After replication of object to another shard, the object remains on the returned disk forever #1079

New issue

Open

opened 2024-04-08 06:40:03 +00:00 by anikeev-yadro · 0 comments

anikeev-yadro commented

2024-04-08 06:40:03 +00:00

Member

https://github.com/nspcc-dev/neofs-node/issues/2087

Expected Behavior

Policer should scan and delete duplicates of objects on local shards

Current Behavior

After replication of object to another shard, the object remains on the returned disk forever.

Steps to Reproduce (for bugs)

Remove disk with selected object in the blobstor
Await of the object's replication to another shard
Return the disk and change the shards state to read-write
All replicated object will not be cleaned.

Regression

Version

FrostFS Storage node
Version: v0.38.2
GoVersion: go1.22.0

Your Environment

Cloud and HW

https://github.com/nspcc-dev/neofs-node/issues/2087 ## Expected Behavior Policer should scan and delete duplicates of objects on local shards ## Current Behavior After replication of object to another shard, the object remains on the returned disk forever. ## Steps to Reproduce (for bugs) 1) Remove disk with selected object in the blobstor 2) Await of the object's replication to another shard 3) Return the disk and change the shards state to read-write All replicated object will not be cleaned. ## Regression No ## Version ``` FrostFS Storage node Version: v0.38.2 GoVersion: go1.22.0 ``` ## Your Environment Cloud and HW