Saturday, 2019-05-04

openstackgerritMerged openstack/swift master: py3: Stop using stdlib's putrequest(); it only does ASCII  https://review.opendev.org/64289300:13
*** _david_sohonet has joined #openstack-swift00:41
*** dr_gogeta86 has quit IRC01:26
*** gmann has quit IRC01:28
*** _david_sohonet has quit IRC02:56
*** nottrobin has quit IRC04:09
*** nottrobin has joined #openstack-swift04:09
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: Do not use o_tmpfile_supported() in kvfile.py  https://review.opendev.org/65700304:11
*** gmann has joined #openstack-swift04:13
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: Resolve statsd module incompatibility  https://review.opendev.org/65704504:14
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: WIP: setup losf policy for in-process functional testing  https://review.opendev.org/65707504:14
timburkewell that's fun: http://logs.openstack.org/00/651100/4/gate/swift-tox-func-s3api/99c9f7e/job-output.txt.gz#_2019-05-03_23_29_47_35992604:43
timburkelooks like i'm brining back https://review.opendev.org/#/c/651277/ !04:43
patchbotpatch 651277 - swift - Fix ostestr invocation - 2 patch sets04:43
timburkenotmyname, yeah, that failure sure looks like eventual consistency issues showing up because of resource constraints...04:46
*** pcaruana has joined #openstack-swift04:53
kota_gah it looks like we need p 651277 asap! all losf gate looks failed with that.04:56
patchbothttps://review.opendev.org/#/c/651277/ - swift - Fix ostestr invocation - 2 patch sets04:56
kota_thanks mattoliverau!04:56
kota_at least, I'd like to port the change after it got merged to the master04:58
timburke:-( not good enough -- the rolling upgrade job runs tests from the *old* checkout!05:10
timburkefrom p 651277 check jobs: http://logs.openstack.org/77/651277/2/check/swift-multinode-rolling-upgrade/ed03b89/job-output.txt.gz#_2019-05-04_05_02_30_16806005:11
patchbothttps://review.opendev.org/#/c/651277/ - swift - Fix ostestr invocation - 2 patch sets05:11
timburke...which means switching to stestr won't help either05:12
kota_oh really?05:14
timburkei wonder if we need to have the rolling upgrade job enforce a suitably-old upper-constraints file for running the old tests...05:17
timburkei don't have a great way of correlating swift versions to https://opendev.org/openstack/requirements/src/branch/master/upper-constraints.txt hashes though...05:26
timburkemaybe it's the sort of thing where we tolerate setting the rolling upgrade non-voting until we can sort this all out? then the ostestr fix should be able to land...05:27
timburkethis also means that the next time i want to do an experimental check to verify rolling upgrade from queens (say), that's not gonna work either :-/05:28
timburke*maybe* we could backport the ostestr change?05:28
kota_hmmm05:30
kota_anyway, the failure gate other than our change is problematic in particular this week.05:31
kota_so making it non-voting sounds easier for a while05:37
kota_what version the rolling-upgrade bumps from?05:38
kota_queens?05:38
timburkelast tag, iirc05:41
kota_so, it's stein if it's the last.05:41
timburkeyeah. when we have a 2.22.0, it'll switch to that05:43
* kota_ is looking at the .zuul.yaml and playbooks.05:43
timburkei think we may have a problem on the stein branch though... it seems to be using master requirements (as was specified in 2.21.0's tox.ini) rather than stable/stein :-( :-( :-(05:44
timburkei'm debating about another idea: propose a change to blacklist the problematic version of stestr....05:45
timburkemaybe the non-voting job is better for now05:46
kota_ic, !=2.3.1?05:48
openstackgerritTim Burke proposed openstack/swift master: Fix ostestr invocation  https://review.opendev.org/65127705:49
timburkeyeah... idk how intentional it was for stestr to get updated though...05:50
timburkewe'll see how that goes, anyway05:50
kota_https://github.com/mtreinish/stestr/issues/231 oh tony05:51
kota_i think the master should follow the way of mattoliverau and in the future, the rolling upgrade test should be back to voting05:53
kota_so... it seems like the easier way is 1. land p 651277 w/o the voting, 2. backport it to stable branch, 3. tag the stable branches, 4. change the non-voting to be back to voting.05:57
patchbothttps://review.opendev.org/#/c/651277/ - swift - Fix ostestr invocation - 3 patch sets05:57
kota_it's still a few steps tho.05:57
*** mikecmpbll has joined #openstack-swift07:46
*** gmann has quit IRC08:02
*** mikecmpbll has quit IRC08:10
*** mikecmpbll has joined #openstack-swift08:10
*** baojg has joined #openstack-swift08:21
openstackgerritKazuhiro MIYAHARA proposed openstack/swift master: Add object-expirer new mode to execute tasks from general task queue  https://review.opendev.org/51738910:38
*** gmann has joined #openstack-swift12:05
*** ccamacho has quit IRC12:30
*** baojg has quit IRC13:23
*** baojg has joined #openstack-swift13:30
openstackgerritKazuhiro MIYAHARA proposed openstack/swift master: Add object-expirer new mode to execute tasks from general task queue  https://review.opendev.org/51738914:00
*** altlogbot_3 has quit IRC14:10
*** altlogbot_2 has joined #openstack-swift14:11
*** gmann has quit IRC14:34
openstackgerritKazuhiro MIYAHARA proposed openstack/swift master: Add object-expirer new mode to execute tasks from general task queue  https://review.opendev.org/51738914:51
*** tdasilva has joined #openstack-swift14:59
*** ChanServ sets mode: +v tdasilva14:59
openstackgerritThiago da Silva proposed openstack/swift master: Add Dockerfile to build a SAIO container image  https://review.opendev.org/65704615:05
*** itlinux has quit IRC15:21
*** e0ne has joined #openstack-swift15:34
openstackgerritMerged openstack/swift master: Fix ostestr invocation  https://review.opendev.org/65127715:35
kota_ok, p 657022 landed15:36
patchbothttps://review.opendev.org/#/c/657022/ - swift - tests: bust md5 of object not footers - 2 patch sets15:36
kota_no15:36
kota_mislink15:36
kota_p 651277 i meant15:36
patchbothttps://review.opendev.org/#/c/651277/ - swift - Fix ostestr invocation (MERGED) - 3 patch sets15:36
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: Merge remote-tracking branch 'remotes/origin/master' into merge-master  https://review.opendev.org/65712315:39
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: Add draft version of Makefile to build golang module  https://review.opendev.org/65109915:43
openstackgerritThiago da Silva proposed openstack/swift master: Add Dockerfile to build a SAIO container image  https://review.opendev.org/65704615:43
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: Update Makefile and README  https://review.opendev.org/65110015:43
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: Add losf package installation for gate job  https://review.opendev.org/65657615:43
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: WIP: setup losf policy for in-process functional testing  https://review.opendev.org/65707515:44
openstackgerritKota Tsuyuzaki proposed openstack/swift master: Enable to configure object-expirer in object-server.conf  https://review.opendev.org/60195015:45
kota_rebase... rebase... rebase...15:45
kota_ah... parhaps, just `recheck` might be ok when the gate merging the patch to the master15:46
*** ccamacho has joined #openstack-swift16:07
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: Do not use o_tmpfile_supported() in kvfile.py  https://review.opendev.org/65700316:09
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: WIP: setup losf policy for in-process functional testing  https://review.opendev.org/65707516:09
*** gmann has joined #openstack-swift16:12
*** ianychoi has quit IRC16:22
*** ccamacho has quit IRC16:30
openstackgerritTim Burke proposed openstack/swift feature/losf: Merge remote-tracking branch 'remotes/origin/master' into merge-master  https://review.opendev.org/65712316:46
kota_oh no? why the rolling upgrade gate job is now voting in losf feature branch? http://logs.openstack.org/23/657123/1/check/swift-multinode-rolling-upgrade/0cdfed3/zuul-info/inventory.yaml16:46
*** zaitcev has joined #openstack-swift16:48
*** ChanServ sets mode: +v zaitcev16:48
kota_oh, thanks timburke. I'm not sure why touching the .zuul.yaml effects to the gate behavior tho.16:51
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: Add draft version of Makefile to build golang module  https://review.opendev.org/65109916:52
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: Update Makefile and README  https://review.opendev.org/65110016:52
claygzaitcev: where did you drop that link?  I don't see it?16:53
zaitcevclayg: into wrong channel, sorry. https://review.opendev.org/60220516:54
patchbotpatch 602205 - swift - Refactor obj.server.ObjectController.PUT (MERGED) - 4 patch sets16:54
mattoliverauhttps://etherpad.openstack.org/p/swift-auto-sharding16:57
kota_alecuyer: http://logs.openstack.org/75/657075/7/check/swift-tox-func-losf/ab3f52e/job-output.txt.gz17:00
alecuyerthanks kota_17:00
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: Add losf package installation for gate job  https://review.opendev.org/65657617:03
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: Do not use o_tmpfile_supported() in kvfile.py  https://review.opendev.org/65700317:06
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: WIP: setup losf policy for in-process functional testing  https://review.opendev.org/65707517:06
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: Resolve statsd module incompatibility  https://review.opendev.org/65704517:14
*** e0ne has quit IRC17:14
*** pcaruana has quit IRC17:18
mattoliverauhttps://bugs.launchpad.net/swift/+bug/178129117:21
openstackLaunchpad bug 1781291 in OpenStack Object Storage (swift) "sharding: container GETs to root container get slow" [Medium,New]17:22
mattoliverauhttps://bugs.launchpad.net/swift/+bug/178129217:22
openstackLaunchpad bug 1781292 in OpenStack Object Storage (swift) "sharding: object reads may return 404s" [Medium,Confirmed]17:22
openstackgerritMerged openstack/swift master: Fix wrong tox env name from func-post-ec to func-ec  https://review.opendev.org/65707617:49
*** pcaruana has joined #openstack-swift18:10
*** pcaruana has quit IRC18:10
timburkeclayg, so tests pass on https://review.opendev.org/#/c/588103/18:15
patchbotpatch 588103 - swift - Unify auth_host/port/ssl/prefix func test options ... - 6 patch sets18:15
openstackgerritMerged openstack/swift master: versioned_writes: stop trying to catch HTTPPreconditionFailed  https://review.opendev.org/63801518:35
*** _david_sohonet has joined #openstack-swift18:36
openstackgerritAndreas Jaeger proposed x/pyeclib master: Update .zuul.yaml  https://review.opendev.org/65714818:59
openstackgerritTimur Alperovich proposed openstack/swift master: Start transition to boto3 in s3api tests.  https://review.opendev.org/65714919:14
timurtimburke: ^^^ that's the first take on moving to boto3 from boto in s3api tests19:15
openstackgerritAlexandre Lécuyer proposed openstack/swift feature/losf: WIP - set grpcio version to 1.3.3 in requirements.txt  https://review.opendev.org/65715019:16
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: Change losf functest to run on saio setup  https://review.opendev.org/65715319:28
kota_thanks alecuyer for p 657150!19:29
patchbothttps://review.opendev.org/#/c/657150/ - swift (feature/losf) - WIP - set grpcio version to 1.3.3 in requirements.txt - 1 patch set19:29
kota_alecuyer: hmm... it looks like no matter on grpcio version (looking at http://zuul.openstack.org/stream/659b49b7b7bb41bc9f73ad9fb6051082?logfile=console.log )19:40
*** gmann has quit IRC19:42
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: WIP: setup losf policy for in-process functional testing  https://review.opendev.org/65707519:46
alecuyerkota_:  oh it's hanging again :(19:51
kota_alecuyer: mine either http://zuul.openstack.org/stream/659b49b7b7bb41bc9f73ad9fb6051082?logfile=console.log however, fortunately it reaches to the functests smoothly even I changes the logic from tox to ansible playbook with saio setup so we will see the logs and setup configs in the gate.19:53
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: Change losf functest to run on saio setup  https://review.opendev.org/65715320:00
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: Change losf functest to run on saio setup  https://review.opendev.org/65715320:02
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: Change losf functest to run on saio setup  https://review.opendev.org/65715320:37
openstackgerritMerged x/pyeclib master: Update .zuul.yaml  https://review.opendev.org/65714820:38
*** altlogbot_2 has quit IRC20:56
*** altlogbot_3 has joined #openstack-swift21:00
timurtimburke: I'm not sure what's entailed in updating requirements, but happy to submit a patch for that21:03
timuris there a pointer to what needs to be done?21:03
*** altlogbot_3 has quit IRC21:04
*** altlogbot_0 has joined #openstack-swift21:06
timuroh, I see. You left a note about where to do that. I can submit a patch21:06
*** _david_sohonet has quit IRC21:06
*** _david_sohonet has joined #openstack-swift21:07
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: Change losf functest to run on saio setup  https://review.opendev.org/65715321:44
*** ianychoi has joined #openstack-swift21:44
*** Fruit has joined #openstack-swift21:47
Fruithi! anyone here familiar with pyeclib? the github repo says that this is the irc channel for it21:49
kota_Fruit: we're in Denver, 3 core maintainer are here.21:50
kota_maintainers21:50
kota_ah, it might be more, swift-core groups might be same with pyeclib's one.21:50
Fruitoh wow :)21:51
kota_what's up on pyeclib?21:51
FruitI'm a novice, trying to get it to work. https://www.fruit.je/m/pyeclib.py21:51
Fruitit doesn't seem to detect any checksum errors21:51
kota_verify_stripe_metadata...21:52
timurtimburke: I uploaded a change to requirements: https://review.opendev.org/#/c/657167/. Does that look right? Is there anything else I need to do?21:53
patchbotpatch 657167 - requirements - Add boto3 to global requirements. - 1 patch set21:53
Fruitaccording to the docstring that should either return a list or None. instead it returns {'status': 0} for me21:53
timurrunning the generate-constraints script didn't show anything aside from changes related to me using python 3.7 (I didn't install 3.6 just for this, but maybe I should?)21:54
zaitcevFruit: the consensus is... auditor does not do this for some reason, so this verification is not tested properly21:56
openstackgerritTimur Alperovich proposed openstack/swift master: Start transition to boto3 in s3api tests.  https://review.opendev.org/65714921:56
Fruitzaitcev: I'm afraid I'm not familiar with "auditor" but doesn't that sound bad?21:57
zaitcevFruit: I suppose I should look at it, but it's going to take forever. If you post a patch, it could be easier. Sorry for blowing you off like that.21:59
zaitcevat the very least the docstring must not lie21:59
Fruitwhelp okay. that will require some diving21:59
zaitcevSince EC was reliable in production, nobody touched PyECLib in more than a year.22:00
zaitcevHonestly I don't remember what it does or does not...22:00
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: Change losf functest to run on saio setup  https://review.opendev.org/65715322:01
timburkelooks like the C docs are more accurate: https://github.com/openstack/pyeclib/blob/master/src/c/pyeclib_c/pyeclib_c.c#L1083-L108522:01
Fruitah! but that means it still didn't detect my corrupted byte22:02
timburkeyup! so there's definitely a bug somewhere :-)22:02
FruitI should try a different ec_type22:03
timburkewhat are you using (or evaluating) pyeclib for? i'd love to help this get wider usage (among other things it'd help root out bugs like this)22:03
FruitI'm writing (yet another) backup tool and would like to be able to detect and reconstruct corrupted file data22:04
zaitcevsomeone is burned by Duplicity's unicode bugs perhaps?22:04
timburkevery cool! at the moment, its primary (only?) consumer is openstack swift; like i said, i'd love to have more people using it22:05
Fruithaha. no, never used duplicity.22:05
Fruitmore of a replacement for backuppc if you're familiar with it22:05
Fruitthis thing of mine has a compression / encryption / etc plugin system for data storage and I'd like to add FEC like functionality to that list22:07
timburkehmm... the naming of the fragment_metadata_list arg makes me think it may not be intended for the whole fragment... that's... not super-obvious :-(22:07
Fruitwell if it's just a matter of "holding it wrong"...22:08
kota_FWIW, test says it returns dict https://github.com/openstack/pyeclib/blob/master/test/test_pyeclib_api.py#L423-L46022:08
claygkota_: alecuyer: I'm seeing `ImportError: No module named google.protobuf.internal` - isn't that something we build/install via the Makefile in p 65657622:08
patchbothttps://review.opendev.org/#/c/656576/ - swift (feature/losf) - Add losf package installation for gate job - 15 patch sets22:08
kota_but {status: 0} seems wiered22:08
Fruitif I change it to check a single buf it segfaults22:09
kota_clayg: https://review.opendev.org/#/c/657075/922:09
patchbotpatch 657075 - swift (feature/losf) - WIP: setup losf policy for in-process functional t... - 9 patch sets22:09
claygusing the tools/playbooks in a vsaio is awesome fun!22:09
clayg https://github.com/swiftstack/vagrant-swift-all-in-one/pull/8022:09
kota_sorry, that's in latter patch.22:10
kota_clayg: we found the issue when we actually run the module at the gate.22:10
timburkeFruit, segfault or stack trace?22:12
Fruit[1]    13732 segmentation fault22:13
timburkewhoa! that *definitely* needs fixing22:13
timburkewhat version of pyeclib is it?22:13
Fruit1.3.1-1ubuntu322:13
Fruitprobably ancient22:13
Fruit04 Aug 201722:14
Fruitthen again most of the commits on github since that date seem to be build system related22:14
FruitI'll check if c564b69731b7330bd491afca4a95c65cae9fdb2a is included since that one seems segfault related22:15
timburkethis one seems promising: https://github.com/openstack/pyeclib/commit/95c742322:16
openstackgerritTimur Alperovich proposed openstack/swift master: Fix the handling of keystone groups in tempauth.  https://review.opendev.org/63041522:16
Fruityeah that's the one I'm checking :)22:16
zaitcevAren't we on pyeclib 1.6 something now?22:16
zaitcevn/m22:17
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: Change losf functest to run on saio setup  https://review.opendev.org/65715322:17
alecuyerclayg:  I forgot to set grpcio to a version that would play nice with eventlet. These later grpcio python packages do not have protobuf as a dependency anymore, which is why you saw that ImportError. Now, Kota has added protobuf explicitely in requirements.txt, so it shouldn't happen again, regardless of the version we use22:17
Fruityep not included. I need a newer version.22:17
timburkefwiw, it raises an exception: "pyeclib.ec_iface.ECInvalidParameter: pyeclib_c_encode ERROR: Invalid arguments. Please inspect syslog for liberasurecode error report."22:18
timburkewhich still isn't terribly obvious/useful22:18
zaitcevalways was like that, since the times of Kevin/Tushar22:19
Fruithmm I'd be content with that22:19
zaitcevReally makes me mad when happens in the gate where I cannot see the syslog.22:20
Fruitanyway I'll update my stuff, retest, and report back when things are still unsatisfactory22:20
Fruits/when/if/22:21
timburke"when" sounds likely ;-)22:21
Fruitwe'll see :)22:22
claygwhy is it sometimes object-rpcmanager and sometimes object-rpc-manager22:25
claygoh oh oh - i know why!22:27
clayghttps://github.com/openstack/swift/blob/master/swift/common/daemon.py#L27122:27
claygObjectRpcManager -> object-rpc-manager22:28
claygi get this now: AttributeError: Invalid or missing volcheck in your config file22:31
claygalthough I'm not sure that we shouldn't do something to change the section name ...22:31
openstackgerritClay Gerrard proposed openstack/swift feature/losf: Change section to [object-rpcmanager]  https://review.opendev.org/65716922:33
openstackgerritClay Gerrard proposed openstack/swift feature/losf: WIP - set grpcio version to 1.3.3 in requirements.txt  https://review.opendev.org/65715022:35
openstackgerritMerged openstack/swift master: Remove py35 unit test job  https://review.opendev.org/65703322:38
Fruittimburke: I'm afraid it turns out to be "when" :) the segfault is fixed by updating to git master (aka 1.6.0) but it still can't into crc it seems22:42
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: Change losf functest to run on saio setup  https://review.opendev.org/65715322:48
openstackgerritMerged openstack/swift master: py3: port s3api  https://review.opendev.org/65189322:49
kota_it looks like zuul doesn't like me :,(  it's just a merge commit https://review.opendev.org/#/c/657123/22:50
patchbotpatch 657123 - swift (feature/losf) - Merge remote-tracking branch 'remotes/origin/maste... - 2 patch sets22:50
openstackgerritTim Burke proposed openstack/swift master: Wait longer for log lines in unit test  https://review.opendev.org/65279122:58
*** tdasilva has quit IRC22:59
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: Change losf functest to run on saio setup  https://review.opendev.org/65715322:59
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: Change section to [object-rpcmanager]  https://review.opendev.org/65716923:12
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: WIP: setup losf policy for in-process functional testing  https://review.opendev.org/65707523:12
openstackgerritKota Tsuyuzaki proposed openstack/swift feature/losf: Change losf functest to run on saio setup  https://review.opendev.org/65715323:18
openstackgerritTim Burke proposed openstack/swift master: py3: start porting for unit/proxy/test_server.py  https://review.opendev.org/65717223:35
alecuyer[object-rpc-manager]23:40
alecuyervolcheck=/usr/local/bin/swift-losf-volume-check23:40
alecuyerlosf_bin=/vagrant/go/bin/swift-rpc-losf23:40
*** zaitcev has quit IRC23:47
openstackgerritMerged openstack/swift feature/losf: Merge remote-tracking branch 'remotes/origin/master' into merge-master  https://review.opendev.org/65712323:52
openstackgerritMerged openstack/swift feature/losf: Add draft version of Makefile to build golang module  https://review.opendev.org/65109923:52
openstackgerritMerged openstack/swift feature/losf: Update Makefile and README  https://review.opendev.org/65110023:52
openstackgerritMerged openstack/swift feature/losf: Add losf package installation for gate job  https://review.opendev.org/65657623:52

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