Friday, 2015-09-11

*** rhagarty__ has joined #openstack-manila00:05
*** rhagarty__ has quit IRC00:07
*** chlong has quit IRC00:19
bswartzcsaba: I just had questions about ganesha for you00:20
bswartzif you're still around00:20
csababswartz: yes I'm here, I'm glad if I can help you with Ganesha00:20
*** rushil has quit IRC00:31
*** jasonsb has quit IRC00:39
*** jasonsb has joined #openstack-manila00:43
*** chen12 has joined #openstack-manila00:47
*** jcsp has quit IRC00:51
openstackgerritCsaba Henk proposed openstack/manila: doc: turn ascii art tables into proper reST grid tables  https://review.openstack.org/22242400:55
*** jasonsb has quit IRC00:55
*** chlong has joined #openstack-manila01:02
*** jasonsb has joined #openstack-manila01:08
*** chlong has quit IRC01:08
*** chlong has joined #openstack-manila01:20
*** cknight has joined #openstack-manila01:30
chen12bswartz, ping01:31
chen12cknight,  ping01:31
cknightchen12: pong01:32
chen12cknight,  just a quick question, does  consistency_group are "MUST" feature for every driver ?01:32
cknightchen12: no, definitely not!01:33
cknightchen12: as of liberty, even snapshots are optional.01:33
chen12cknight, ok, great ! Thanks!  Then I can safely  disable the related tests to make HDFS CI pass....01:33
chen12cknight, do we have a min required function list for each driver ?01:34
cknightchen12: yes, and yes.  the minimum feature set was discussed at length at the mid-cycle meetup.01:34
cknightchen12: let me check on the status of that01:35
cknightchen12: will you be in tokyo?01:35
chen12cknight, nope!  :(01:35
cknightchen12: next time perhaps!01:35
bswartzchen12: are you working on getting the HDFS tests to pass?01:40
bswartzI have seen them all failing for a long time01:40
bswartzI hope the issues are minor and easy to fix01:41
chen12bswartz, just a week... it caused by 1, you're starting to using tempest plugin  2, the consistency_group test is failing01:41
chen12bswartz, I'm working on it01:41
bswartzokay01:41
chen12bswartz, by the way, all drivers "must" adapt the "share instance" in the future, right ?01:42
bswartzchen12: that feature should be transparent to drivers01:43
bswartzdrivers no longer see share they see share instances01:43
bswartzbut the python object looks the same01:43
bswartzthat change should not have broken any existing drivers01:43
chen12bswartz,  just wonder do I need to update HDFS driver for this ?01:43
chen12bswartz, because I see lots of "warnings" as things are deprecated..01:45
bswartzwhich warnings?01:46
bswartzyou shouldn't need to update the driver01:46
*** u_glide has quit IRC01:47
bswartzwe should take a look at any warnings01:48
bswartzalthough if they're related to share instances it might be better to ask u_glide about it when he wakes up01:48
bswartzI'm about to go to sleep01:48
chen12bswartz, ok, I will check it later and ask .01:50
chen12bswartz, :)01:50
cknightchen12: After reviewing the mid-cycle notes as well as the driver interface, I would think a DHSS == False driver would need share create, delete, allow_access, deny_access, _update_share_stats, and get_pool (if you support pools).  In Mitaka, I think we'll also ask drivers to support extend_share (shrink will always remain optional).  If you have further questions, I suggest you ask the Manila mailing list since other CI maintai01:54
chen12cknight, Thanks! understand this now. :)01:56
*** bill_az has quit IRC01:57
cknightchen12: To add a bit more color, we'd love to see more feature-rich drivers in Manila, but we've realized there are also use cases where all that is required (or feasible) is a backend that offers lots of file shares without all the fancy stuff.  I think all the drivers we have now can meet the minimum standard we have established.  Like Ben, I look forward to seeing your CI system reporting successful runs.02:00
*** EinstCrazy has joined #openstack-manila02:00
*** baojg has joined #openstack-manila02:01
chen12cknight, sure. I will make the CI pass before the end of this week.02:04
cknightchen12: Great!02:04
*** gouthamr_ has quit IRC02:19
*** cknight has quit IRC02:53
*** gouthamr has joined #openstack-manila03:00
openstackgerritMerged openstack/manila: Ignore git backup merge files  https://review.openstack.org/21812303:10
*** vbellur has quit IRC03:16
openstackgerritli,chen proposed openstack/manila: Enable extend_share in HDFS driver  https://review.openstack.org/22244703:19
*** gouthamr_ has joined #openstack-manila03:20
openstackgerritMerged openstack/manila: Fix the typo "version"  https://review.openstack.org/21539403:21
openstackgerritMerged openstack/manila: Replaces xrange() with range() for py 2/3 compatibility  https://review.openstack.org/21238503:21
*** gouthamr has quit IRC03:23
*** haomaiwang has joined #openstack-manila03:25
*** haomaiwang has quit IRC03:25
*** bswartz has quit IRC03:29
*** haomaiwang has joined #openstack-manila03:34
*** bswartz has joined #openstack-manila03:36
*** haomaiwang has quit IRC04:01
*** haomaiwa_ has joined #openstack-manila04:01
*** EinstCrazy has quit IRC04:03
*** rraja has joined #openstack-manila04:15
*** ktolstoy has joined #openstack-manila04:24
*** ashokchandran has quit IRC04:25
*** ktolstoy has quit IRC04:25
*** BharatK has joined #openstack-manila04:47
*** haomaiwa_ has quit IRC05:01
*** haomaiwang has joined #openstack-manila05:01
*** ashokchandran has joined #openstack-manila05:12
*** ashokchandran has quit IRC05:13
*** jwang_ has quit IRC05:13
*** BharatK has quit IRC05:20
*** EinstCrazy has joined #openstack-manila05:33
*** BharatK has joined #openstack-manila05:33
*** shausy has joined #openstack-manila05:37
*** deepakcs has joined #openstack-manila05:41
*** nkrinner has joined #openstack-manila05:48
*** shausy has quit IRC05:48
*** gouthamr_ has quit IRC05:58
*** haomaiwang has quit IRC06:01
*** haomaiwang has joined #openstack-manila06:01
*** vbellur has joined #openstack-manila06:07
*** jasonsb has quit IRC06:13
openstackgerritli,chen proposed openstack/manila: Check the snapshot directory before copy it  https://review.openstack.org/22247906:16
*** jasonsb has joined #openstack-manila06:30
*** jasonsb has quit IRC06:36
*** anubhav has joined #openstack-manila06:41
*** haomaiw__ has joined #openstack-manila06:53
*** haomaiwang has quit IRC06:55
*** haomaiw__ has quit IRC07:01
*** haomaiwang has joined #openstack-manila07:01
*** EinstCrazy has quit IRC07:07
openstackgerritli,chen proposed openstack/manila: Check the snapshot directory before copy it  https://review.openstack.org/22247907:13
*** breitz has quit IRC07:16
*** breitz has joined #openstack-manila07:17
openstackgerritzhongjun proposed openstack/manila: Fix create snapshot API in Huawei driver  https://review.openstack.org/22250107:28
*** anubhav has quit IRC07:30
openstackgerritSilvan Kaiser proposed openstack/manila: Corrects capabilities returned by Quobyte Manila driver  https://review.openstack.org/21555607:31
*** lpetrut has joined #openstack-manila07:45
*** haomaiwang has quit IRC07:49
*** baojg has quit IRC08:05
*** baojg has joined #openstack-manila08:09
*** ktolstoy has joined #openstack-manila08:14
*** anubhav has joined #openstack-manila08:22
openstackgerritValeriy Ponomaryov proposed openstack/manila: Add 'snapshot_support' attr to share details  https://review.openstack.org/22123108:30
*** haomaiwang has joined #openstack-manila08:44
*** haomaiwang has quit IRC09:01
*** haomaiwang has joined #openstack-manila09:01
*** anubhav has quit IRC09:10
*** chlong has quit IRC09:16
*** jcsp has joined #openstack-manila09:56
*** haomaiwang has quit IRC10:01
*** haomaiwang has joined #openstack-manila10:01
*** jcsp has quit IRC10:06
*** jcsp has joined #openstack-manila10:06
*** baojg has quit IRC10:14
*** baojg has joined #openstack-manila10:14
*** baojg has quit IRC10:19
*** resker has joined #openstack-manila10:23
*** esker has quit IRC10:26
*** ociuhandu has quit IRC10:29
openstackgerritMerged openstack/puppet-manila: acceptance: switch RDO to CBS repositories  https://review.openstack.org/22227510:29
*** akerr has joined #openstack-manila10:52
openstackgerritRodrigo Barbieri proposed openstack/manila: Fix Share Migration tempest tests  https://review.openstack.org/22191610:54
*** akerr_ has joined #openstack-manila10:58
*** ktolstoy has quit IRC10:59
*** ktolstoy has joined #openstack-manila11:00
*** haomaiwang has quit IRC11:01
*** 18VAANNOW has joined #openstack-manila11:01
*** akerr has quit IRC11:01
*** ktolstoy has quit IRC11:03
gansovponomaryov: ping11:06
*** anubhav has joined #openstack-manila11:07
*** ktolstoy has joined #openstack-manila11:09
*** alyson_ has joined #openstack-manila11:09
*** ktolstoy has quit IRC11:10
*** ktolstoy has joined #openstack-manila11:11
*** akerr_ has quit IRC11:14
*** ociuhandu has joined #openstack-manila11:17
*** ktolstoy has quit IRC11:17
*** ktolstoy has joined #openstack-manila11:17
*** ktolstoy has quit IRC11:21
*** rraja has quit IRC11:22
vponomaryovganso: pong11:23
*** ktolstoy has joined #openstack-manila11:23
gansoHello, do you know any way to disable Shrink tests for vendor CIs that do not support it?11:24
gansovponomaryov: besides using an exclusion regex11:24
vponomaryovganso: set opt "run_shrink_tests" to False11:25
vponomaryovTempest opt in group [share]11:25
vponomaryovganso: https://github.com/openstack/manila/blob/master/manila_tempest_tests/config.py11:25
gansovponomaryov: hmmmm I was looking for env vars11:26
gansovponomaryov: shouldn't there be one?11:26
vponomaryovganso: why?11:26
vponomaryovganso: it can be added, exactly in your Ci system config11:27
gansovponomaryov: CG and multibackend done this way11:27
gansovponomaryov: through env vars11:28
vponomaryovganso: if you want, you can add it11:28
gansovponomaryov: also, multitenancy11:28
vponomaryovganso: I do not mind11:28
gansovponomaryov: ok, I think I will submit a patch for it later11:28
gansovponomaryov: also, I am trying to test manila_tempest_tests.tests.scenario.test_share_basic_ops.TestShareBasicOpsCIFS.test_read_write_two_vms[compute,network]11:29
gansovponomaryov: it was working before, now I am getting this error on a fresh devstack: http://paste.openstack.org/show/456622/11:29
gansovponomaryov: any idea how to fix it? I added several rules to the default security group manually11:30
vponomaryovganso: it can be our known bug - https://bugs.launchpad.net/manila/+bug/148087011:30
openstackLaunchpad bug 1480870 in Manila "Service instance does not get ssh key with Generic driver" [Medium,Confirmed]11:30
vponomaryovganso: enable password for service instance as it is done for CI11:31
vponomaryovganso: so, there is some concurrency bug, that leads to some breakage of routing rules of a service VMs11:32
vponomaryovganso: 1 thread for scenario tests workarounds it11:32
gansovponomaryov: I added service_instance_password entries, that fixed the first error I got which was when trying to SSH to share-servers, but then the error I get is when the test tries to SSH into guest instances to mount shares11:32
vponomaryovganso: right11:33
gansoI am already running only one test, and only one thread: tox -eall-plugin manila_tempest_tests.tests.scenario.test_share_basic_ops.TestShareBasicOpsCIFS.test_read_write_two_vms -- --concurrency=1 | tee test2.log11:33
vponomaryovganso: routing rules are messed up for some reason11:33
gansovponomaryov: are there any workarounds I can do manually just to run my tests?11:33
vponomaryovganso: my experience shown that if error appeared then it will continue appear11:34
vponomaryovganso: have you restarted neutron services?11:34
vponomaryovganso: l3 agent, dhcp agent?11:34
gansovponomaryov: I have not tried restarting services yet, but attempting to run the test several times ends up in the same error11:34
vponomaryovganso: after hitting bug11:34
gansovponomaryov: I will try restarting services11:34
gansovponomaryov: is this why our scenario tests are unstable?11:35
vponomaryovganso: yes11:35
vponomaryovganso: adding password support we workarounded our voting jobs with API tests11:36
vponomaryovganso: but root cause stil lexist11:36
gansovponomaryov: hmm right11:37
vponomaryovganso: it is related to module "service_instance" and its setup of connectivity11:37
vponomaryovganso: so, two drivers are affected - Generic and Windows11:38
gansovponomaryov: in fact... all drivers are unable to run scenario tests properly due to this11:38
vponomaryovganso: there are VMs that are being created in parallel11:39
openstackgerritJulia Varlamova proposed openstack/manila: Make scenario tests able to run with DHSS=False  https://review.openstack.org/21880111:49
*** 18VAANNOW has quit IRC12:01
*** haomaiwang has joined #openstack-manila12:01
bswartzvponomaryov: https://bugs.launchpad.net/manila/+bug/148087012:02
openstackLaunchpad bug 1480870 in Manila "Service instance does not get ssh key with Generic driver" [Medium,Confirmed]12:02
bswartzvponomaryov: this bug is now unassigned12:03
*** resker has quit IRC12:04
*** ChanServ changes topic to "Time to fix bugs for RC1. 16 bugs left"12:04
*** porrua has joined #openstack-manila12:05
bswartzthere is another unassigned bug up for grabs: https://bugs.launchpad.net/manila/+bug/145459012:05
openstackLaunchpad bug 1454590 in Manila "os_region_name defined twice in manila.conf" [Undecided,New]12:05
*** ktolstoy has quit IRC12:06
vponomaryovbswartz: last one can be postponed for Mitaka12:07
bswartzvponomaryov: it can be, but if someone steps up to fix it in the next week that would be better12:08
bswartzI don't like bugs hanging around forever12:08
bswartzvponomaryov: about the earlier one, we need to decide how much more effort to sink into solving it12:09
bswartzwill another day or 2 of research make a big difference?12:09
vponomaryovbswartz: I do not have answer about efforts, but we definitly suffer from it12:10
bswartzvponomaryov: how many days have you already spend trying to fix it?12:11
bswartzvponomaryov: I'm wondering if a fresh set of eyes might have better luck or if you feel like you were getting close to a solution12:11
gansomarkstur: ping12:11
vponomaryovbswartz: fresh eyes can help, making statistics12:12
*** ociuhandu has quit IRC12:13
*** ociuhandu has joined #openstack-manila12:14
vponomaryovganso: you said that scenario tests failed for you even using different driver?12:18
vponomaryovganso: with "connection errors"?12:19
*** haomaiwang has quit IRC12:20
gansovponomaryov: yes, it is the error I posted to you http://paste.openstack.org/show/456622/12:20
gansovponomaryov: the error is trying to SSH into the guest instance, not the share server, guest instances are used for all drivers12:20
*** ktolstoy has joined #openstack-manila12:21
vponomaryovganso: We can try to set "lock" for creating instances in scenario tests, but it will be workaround and real bug will exist12:24
gansovponomaryov: at least it does not block us anymore, like keep rechecking jenkins until scenario passes, or restarting all services in our labs (which I am still testing)12:25
*** anubhav has quit IRC12:26
*** akerr has joined #openstack-manila12:27
*** BharatK has quit IRC12:40
*** dustins has joined #openstack-manila12:43
lpetrutvponomaryov: Hi12:46
vponomaryovlpetrut: Hello12:47
lpetrutvponomaryov: I've seen that at some point, you guys were considering using stable Nova for the CI12:47
openstackgerritCsaba Henk proposed openstack/manila: doc: turn ascii art tables into proper reST grid tables  https://review.openstack.org/22242412:48
bswartzthere is a downside to using stable nova12:49
*** gouthamr has joined #openstack-manila12:49
lpetrutbswartz: could you please tell me what the downside is? We are considering doing this for the Windows CI, lowering the time required to setup our Hyper-V nodes.12:50
bswartzI'm not in favor of using stable branches12:50
bswartzthe downside is that you don't discover breakage until after it's happened12:50
bswartzif nova makes a change that breaks your project, you want to file a bug and fix it immediately12:51
bswartzotherwise users who are running master will have a broken system12:51
bswartzyou can argue that most users don't run master, but the openstack community wants to support it12:52
lpetrutgot it. but would it make sense to do this in case of 3rd party CIs? as the 'upstream' one would get this kind of issues12:52
bswartzlpetrut: it depends whether the breakage is specific to your driver or not12:52
bswartzright now the only part of manila that depends on nova is a few drivers12:53
bswartzincluding generic driver and yours12:53
bswartzif nova breaks in a way that doesn't affect generic driver we won't notice it12:53
lpetrutright. thanks for clearing this out!12:53
bswartzI know it's painful to run on the bleeding edge, but because everyone does it we share the pain and get things fixed quickly12:54
lpetrutour main issue is the time required to prepare the setup and run the tests.12:56
*** vbellur has quit IRC12:56
gansoHello bswartz, u_glide1, could you please take a look at https://review.openstack.org/#/c/219851/ ? Thanks12:57
bswartzganso: this change fixes a bug that's not targeted to RC112:57
bswartzhttps://bugs.launchpad.net/manila/+bug/149155012:57
openstackLaunchpad bug 1491550 in Manila "Manila retry function does not accept random periods" [Undecided,In progress] - Assigned to Alyson (alyson-rodrigues-rosa)12:57
bswartzdoes it need to be targeted?12:58
gansoforgot to target it, sorry12:58
*** ChanServ changes topic to "Time to fix bugs for RC1. 17 bugs left"12:59
lpetrutanyway, our CI should become stable soon. We were a little limited by our Hyper-V nodes resources, so the recent failures are related to not having enough resources to spawn all the Windows share server instances :)13:00
openstackgerritIgor Malinovskiy proposed openstack/manila: Return share_type UUID instead of name in Share API  https://review.openstack.org/22004613:00
*** martyturner has joined #openstack-manila13:01
*** tpsilva has joined #openstack-manila13:05
gansovponomaryov: I attempted to restart all neutron and nova services and run the test again, several times, this workaround has not worked for me13:05
alyson_ping bswartz13:09
bswartzalyson_: hi13:10
*** martyturner has quit IRC13:10
*** ayma has joined #openstack-manila13:11
*** nkrinner has quit IRC13:11
*** eharney has joined #openstack-manila13:14
alyson_bswartz: about your comment in bug fix https://review.openstack.org/#/c/219851/. Will be ok if I just delete that line? https://docs.python.org/2/library/random.html#random.seed13:15
*** erlon has joined #openstack-manila13:15
*** logan2 has quit IRC13:15
bswartzhow will the rng get seeded?13:15
*** nkrinner has joined #openstack-manila13:16
bswartzalyson_: is there a static initializer somewhere you can use to seed it once?13:18
alyson_bswartz: it uses the system time when the module is imported, according to https://docs.python.org/2/library/random.html#random.seed13:18
bswartzoh I see that13:18
bswartzokay yeah then just remove it13:19
alyson_ok!13:19
alyson_thanks13:19
*** bill_az_ has joined #openstack-manila13:24
*** esker has joined #openstack-manila13:27
*** cknight has joined #openstack-manila13:29
*** eharney has quit IRC13:30
*** eharney has joined #openstack-manila13:31
*** timcl has joined #openstack-manila13:32
*** u_glide has joined #openstack-manila13:36
*** logan2 has joined #openstack-manila13:43
*** haomaiwang has joined #openstack-manila13:48
*** ayma has quit IRC13:51
openstackgerritMerged openstack/manila: Enhance base driver checking if a method was implemented  https://review.openstack.org/22051113:51
*** timcl has quit IRC13:54
*** deepakcs has quit IRC14:00
*** haomaiwang has quit IRC14:01
*** haomaiwang has joined #openstack-manila14:01
*** timcl has joined #openstack-manila14:03
*** esker has quit IRC14:03
*** esker has joined #openstack-manila14:04
openstackgerritValeriy Ponomaryov proposed openstack/manila: Add 'snapshot_support' attr to share details  https://review.openstack.org/22123114:04
openstackgerritValeriy Ponomaryov proposed openstack/manila: Add 'snapshot_support' attr to share details  https://review.openstack.org/22123114:07
*** izadorozhna has left #openstack-manila14:09
openstackgerritValeriy Ponomaryov proposed openstack/manila: Add 'snapshot_support' attr to share details  https://review.openstack.org/22123114:11
gouthamrvponomaryov: Requesting clarification on https://review.openstack.org/221231. When we added snapshot_support to the share, shouldn't we add it on the horizon/manila_ui? Why would we pop it out of the share details?14:11
vponomaryovgouthamr: see microversion decorators14:13
vponomaryovgouthamr: microversion 2.2+ will have it14:13
vponomaryovgouthamr: 2.1, 2.0 and 1.0 will now14:14
vponomaryovs/now/not/14:14
openstackgerritIgor Malinovskiy proposed openstack/manila: Return share_type UUID instead of name in Share API  https://review.openstack.org/22004614:14
gouthamrvponomaryov: Ah.. thanks for the clarification. Maybe you could add that to the commit message :)14:15
openstackgerritChuck Fouts proposed openstack/python-manilaclient: Fix client API version to support Manila API  https://review.openstack.org/22262614:16
cknightu_glide, vponomaryov: ^^ This is Chuck's client microversion bump that should work for the L.3 client release.14:18
*** vbellur has joined #openstack-manila14:21
*** baojg has joined #openstack-manila14:27
*** nkrinner has quit IRC14:27
*** baojg has quit IRC14:32
openstackgerritAlyson proposed openstack/manila: Fix get_stats to return real used space in HNAS  https://review.openstack.org/22263814:46
openstackgerritAlyson proposed openstack/manila: Adds retry function to HNAS driver  https://review.openstack.org/21985114:54
*** ayma has joined #openstack-manila14:55
*** xyang1 has joined #openstack-manila14:55
openstackgerritValeriy Ponomaryov proposed openstack/manila-ui: Updated from global requirements  https://review.openstack.org/22265215:00
*** haomaiwang has quit IRC15:01
*** haomaiwang has joined #openstack-manila15:01
bswartzvponomaryov: why doesn't proposal bot do this for us? ^15:15
*** mctaylor has joined #openstack-manila15:22
*** u_glide has quit IRC15:25
openstackgerritClinton Knight proposed openstack/manila: Remove support for 'latest' microversion  https://review.openstack.org/22267715:31
*** vbellur has quit IRC15:37
*** timcl has quit IRC15:39
*** martyturner has joined #openstack-manila15:39
*** ktolstoy has quit IRC15:42
*** ktolstoy has joined #openstack-manila15:46
openstackgerritMerged openstack/manila: Remove HDS SOP driver due to lack of CI  https://review.openstack.org/22231715:51
*** jasonsb has joined #openstack-manila15:57
*** haomaiwang has quit IRC16:01
*** 32NAAC9FD has joined #openstack-manila16:01
*** chlong has joined #openstack-manila16:04
openstackgerritSebastien Badia proposed openstack/puppet-manila: Initial msync run for all Puppet OpenStack modules  https://review.openstack.org/18987216:04
*** devlaps has joined #openstack-manila16:06
*** timcl has joined #openstack-manila16:26
*** vbellur has joined #openstack-manila16:44
*** esker has quit IRC16:55
*** rushil has joined #openstack-manila16:56
*** 32NAAC9FD has quit IRC17:01
*** haomaiwang has joined #openstack-manila17:01
*** ktolstoy has quit IRC17:02
*** adrianofr has joined #openstack-manila17:06
*** rushil has quit IRC17:06
*** rushil has joined #openstack-manila17:07
*** haomaiwang has quit IRC17:16
openstackgerritAlyson proposed openstack/manila: Fix get_stats to return real used space in HNAS  https://review.openstack.org/22263817:16
gansomarkstur: ping17:20
marksturganso, pong17:21
openstackgerritChuck Fouts proposed openstack/python-manilaclient: Fix client API version to support Manila API  https://review.openstack.org/22262617:22
gansomarkstur: Hi Mark. I have re-examining patch https://review.openstack.org/#/c/221916 and noticed that all the changes it contains should be included regardless of https://review.openstack.org/#/c/221880. Even if renaming jobs patch is aborted, they should be merged because they should have been included in the Share Migration patch in the first place.17:22
gansomarkstur: So, if renaming the jobs break anything, it should be fixed regardless of the changes included in the proposed patch.17:22
gansomarkstur: I see no point in delaying those fixes even more if renaming jobs patch break. The same changes, along with more changes, will need to be included, that could be a separate patch, I do not see the current changes being reverted as they apply to the current state of Jenkins.17:24
marksturganso, Yes. It should be good to merge. But if these new tests fail they will be gate-breakers.17:24
marksturganso, It sure would be nice to see them pass at least once before they stop everyone else from passing17:25
gansomarkstur: they will require a new patch, in the same change, or a new one, to fix them. So it is the same.17:25
gansomarkstur: I know there is a medium probability of "no-share-servers" job breaking after that patch. To fix it, I will need to see the error first, although I have an idea of how to fix it, it does not make much sense of submitting the fix earlier. Also, the jobs being renamed are non-voting.17:27
marksturganso, Hmmm.  The old default was true.17:27
gansomarkstur: If you agree to +2 it, I need to remove the "depends-on" from the message.17:27
openstackgerritMerged openstack/manila: Corrects capabilities returned by Quobyte Manila driver  https://review.openstack.org/21555617:27
marksturganso, Yeah. So the no-ss only runs the "empty" test and only scenario is at risk with new behavior.  That would be OK17:31
marksturganso, Would be better if we could get the project-config done17:31
gansomarkstur: in fact, no. The no-ss is the one that has a medium probability of breaking, and I believe scenario one is the one that should pass.17:32
marksturganso, no-ss already passed with this patch.  The 'empty' test.17:32
gansomarkstur: it is skipping it17:33
*** chlong has quit IRC17:34
gansomarkstur: it is *being skipped17:36
*** jasonsb has quit IRC17:36
marksturganso, What do you think about taking out the dependency and adding if/elif/elif so that it works with the current names (or multibackend) and then we'd just have some junk to delete later?17:39
marksturganso, not sure if other reviewers would like that, but it's strange to merge new test that will start affecting our gate at some later time.17:39
*** haomaiwang has joined #openstack-manila17:39
gansomarkstur: I can apply the same effect of the patch it depends on in it to run right now. Like, configure as multibackend if "no-share-servers" or "scenario". This would have the same effect and we will be able to see what happens now. It is dirty, but will work.17:42
gansomarkstur: even better, I can submit a patch that changes exactly that, and abandon it later.17:43
*** jasonsb has joined #openstack-manila17:43
gansomarkstur: what do you think?17:45
*** haomaiwang has quit IRC17:45
marksturganso,  Good idea.  I hate to waste your time on this, but that seems to be better than waiting for project-config merges17:48
gansomarkstur: the earlier we address this, the better.17:49
marksturganso, Yes. I'd be happy to merge your patch if that dependent/temporary one works.17:50
*** bill_az_ has quit IRC17:54
*** tpsilva has quit IRC17:54
*** u_glide has joined #openstack-manila17:54
*** lpetrut has quit IRC17:55
*** ayma1 has joined #openstack-manila18:06
*** ayma has quit IRC18:07
cFoutscknight, vponomaryov https://review.openstack.org/#/c/222626/ <-- Jenkins just verified last patch set18:09
cFoutsvponomaryov: your review of ^^ would be appreciated.18:09
openstackgerritRodrigo Barbieri proposed openstack/manila: TEST CHANGE, DO NOT MERGE!!!  https://review.openstack.org/22272018:09
cFoutsvponomaryov: cknight added a comment regarding needing to move from manilaclient/v1 to manilaclient/v2. Thanks ahead of time18:10
*** porrua has quit IRC18:10
openstackgerritRodrigo Barbieri proposed openstack/manila: Fix Share Migration tempest tests  https://review.openstack.org/22191618:11
*** lpabon has joined #openstack-manila18:13
*** porrua has joined #openstack-manila18:15
cFoutsmarkstur: the docstring in manilaclient/v2/client.py does need to be updated18:21
cFoutsmarkstur: the api_version is determined in manilaclient/shell.py either based on the constant or what is passed in via --os-share-api-version18:23
marksturcFouts, Sure. But also I think out-of-the-box horizon/manila is broken.18:23
cFoutsthe service type that the client uses now is 'sharev2' so that it uses the v2 endpoint18:23
cFoutshorizon should still be using v1 with the 'share' service type18:24
cknightmarkstur: when did you notice the horizon breakage?  are you certain it began with the v2 merge yesterday?  the API server should still be responding to v1, and the devstack scripts should be creating both v1 and v2 endpoints.18:24
marksturcknight, cFouts Actually I think it broke last week or maybe Tues AM18:25
marksturcknight, cFouts That new param api_version is being called with "admin"18:25
cknightmarkstur: OK, then it's probably unrelated to microversions.  Where is this parameter you speak of?18:26
marksturcknight, manilaclient/v1/client.py __init__18:26
cFoutsI think he is talking about the first parameter used in Client. api_versions was added as the first parameter18:26
cFoutsinstead of as a named parameter18:26
*** tpsilva has joined #openstack-manila18:27
cknightmarkstur: i see, so admin should be mapped to username?18:27
marksturyes18:27
marksturcknight, cFouts I'm not sure if these devstacks are valid, but we've seen it twice and it seems like a basic setup hits this problem18:29
cknightmarkstur: yes, we've looked and acknowledge the issue.  chuck will push up a separate fix later today.  I think the best solution is to move api_version to the end of the list and make it a kwarg.18:30
marksturcknight, cFouts OK. Yes. I tried that hack and seems to work. I just wasn't sure if I was out-of-sync on something our new versioning was supposed to solve18:31
cFoutsmarkstur: no, it was me that was out of sync. I didn't consider that another entity was using the Client class18:32
marksturGood. Problem solved then. We should have caught that. Too bad horizon isn't a check/gate thing.18:33
cknightmarkstur: Why is horizon using manilaclient/v1/client instead of manilaclient/client?18:34
cknightmarkstur: If it were, it wouldn't have broken.18:34
*** dzamboni has joined #openstack-manila18:36
cknightmarkstur: In the client patch that's in review now, all the code in manilaclient/v1 is moved to manilaclient/v2.  So clients should come in via the top-level manilaclient/client, and everything will seamlessly transition to /v2.  Can you try that in horizon before we change anything?18:37
*** timcl has quit IRC18:38
cknightmarkstur: This sounds more like an issue with horizon's use of the client.18:38
*** lpetrut has joined #openstack-manila18:41
bswartzneed a core +2 on https://review.openstack.org/#/c/222626/18:48
bswartzsupporting newest microversion will allow us to make a client release18:49
openstackgerritAndrew Kerr proposed openstack/manila: Non-admin user can perform 'extra-specs-list'  https://review.openstack.org/22273818:54
*** esker has joined #openstack-manila18:58
marksturcknight, The manila-ui looks like it has been using v1/Client since Kilo18:58
marksturcknight, Shouldn't v1 stay compatible?18:59
bswartzmarkstur, cknight: that's a good question19:00
bswartzanything that imports the library and uses v1 won't deal well with v1 being renamed19:01
marksturbswartz, Yeah. It doesn't deal well with signature changes either.19:02
*** ayma1 has quit IRC19:03
*** ayma has joined #openstack-manila19:03
marksturbswartz, Moving forward to V2 sounds good, but the breaking of existing things has me confused. Does someone have that figured out?19:03
*** lpabon has quit IRC19:04
cFoutsmarkstur: I've brought it up in discussion but it hasn't really been figured out19:04
bswartzmarkstur: we have 2 goals19:04
cFoutsmarkstur, bswartz the way the client currently loads the extensions requires that we rename the directory to v219:05
bswartzmarkstur: long term we want to support all versions with the client19:05
bswartzmarkstur: however we need to make a python-manilaclient release that's compatible with the new microversioned features asap19:05
bswartzI think the concern about the existing horizon code depending on v1 is something we need to address19:05
bswartzI made a comment in the review19:06
bswartzperhaps we can simply keep both v1 and v2 side by side until we have the proper multi-version support19:06
marksturbswartz, Yes. So we could fix the Liberty Horizon.  But installing the new client would break existing Horizons.19:06
cFoutsbswartz considering that there isn't really a defined version for v1 is a bit of a problem.19:06
cFoutsbswartz: we've had several 1.x and now 2.x microversions during the Liberty release19:07
cFoutsbswartz: are you suggest that I pull manilaclient/v1 from Kilo and hook it into the manila client/19:07
cFouts?19:07
bswartzcFouts: I'm talking about simply leaving the v1 code exactly as it was, and also adding the v2 stuff you've added19:07
cFoutsbswartz: but the problem with that is that the v1 code is not really v1.19:08
cFoutsbswartz: it is Kilo + everything done in Liberty up to this point19:08
bswartzhow much delta is that?19:09
bswartzcould we ressurrect a working v1 implemetnation from the git history and place that alongside v2?19:09
bswartzv1 without microversions I mean19:09
marksturcFouts, I think we'd have at least one fix to make v1 work.  At least splitting v1/v2 should make that possible-ish19:09
bswartzthe v1 stuff would be the only way to talk to kilo and earlier19:09
bswartzcFouts: it might be next week before we get it all perfectly working but I'm still looking for a shortcut to get a working release this week19:10
cFoutsmarkstur: for a quick fix we could have v1 point to the current v219:10
marksturcFouts, We'd at least that __init__ fix for v119:10
cFoutsmarkstur, bswartz : is changing Horizon out of the question?19:11
*** csaba has quit IRC19:11
marksturHorizon should be changed to work with v2, but the client should work with unchanged Horizon19:11
*** ociuhandu has quit IRC19:13
*** rushil has quit IRC19:16
cknightmarkstur: where is the code for the manila-ui horizon plug-in?19:16
*** rushil has joined #openstack-manila19:17
*** rushil has quit IRC19:17
*** u_glide has quit IRC19:18
*** lpabon has joined #openstack-manila19:20
*** rushil has joined #openstack-manila19:22
cFoutsmarkstur: have you looked at https://review.openstack.org/#/c/222626/19:22
cFoutsI talked to cknight and we can move things back to v1 and hardcode the extension discovery to look in the v1 directory instead of using 'v' + api_version.major19:23
cFoutsmarkstur: and then address the v2 and v1 support in the client at a later date19:24
*** csaba has joined #openstack-manila19:24
cFoutsmarkstur: after I make that change and push it upstream would you +2 and workflow 222626 to get the change in?19:25
*** rushil has quit IRC19:25
*** porrua has quit IRC19:26
*** akerr has quit IRC19:26
bswartzcFouts: I think the concern is that anything that drops v1 will immediately break manila-ui19:30
bswartzso we can't merge that even as a temporary change19:31
cFoutsaight19:31
bswartzthe upshot may be that we need to wait for the fully working solution that supports v1 and v2 and all the microversions19:31
cFoutsmaking the change to maintain backwards compatiblity19:31
bswartzand just suffer not having a client users can test with until that time19:31
marksturcknight, The manila-ui code is git clone https://review.openstack.org/openstack/manila-ui19:34
marksturcknight, You are a core reviewer for manila-ui ;)19:34
*** csaba has quit IRC19:34
cknightmarkstur: I know, and I've been focused elsewhere.  I don't know django.19:34
*** csaba has joined #openstack-manila19:35
marksturcknight, understood.19:35
cknightmarkstur: So do you agree that manilaclient/v1/client should be encapsulated behind manilaclient/client?19:35
cknightmarkstur: if horizon were using the top-level client, it wouldn't break with chuck's change.19:35
marksturcknight, I'm not sure. Seems like we are talking about keeping a V1 that doesn't change/break. And adding a V2 that uses microversioning to change19:36
marksturcknight, Not sure what the "encapsulated" idea is.  But if it accomplishes old stuff working while we move forward then great19:37
openstackgerritMerged openstack/manila: Missing check in ShareManager::manage_existing()  https://review.openstack.org/22225719:37
cknightmarkstur: Well, if horizon uses the top-level client, then we have the freedom to move stuff around underneath and nothing breaks.  horizon doesn't need to know we're using v2 underneath.19:38
marksturcknight, I see.19:38
marksturcknight, That would require changing horizon in kilo and somehow not allowing unpatched horizons to run with new clients19:39
cknightmarkstur: I think we can defer fixing horizon and push up a client that just works.19:40
*** csaba has quit IRC19:41
*** csaba has joined #openstack-manila19:41
marksturcknight, Good.  I like "just works".19:41
cknightmarkstur: We didn't realize horizon wasn't using the top-level published client.  That's something we'll need to fix when convenient.19:41
openstackgerritRodrigo Barbieri proposed openstack/manila: TEST CHANGE, DO NOT MERGE!!!  https://review.openstack.org/22272019:42
marksturcknight, Yes. If you could study django this weekend...19:42
cknightmarkstur: "just works" is easy when clients aren't peeking beneath the interface ;-P19:42
*** lpetrut has quit IRC19:43
openstackgerritRodrigo Barbieri proposed openstack/manila: Fix Share Migration tempest tests  https://review.openstack.org/22191619:43
marksturcknight, Yeah. That "v1" in the code doesn't look good.19:44
*** u_glide has joined #openstack-manila19:44
marksturganso, Your test patch isn't dependent on the other one19:45
*** jwang has joined #openstack-manila19:47
marksturganso, Oh. I see. You did it the other way around.19:47
gansomarkstur: :)19:48
gansomarkstur: scenario failed due to v2 update19:48
gansomarkstur: and no-share-servers failed exactly as I expected19:48
gansomarkstur: but that helped me find a bug in Renaming patch, I just submitted a new one19:49
gansomarkstur: I am glad that did not merge19:49
marksturganso, If you make that "TEST CHANGE" one dependent on the real one, then we can look at its tests and you won't have to fix your depends-on to get the first one to merge19:51
*** tpsilva has quit IRC19:52
marksturganso, I mean the first patch can not merge while it is depending on the test one. So the other way around would be better if it is ready to merge.19:53
*** martyturner has quit IRC20:01
*** martyturner has joined #openstack-manila20:02
*** u_glide has quit IRC20:07
*** erlon has quit IRC20:08
*** lpetrut has joined #openstack-manila20:08
*** rushil has joined #openstack-manila20:13
*** u_glide has joined #openstack-manila20:16
*** bill_az has joined #openstack-manila20:22
*** lpabon has quit IRC20:24
*** martyturner has quit IRC20:29
*** cknight has quit IRC20:30
*** martyturner has joined #openstack-manila20:31
*** csaba has quit IRC20:37
*** csaba has joined #openstack-manila20:37
*** lpetrut has quit IRC20:41
*** rushil has quit IRC20:52
*** rushil has joined #openstack-manila20:52
*** u_glide has quit IRC20:53
openstackgerritMerged openstack/manila: Remove support for 'latest' microversion  https://review.openstack.org/22267720:56
openstackgerritMerged openstack/manila: Check the snapshot directory before copy it  https://review.openstack.org/22247920:56
*** cknight has joined #openstack-manila21:00
*** bill_az__ has joined #openstack-manila21:01
*** cknight1 has joined #openstack-manila21:04
*** cknight has quit IRC21:07
openstackgerritValeriy Ponomaryov proposed openstack/manila: Add 'snapshot_support' attr to share details  https://review.openstack.org/22123121:08
*** dustins has quit IRC21:09
gansomarkstur: it worked! :D https://review.openstack.org/#/c/221916/21:10
*** martyturner has quit IRC21:10
marksturganso, woo hoo!21:13
gansomarkstur: I will upload a new patch, I saved the log links here: http://paste.openstack.org/show/457284/21:13
marksturganso, scenario skipped one test_migration_files and passed one test_migration_files21:14
gansomarkstur: the one skipped is CIFS21:14
marksturganso, I see21:14
gansomarkstur: that is not supported at this moment21:14
gansomarkstur: I will just remove the "depends-on" now21:14
marksturganso, Cool. Should be good now.  Of course you happened to not pass Jenkins anyway.  Next time!21:15
gansomarkstur: the TEST CHANGE messed up regular neutron job, it was expected21:15
marksturganso, So you just need to edit the commit message in gerrit now, right?21:16
openstackgerritRodrigo Barbieri proposed openstack/manila: Fix Share Migration tempest tests  https://review.openstack.org/22191621:16
gansoyes ^21:16
*** mctaylor has quit IRC21:17
*** esker has quit IRC21:19
*** ayma has quit IRC21:21
*** lpabon has joined #openstack-manila21:24
*** lpabon has quit IRC21:29
*** cknight1 has quit IRC21:34
*** cknight has joined #openstack-manila21:34
*** gouthamr has quit IRC21:47
*** cknight has quit IRC21:50
*** dzamboni has quit IRC22:02
*** eduardo has quit IRC22:03
*** alyson_ has quit IRC22:10
*** cknight has joined #openstack-manila22:15
*** ChanServ changes topic to "Time to fix bugs for RC1. 14 bugs left"22:15
*** xyang1 has quit IRC22:15
-openstackstatus- NOTICE: 30 minute warning, Gerrit will be offline from 23:00 to 23:30 UTC while some projects are renamed http://lists.openstack.org/pipermail/openstack-dev/2015-September/074235.html22:31
*** cknight has quit IRC22:35
*** esker has joined #openstack-manila22:46
-openstackstatus- NOTICE: Gerrit is offline from 23:00 to 23:30 UTC while some projects are renamed. http://lists.openstack.org/pipermail/openstack-dev/2015-September/074235.html23:02
*** ChanServ changes topic to "Gerrit is offline from 23:00 to 23:30 UTC while some projects are renamed. http://lists.openstack.org/pipermail/openstack-dev/2015-September/074235.html"23:02
*** rushil has quit IRC23:12
*** esker has quit IRC23:20
*** eharney has quit IRC23:20
*** esker has joined #openstack-manila23:27
*** adrianofr has quit IRC23:27
*** bill_az has quit IRC23:34
*** ChanServ changes topic to "Time to fix bugs for RC1. 14 bugs left"23:39
*** bill_az__ has quit IRC23:42
*** openstackgerrit has quit IRC23:46
*** openstackgerrit has joined #openstack-manila23:46
*** rushil has joined #openstack-manila23:48
*** esker has quit IRC23:49
*** cknight has joined #openstack-manila23:59

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!