Project

General

Profile

Actions

Bug #21669

closed

osd: scrub scheduling can deadlock

Added by Sage Weil over 6 years ago. Updated about 6 years ago.

Status:
Rejected
Priority:
Urgent
Category:
-
Target version:
-
% Done:

0%

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

Description

both local and remote scrub reservations pull from the same pool, and the default limit of osd_max_scrubs is 1. If all osds try to scrub at once it will deadlock.

/a/sage-2017-10-03_21:58:15-rados-wip-sage-testing-2017-10-03-1358-distro-basic-smithi/1700216

Actions #1

Updated by Jesse Williamson over 6 years ago

  • Assignee set to Jesse Williamson
Actions #2

Updated by Jesse Williamson about 6 years ago

@Sage Weil Any further information on this?

Actions #3

Updated by Sage Weil about 6 years ago

  • Status changed from 12 to Rejected

I misunderstood this problem; the scrub scheduler also backs off after some time, so strictly speaking it isn't a deadlock.

Actions #4

Updated by Jesse Williamson about 6 years ago

As per conversation with Sage, this is likely Not a Bug (NaB).

Actions

Also available in: Atom PDF