (map[ui
type SelectOneVolumeFunc func(replicas []*VolumeReplica, replicaPlacement *super_block.ReplicaPlacement) *VolumeReplica
func checkOneVolume(a *VolumeReplica, b *VolumeReplica, writer io.Writer, grpcDialOption grpc.DialOption) (err error) {
func checkOneVolume(a *VolumeReplica, b *VolumeReplica, writer io.Writer, commandEnv *CommandEnv) (err error) {
aDB, bDB := needle_map.NewMemDb(), needle_map.NewMemDb()
defer func() {
aDB.Close()
bDB.Close()
}()
vcd := &volumeCheckDisk{
writer: writer,
commandEnv: commandEnv,
now: time.Now(),
verbose: false,
applyChanges: true,
syncDeletions: false,
nonRepairThreshold: float64(1),
}
// read index db
readIndexDbCutoffFrom := uint64(time.Now().UnixNano())
if err = readIndexDatabase(aDB, a.info.Collection, a.info.Id, pb.NewServerAddressFromDataNode(a.location.dataNode), false, writer, grpcDialOption); err != nil {
if err = vcd.readIndexDatabase(aDB, a.info.Collection, a.info.Id, pb.NewServerAddressFromDataNode(a.location.dataNode)); err != nil {
return fmt.Errorf("readIndexDatabase %s volume %d: %v", a.location.dataNode, a.info.Id, err)
}
if err := readIndexDatabase(bDB, b.info.Collection, b.info.Id, pb.NewServerAddressFromDataNode(b.location.dataNode), false, writer, grpcDialOption); err != nil {
if err := vcd.readIndexDatabase(bDB, b.info.Collection, b.info.Id, pb.NewServerAddressFromDataNode(b.location.dataNode)); err != nil {
return fmt.Errorf("readIndexDatabase %s volume %d: %v", b.location.dataNode, b.info.Id, err)
}
if _, err = doVolumeCheckDisk(aDB, bDB, a, b, false, writer, true, false, float64(1), readIndexDbCutoffFrom, grpcDialOption); err != nil {
if _, err = vcd.doVolumeCheckDisk(aDB, bDB, a, b); err != nil {
return fmt.Errorf("doVolumeCheckDisk source:%s target:%s volume %d: %v", a.location.dataNode.Id, b.location.dataNode.Id, a.info.Id, err)
}
return
@@ -271,7 +280,7 @@ func (c *commandVolumeFixReplication) deleteOneVolume(commandEnv *CommandEnv, wr
if replicaB.location.dataNode == replica.location.dataNode {
continue
}
if checkErr = checkOneVolume(replica, replicaB, writer, commandEnv.option.GrpcDialOption); checkErr != nil {
if checkErr = checkOneVolume(replica, replicaB, writer, commandEnv); checkErr != nil {
fmt.Fprintf(writer, "sync volume %d on %s and %s: %v\n", replica.info.Id, replica.location.dataNode.Id, replicaB.location.dataNode.Id, checkErr)
break
}