[Openstack] [swift] Object replication failure counts confusing in 2.7.0

Mark Kirkwood mark.kirkwood at catalyst.net.nz
Fri May 20 08:22:57 UTC 2016

On 20/05/16 20:03, Pete Zaitcev wrote:
> On Wed, 18 May 2016 16:46:05 +1200
> Mark Kirkwood <mark.kirkwood at catalyst.net.nz> wrote:
>> May 18 04:31:17 markir-dev-ostor002 object-server: object replication
>> failure 4, detail Traceback (most recent call last):#012  File
>> "/opt/cat/openstack/swift/local/lib/python2.7/site-packages/swift/obj/replicator.py",
>> line 622, in build_replication_jobs#012 int(partition))#012ValueError:
>> invalid literal for int() with base 10: 'auditor_status_ALL.json'#012
> Mark, I saw the patch you attached to the bug 1583305, but it only deals
> with counting of failures. It does nothing to ignore auditor's files, it seems.
> Would you be willing to cook up something like Tim's fix in commit
> ad16e2c77bb61bdf51a7d3b2c258daf69bfc74da
> ?
> -- Pete

Hi Peter,

My patch was meant to be in addition to Clay's (attached earlier to 
bug), which does have logic to handle the auditor's files, sorry if that 
wasn't clear.And I agree with what you are saying.

Doing some testing with just Clay's patch applied, I'm not seeing any 
'extra' errors from the timeout sections that I was seeing 
previously...now it might be that my test setup is just a bit different 
now, or that Clay's changes stop the othererrors happening (I'll watch a 
bit more).

My next head scratching moment is the *success* count seems too high(I 
know...I'm complaining about too much success- heh...some folk never 
satisfied) - I have a 3 device system (3 replicas) with 2 objects 
uploaded...and the replication stats are saying "12 successes"...err- 
was expecting *6 successes* - wtf? I need to examine how these counts 
are being calculated, as it seems theycould possibly be not quite right.



More information about the Openstack mailing list