Tuesday, 2019-05-14

*** itlinux has joined #openstack-swift00:14
*** itlinux has quit IRC00:22
*** itlinux has joined #openstack-swift00:23
*** itlinux has quit IRC00:25
*** itlinux has joined #openstack-swift00:26
*** itlinux has quit IRC00:41
*** itlinux has joined #openstack-swift00:43
*** itlinux has quit IRC00:47
zaitcev_timburke: forgot pep8 again? mind if I test and update it for you?01:01
*** baojg has joined #openstack-swift01:27
openstackgerritPete Zaitcev proposed openstack/swift master: py3: bulk -- alternate approach  https://review.opendev.org/65750501:31
*** baojg has quit IRC01:32
openstackgerritTimur Alperovich proposed openstack/swift master: Set Content-Type with s3api metadata updates.  https://review.opendev.org/65894501:40
openstackgerritPete Zaitcev proposed openstack/swift master: py3: slo  https://review.opendev.org/65050702:17
*** BjoernT has joined #openstack-swift02:32
*** BjoernT has quit IRC02:33
*** BjoernT has joined #openstack-swift02:34
*** BjoernT has quit IRC02:37
*** mvkr has quit IRC02:41
*** m75abrams has joined #openstack-swift02:49
*** m75abrams has quit IRC02:49
*** BjoernT has joined #openstack-swift02:51
*** mvkr has joined #openstack-swift02:55
*** gkadam has joined #openstack-swift03:28
*** mvkr has quit IRC04:03
*** pcaruana|afk| has joined #openstack-swift04:25
*** pcaruana|afk| has quit IRC04:35
openstackgerritMerged openstack/swift master: py3: port unit/proxy/test_server.py  https://review.opendev.org/65769904:43
openstackgerritMerged openstack/swift master: py3: finish porting proxy/test_server.py  https://review.opendev.org/65770004:43
openstackgerritMerged openstack/swift master: py3: port proxy/test_mem_server.py  https://review.opendev.org/65770504:43
*** BjoernT has quit IRC04:48
*** BjoernT has joined #openstack-swift04:48
*** BjoernT has quit IRC05:47
*** tesseract has joined #openstack-swift07:05
*** rcernin has quit IRC07:06
*** pcaruana has joined #openstack-swift07:12
*** hseipp has joined #openstack-swift07:12
*** rdejoux has joined #openstack-swift07:14
*** mikecmpbll has joined #openstack-swift07:56
*** tkajinam has quit IRC08:14
*** hseipp has quit IRC08:27
alecuyerclayg: You're probably busy with other things now, but just to let you know, with only a few changes, the work you did at the PTG to run the LOSF ansible playbooks works with a small change : https://github.com/alecuyer/vagrant-swift-all-in-one/tree/losf_playbook (I don't think I can push a PR to swiftstack's repo). That will start swift and the index server, and run functests :)09:05
*** e0ne has joined #openstack-swift09:13
*** [diablo] has quit IRC09:16
*** [diablo] has joined #openstack-swift09:24
kota_`AssertionError: False is not true : Status returned: 202 Expected: 201` :/09:44
alecuyerkota_:  hello! hmm I see the job output. Infra issue you think ?09:54
alecuyerRunning the functests here with the latest patch version let's see09:55
kota_alecuyer: o/09:55
kota_I think it's just wrong assertion or wrong setup of functests situation that it could happen periodically.09:56
kota_so just added `recheck` call.09:56
alecuyerok09:56
alecuyerI'll try to reproduce hangs with 1.10.0 today and get stacks, following your comments in the review09:57
kota_I don't think just a setup playbook for losf affects encryption in-process func.09:57
alecuyeryes..09:57
kota_alecuyer: sure, thanks.09:57
kota_i just mind the object-rpcmanager is using time.sleep instead of eventlet.sleep https://github.com/openstack/swift/blob/feature/losf/swift/obj/objectrpcmanager.py#L14009:59
kota_that might get stuck because the time module might not be maintained by eventlet greanthred if i understand correctly.10:00
kota_if all the grapc and rpcmanager are in green thread, it seems to work well, imo.10:03
*** takerian has joined #openstack-swift10:03
takerianHello, I came across one bug - in swauth there can occure race condition when creating a new token. All threads except for the first will return error.10:13
takerianI had written fix but Zuul failed because of readthedocs and SHA1 (not in my new code). Can someone please advise me how to proceed?10:13
takerian My commit is here: https://review.opendev.org/#/c/657621/10:14
patchbotpatch 657621 - x/swauth - Fix conflict error when creating new token - 2 patch sets10:14
alecuyerkota_: uh right, I'll change this. But issues we saw where in the object-server, getting messages such as "cannot switch thread", or otherwise erratic performance10:55
*** m75abrams has joined #openstack-swift11:09
*** e0ne has quit IRC11:31
*** e0ne has joined #openstack-swift11:31
openstackgerritMerged openstack/swift feature/losf: Setup losf functional test with saio setup  https://review.opendev.org/65707511:50
*** e0ne has quit IRC12:02
*** mahatic has joined #openstack-swift12:26
*** ChanServ sets mode: +v mahatic12:26
*** gkadam has quit IRC12:40
*** baojg has joined #openstack-swift12:48
*** baojg has quit IRC12:48
*** e0ne has joined #openstack-swift12:50
*** jistr is now known as jistr|call12:59
*** baojg has joined #openstack-swift13:22
*** jistr|call is now known as jistr13:30
*** BjoernT has joined #openstack-swift13:34
*** BjoernT has quit IRC13:36
*** BjoernT_ has joined #openstack-swift13:36
*** BjoernT has joined #openstack-swift13:37
*** BjoernT_ has quit IRC13:40
*** baojg has quit IRC13:44
*** persia has left #openstack-swift14:40
openstackgerritAlexandre Lécuyer proposed openstack/swift feature/losf: Use eventlet.sleep rather than time.sleep in objectrpcmanager.py  https://review.opendev.org/65910815:49
*** openstackgerrit has quit IRC15:54
*** e0ne has quit IRC15:57
*** gyee has joined #openstack-swift16:01
timburkegood morning16:05
timburkezaitcev_, thanks for getting that fixed up for me16:06
zaitcev_timburke, it adds argument checking into py2, but I observed that you did that in other places, so I went with that.16:21
*** mikecmpbll has quit IRC16:32
*** openstackgerrit has joined #openstack-swift16:48
openstackgerritTim Burke proposed openstack/swift master: Set Content-Type with s3api metadata updates.  https://review.opendev.org/65894516:48
timburkeit's the sort of thing that carries some risk (see also: https://bugs.launchpad.net/keystonemiddleware/+bug/1794179) but should but fine in bulk since we know we're expecting bytes from the wsgi.input16:55
openstackLaunchpad bug 1794179 in keystonemiddleware "Inserted auth_token headers aren't WSGI compliant" [High,In progress] - Assigned to Morgan Fainberg (mdrnstm)16:55
openstackgerritTim Burke proposed openstack/swift master: py3: port obj/reconstructor  https://review.opendev.org/65281917:33
claygalecuyer: that vsaio change looks solid!  can you do a PR against the swiftstack repo?17:34
*** itlinux has joined #openstack-swift17:42
*** itlinux has quit IRC17:48
*** itlinux has joined #openstack-swift17:54
*** e0ne has joined #openstack-swift17:57
*** mvkr has joined #openstack-swift17:58
*** e0ne has quit IRC18:05
*** BjoernT has quit IRC18:22
*** rdejoux has quit IRC18:37
*** e0ne has joined #openstack-swift18:38
*** BjoernT has joined #openstack-swift18:44
*** mikecmpbll has joined #openstack-swift18:59
*** itlinux has quit IRC19:16
*** pcaruana has quit IRC19:19
*** cwright has joined #openstack-swift19:38
*** m75abrams has quit IRC19:57
timburketeam photos from the PTG! https://www.dropbox.com/sh/fydqjehy9h5y728/AAAyDM_CxPvKIg5V1brySupna/Swift?dl=0&subfolder_nav_tracking=119:58
*** tesseract has quit IRC20:06
*** e0ne has quit IRC20:20
*** itlinux has joined #openstack-swift20:27
alecuyerclayg:  ok I did, see the note, if you'd like to just start the index server without running functests I think we would want a chef recipe instead of using the ansible run.yaml20:53
alecuyerkota_:  so, grpc 10.20.1 does run functest fine on SAIO as well. 10.10.1 will hang almost immediately, nothing in the log but "Timeout". As I recall, I think we did think for a while that from a certain version, it would work again, but we had hangs in production. I'll run 10.20.1 on a preprod machine tomorrow see if it still works with more concurrency20:55
openstackgerritTim Burke proposed openstack/swift master: Add operator tool to async-delete some or all objects in a container  https://review.opendev.org/65916220:59
timburkeneeds tests and a decent usage string, but ^^^ might actually be something we could have that'd be both useful and not-terrible21:00
*** itlinux has quit IRC21:07
*** BjoernT has quit IRC21:14
*** rcernin has joined #openstack-swift22:00
*** BjoernT has joined #openstack-swift22:35
*** mikecmpbll has quit IRC22:38
*** BjoernT has quit IRC22:39
*** BjoernT has joined #openstack-swift22:40
*** tkajinam has joined #openstack-swift22:52
mattoliveraumorning22:59
openstackgerritMerged openstack/swift feature/losf: Use eventlet.sleep rather than time.sleep in objectrpcmanager.py  https://review.opendev.org/65910823:04
*** BjoernT has quit IRC23:42
*** irclogbot_1 has quit IRC23:45
*** irclogbot_3 has joined #openstack-swift23:49
openstackgerritTim Burke proposed openstack/swift master: Add operator tool to async-delete some or all objects in a container  https://review.opendev.org/65916223:49
*** itlinux has joined #openstack-swift23:50
openstackgerritPete Zaitcev proposed openstack/swift master: py3: Port the tempurl middleware  https://review.opendev.org/65292823:59

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!