Project

General

Profile

Actions
Copy link

Bug #14595

closed

Race between releasing exclusive lock and incoming write ops

Added by Jason Dillaman about 8 years ago. Updated about 8 years ago.

Status:
Resolved
Priority:
Normal
Assignee:
-
Target version:
-
% Done:

0%

Source:
other
Tags:
Backport:
hammer, infernalis
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Pull request ID:
Crash signature (v1):
Crash signature (v2):

Description

A thread releasing the exclusive lock (and flushing the cache) might race with a thread submitting a new write operation. The write IO should be queued but instead it is processed. This can result in the cache having writeback bufferheads while it doesn't own the exclusive lock -- which will get flushed at some point.

Actions
Copy link
 #1

Updated by Jason Dillaman about 8 years ago

  • Status changed from Pending Backport to New

Should verify it's not an issue in master due to the major refactoring of librbd.

Actions
Copy link
 #2

Updated by Jason Dillaman about 8 years ago

  • Status changed from New to Resolved

This is not an issue within jewel due to the new block-writes ability and exclusive lock state machine.

Actions
Copy link

Also available in: Atom PDF