MyFavMirrors/minio
1
0
Fork 0
mirror of https://github.com/minio/minio.git synced 2025-11-06 20:33:07 -05:00
Code Issues Packages Projects Releases Wiki Activity

fix: initialize new drwMutex for each attempt in 'for {' loop. (#14009)

Browse Source 
It is possible that GetLock() call remembers a previously
failed releaseAll() when there are networking issues, now
this state can have potential side effects.

This PR tries to avoid this side affect by making sure
to initialize NewNSLock() for each GetLock() attempts
made to avoid any prior state in the memory that can
interfere with the new lock grants.
This commit is contained in:
Harshavardhana
2022-01-02 09:15:34 -08:00
committed by GitHub
parent f527c708f2
commit 42ba0da6b0
3 changed files with 24 additions and 28 deletions
Download Patch File Download Diff File Expand all files Collapse all files

14
cmd/server-main.go
View File

@@ -297,13 +297,6 @@ func initServer(ctx context.Context, newObject ObjectLayer) ([]BucketInfo, error
// Once the config is fully loaded, initialize the new object layer.
setObjectLayer(newObject)
// Make sure to hold lock for entire migration to avoid
// such that only one server should migrate the entire config
// at a given time, this big transaction lock ensures this
// appropriately. This is also true for rotation of encrypted
// content.
txnLk := newObject.NewNSLock(minioMetaBucket, minioConfigPrefix+"/transaction.lock")
// **** WARNING ****
// Migrating to encrypted backend should happen before initialization of any
// sub-systems, make sure that we do not move the above codeblock elsewhere.
@@ -320,6 +313,13 @@ func initServer(ctx context.Context, newObject ObjectLayer) ([]BucketInfo, error
default:
}
// Make sure to hold lock for entire migration to avoid
// such that only one server should migrate the entire config
// at a given time, this big transaction lock ensures this
// appropriately. This is also true for rotation of encrypted
// content.
txnLk := newObject.NewNSLock(minioMetaBucket, minioConfigPrefix+"/transaction.lock")
// let one of the server acquire the lock, if not let them timeout.
// which shall be retried again by this loop.
lkctx, err := txnLk.GetLock(ctx, lockTimeout)