MyFavMirrors/minio
1
0
Fork 0
mirror of https://github.com/minio/minio.git synced 2025-11-20 01:50:24 -05:00
Code Issues Packages Projects Releases Wiki Activity

re-attach offline drive after new drive replacement (#10416)

Browse Source 
inconsistent drive healing when one of the drive is offline
while a new drive was replaced, this change is to ensure
that we can add the offline drive back into the mix by
healing it again.
This commit is contained in:
Harshavardhana
2020-09-04 17:09:02 -07:00
committed by GitHub
parent eb19c8af40
commit b0e1d4ce78
7 changed files with 166 additions and 146 deletions
Download Patch File Download Diff File Expand all files Collapse all files

43
cmd/erasure-zones.go
View File

@@ -2056,6 +2056,25 @@ func (z *erasureZones) Health(ctx context.Context, opts HealthOptions) HealthRes
writeQuorum++
}
var aggHealStateResult madmin.BgHealState
if opts.Maintenance {
// check if local disks are being healed, if they are being healed
// we need to tell healthy status as 'false' so that this server
// is not taken down for maintenance
var err error
aggHealStateResult, err = getAggregatedBackgroundHealState(ctx)
if err != nil {
logger.LogIf(logger.SetReqInfo(ctx, reqInfo), fmt.Errorf("Unable to verify global heal status: %w", err))
return HealthResult{
Healthy: false,
}
}
if len(aggHealStateResult.HealDisks) > 0 {
logger.LogIf(logger.SetReqInfo(ctx, reqInfo), fmt.Errorf("Total drives to be healed %d", len(aggHealStateResult.HealDisks)))
}
}
for zoneIdx := range erasureSetUpCount {
for setIdx := range erasureSetUpCount[zoneIdx] {
if erasureSetUpCount[zoneIdx][setIdx] < writeQuorum {
@@ -2063,10 +2082,11 @@ func (z *erasureZones) Health(ctx context.Context, opts HealthOptions) HealthRes
fmt.Errorf("Write quorum may be lost on zone: %d, set: %d, expected write quorum: %d",
zoneIdx, setIdx, writeQuorum))
return HealthResult{
Healthy: false,
ZoneID: zoneIdx,
SetID: setIdx,
WriteQuorum: writeQuorum,
Healthy: false,
HealingDrives: len(aggHealStateResult.HealDisks),
ZoneID: zoneIdx,
SetID: setIdx,
WriteQuorum: writeQuorum,
}
}
}
@@ -2081,21 +2101,6 @@ func (z *erasureZones) Health(ctx context.Context, opts HealthOptions) HealthRes
}
}
// check if local disks are being healed, if they are being healed
// we need to tell healthy status as 'false' so that this server
// is not taken down for maintenance
aggHealStateResult, err := getAggregatedBackgroundHealState(ctx)
if err != nil {
logger.LogIf(logger.SetReqInfo(ctx, reqInfo), fmt.Errorf("Unable to verify global heal status: %w", err))
return HealthResult{
Healthy: false,
}
}
if len(aggHealStateResult.HealDisks) > 0 {
logger.LogIf(logger.SetReqInfo(ctx, reqInfo), fmt.Errorf("Total drives to be healed %d", len(aggHealStateResult.HealDisks)))
}
return HealthResult{
Healthy: len(aggHealStateResult.HealDisks) == 0,
HealingDrives: len(aggHealStateResult.HealDisks),