Bug 1578509 - entries_behind_master metric output but "rbd mirror image status " never reduces to zero.
Summary: entries_behind_master metric output but "rbd mirror image status " never redu...
Keywords:
Status: CLOSED ERRATA
Alias: None
Product: Red Hat Ceph Storage
Classification: Red Hat Storage
Component: RBD-Mirror
Version: 3.0
Hardware: Unspecified
OS: Linux
medium
medium
Target Milestone: z4
: 3.0
Assignee: Jason Dillaman
QA Contact: Vasishta
Erin Donnelly
URL:
Whiteboard:
Depends On:
Blocks: 1585192
TreeView+ depends on / blocked
 
Reported: 2018-05-15 18:14 UTC by Bob Emerson
Modified: 2021-06-10 16:11 UTC (History)
9 users (show)

Fixed In Version: ceph-12.2.4-24.el7cp
Doc Type: Bug Fix
Doc Text:
Previously, the "entries_behind_master" metric output from the "rbd mirror image status" CLI tool did not always reduce to zero under synthetic workloads. This metric would only be updated after 32 write operations, or after an I/O flush was requested. This could cause a false alarm that there is an issue with RBD mirroring replications even though it is properly functioning. With this update, the metric is now updated periodically without the need for an explicit I/O flush in the workload.
Clone Of:
: 1585192 (view as bug list)
Environment:
Last Closed: 2018-07-11 18:11:10 UTC
Embargoed:

Attachments (Terms of Use)


Links
System ID Private Priority Status Summary Last Updated
Ceph Project Bug Tracker 23516 0 None None None 2018-05-16 23:00:14 UTC
Red Hat Product Errata RHSA-2018:2177 0 None None None 2018-07-11 18:12:03 UTC

Description Bob Emerson 2018-05-15 18:14:28 UTC 
Description of problem:

Customer is running into the following upstream issue:

Ref upstream bug: https://tracker.ceph.com/issues/23457

entries_behind_master metric output but "rbd mirror image status " never reduces to zero.



Version-Release number of selected component (if applicable):

12.2.1-40.el7cp


How reproducible:


Customer can reproduce on demand
Comment 3 Bob Emerson 2018-05-15 18:17:12 UTC 
Customer would like to know when this will be addressed in the downstream version.
Comment 4 Jason Dillaman 2018-05-16 23:04:32 UTC 
It hasn't been fixed upstream yet since it has always been deemed as a low priority. The commit position (entries behind master is derived from this) is updated every 32 IOs or after each flush request. Was this from a synthetic workload that is just writing and never flushing?
Comment 16 Vasishta 2018-06-22 14:38:17 UTC 
Working fine with 12.2.4-27

Moving to VERIFIED state, please let me know if there are any concerns.

Regards,
Vasishta Shatsry
AQE, Ceph
Comment 18 errata-xmlrpc 2018-07-11 18:11:10 UTC 
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHSA-2018:2177
Note You need to log in before you can comment on or make changes to this bug.