[LU-14594] reply to the original RPC can land to the resent RPC reply buffer Created: 08/Apr/21  Updated: 08/Sep/23  Resolved: 27/May/21

Status: Resolved
Project: Lustre
Component/s: None
Affects Version/s: None
Fix Version/s: Lustre 2.15.0

Type: Bug Priority: Minor
Reporter: Vitaly Fertman Assignee: Vitaly Fertman
Resolution: Fixed Votes: 0
Labels: None

Severity: 3
Rank (Obsolete): 9223372036854775807

 Description   

whereas the server is able to filter by the connection ID, and drop late coming RPCs of previous connections, it does not work for replies. At the same time, this is a problem in some cases and we have to invent another fix each time.
one of the last cases was a late came enqueue resent which was handled as a new req (original lock was already cancelled) and was sent back with AST_SENT flag, as client does not wait for this reply anymore (the reply on the original RPC matched), it led to the eviction.



 Comments   
Comment by Gerrit Updater [ 08/Apr/21 ]

Vitaly Fertman (vitaly.fertman@hpe.com) uploaded a new patch: https://review.whamcloud.com/43242
Subject: LU-14594 ptlrpc: do not match reply with resent RPC
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: 86c19b0e4fc062546038e2c02fe355d93af8b4b0

Comment by Gerrit Updater [ 27/May/21 ]

Oleg Drokin (green@whamcloud.com) merged in patch https://review.whamcloud.com/43242/
Subject: LU-14594 ptlrpc: do not match reply with resent RPC
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: 057fafc018d7369dd9ec74641286446f5ba38491

Comment by Peter Jones [ 27/May/21 ]

Landed for 2.15

Generated at Sat Feb 10 03:11:05 UTC 2024 using Jira 9.4.14#940014-sha1:734e6822bbf0d45eff9af51f82432957f73aa32c.