Project

General

Profile

Actions
Copy link

Bug #7165

closed

high ping latency in vm using rbd writeback cache in qemu 1.7 with async flush

Added by Josh Durgin over 10 years ago. Updated about 10 years ago.

Status:
Resolved
Priority:
Urgent
Assignee:
Josh Durgin
Target version:
-
% Done:

0%

Source:
Community (user)
Tags:
Backport:
Regression:
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Pull request ID:
Crash signature (v1):
Crash signature (v2):

Description

See thread on qemu-devel: http://lists.nongnu.org/archive/html/qemu-devel/2014-01/msg01993.html

It happens only when the cache is in writeback mode, whether qemu is set to cache=none or cache=writeback.

Files

test-logs.tgz (32.3 MB) test-logs.tgz Josh Durgin, 01/15/2014 11:05 PM
Actions
Copy link
 #1

Updated by Ian Colle about 10 years ago

  • Priority changed from High to Urgent
Actions
Copy link
 #2

Updated by Josh Durgin about 10 years ago

  • Status changed from New to Resolved

From the logs it was caused by the ObjectCacher's flusher thread holding the lock too long. Backported the fixes for that to dumpling (commit:f3b5ba6f25010291a2918bdd286f1b39570bb90 commit:48895a46015c9d6d67543816f5a400c21aa206b1) and emperor (commit:25c32482046a3925024ef996ecc98053a7358e61).

Actions
Copy link

Also available in: Atom PDF