| *** hoonetorg has quit IRC | 00:09 | |
| openstackgerrit | Dhinesh Balasubramaniam proposed openstack/cinder master: Hedvig Cinder driver implementation https://review.openstack.org/433341 | 00:11 |
|---|---|---|
| openstackgerrit | Dhinesh Balasubramaniam proposed openstack/cinder master: Hedvig Cinder driver implementation https://review.openstack.org/433341 | 00:17 |
| *** hoonetorg has joined #openstack-cinder | 00:22 | |
| *** lixiaoy1 has joined #openstack-cinder | 00:27 | |
| *** Liang__ has joined #openstack-cinder | 00:27 | |
| *** erlon has joined #openstack-cinder | 00:32 | |
| *** brinzhang has joined #openstack-cinder | 00:43 | |
| openstackgerrit | Merged openstack/cinder stable/rocky: Handling unexpected python error "NoneType object is not iterable" https://review.openstack.org/618978 | 00:50 |
| openstackgerrit | Merged openstack/cinder master: doc: Change variable name DEFAULT_API_VERSION to _MIN_API_VERSION https://review.openstack.org/617508 | 00:50 |
| *** bhagyashris has joined #openstack-cinder | 00:57 | |
| *** Liang__ is now known as LiangFang | 01:31 | |
| *** hedvig_01 has quit IRC | 01:46 | |
| *** erlon has quit IRC | 01:54 | |
| *** dhineshb has quit IRC | 02:24 | |
| *** Dinesh_Bhor has joined #openstack-cinder | 02:33 | |
| *** psachin has joined #openstack-cinder | 02:41 | |
| *** mhen has quit IRC | 02:42 | |
| *** mhen has joined #openstack-cinder | 02:45 | |
| openstackgerrit | Brian Rosmaita proposed openstack/cinder master: Correct default policy file https://review.openstack.org/620450 | 03:53 |
| *** chhagarw has joined #openstack-cinder | 03:59 | |
| *** rosmaita has left #openstack-cinder | 04:08 | |
| *** hoangcx has joined #openstack-cinder | 04:11 | |
| *** lpetrut has joined #openstack-cinder | 04:51 | |
| *** udesale has joined #openstack-cinder | 04:54 | |
| *** pcaruana has joined #openstack-cinder | 05:09 | |
| *** lpetrut has quit IRC | 05:28 | |
| *** imacdonn has quit IRC | 05:30 | |
| *** imacdonn has joined #openstack-cinder | 05:30 | |
| *** e0ne has joined #openstack-cinder | 05:37 | |
| *** imacdonn has quit IRC | 05:38 | |
| *** imacdonn has joined #openstack-cinder | 05:38 | |
| *** hedvig_01 has joined #openstack-cinder | 05:39 | |
| openstackgerrit | Yingxin Cheng proposed openstack/os-brick master: Support RSD scenario of nvme connector https://review.openstack.org/620250 | 05:40 |
| *** e0ne has quit IRC | 05:42 | |
| *** e0ne has joined #openstack-cinder | 05:44 | |
| *** e0ne has quit IRC | 05:45 | |
| *** brinzhang has quit IRC | 05:45 | |
| *** brinzhang has joined #openstack-cinder | 05:46 | |
| *** hedvig_01 has quit IRC | 05:47 | |
| *** caixiaoyu has joined #openstack-cinder | 05:53 | |
| *** caixiaoyu_1 has joined #openstack-cinder | 06:15 | |
| *** caixiaoyu has quit IRC | 06:18 | |
| openstackgerrit | Vieri proposed openstack/python-brick-cinderclient-ext master: add python 3.6 unit test job https://review.openstack.org/620468 | 06:31 |
| openstackgerrit | Vieri proposed openstack/python-cinderclient master: add python 3.6 unit test job https://review.openstack.org/620470 | 06:32 |
| openstackgerrit | Yachika_Ralhan proposed openstack/cinder master: Driver QuantaStor added https://review.openstack.org/617503 | 06:35 |
| *** udesale has quit IRC | 06:43 | |
| *** udesale has joined #openstack-cinder | 06:44 | |
| *** udesale has quit IRC | 06:45 | |
| *** udesale has joined #openstack-cinder | 06:46 | |
| *** psachin has quit IRC | 06:47 | |
| *** brinzh has joined #openstack-cinder | 06:52 | |
| *** caixiaoyu_1 has quit IRC | 06:54 | |
| *** brinzhang has quit IRC | 06:56 | |
| *** udesale has quit IRC | 06:57 | |
| *** rcernin has quit IRC | 06:57 | |
| *** Luzi has joined #openstack-cinder | 07:00 | |
| *** gkadam has joined #openstack-cinder | 07:00 | |
| *** slaweq has joined #openstack-cinder | 07:04 | |
| *** dpawlik has joined #openstack-cinder | 07:16 | |
| openstackgerrit | Yachika_Ralhan proposed openstack/cinder master: Driver QuantaStor added https://review.openstack.org/617503 | 07:18 |
| *** dpawlik has quit IRC | 07:20 | |
| openstackgerrit | Rajat Dhasmana proposed openstack/cinder master: DNM: Test jobs running on bionic instead of xenial https://review.openstack.org/620501 | 07:23 |
| *** dpawlik has joined #openstack-cinder | 07:24 | |
| *** caixiaoyu_1 has joined #openstack-cinder | 07:24 | |
| openstackgerrit | Yingxin Cheng proposed openstack/os-brick master: Support RSD scenario of nvme connector https://review.openstack.org/620250 | 07:31 |
| *** caixiaoyu_2 has joined #openstack-cinder | 07:32 | |
| *** dcdamien has joined #openstack-cinder | 07:34 | |
| *** caixiaoyu_1 has quit IRC | 07:36 | |
| *** LiangFang has quit IRC | 07:50 | |
| *** udesale has joined #openstack-cinder | 07:53 | |
| *** e0ne has joined #openstack-cinder | 07:54 | |
| *** udesale has quit IRC | 07:58 | |
| *** udesale has joined #openstack-cinder | 08:11 | |
| *** e0ne has quit IRC | 08:37 | |
| *** lemko has joined #openstack-cinder | 08:38 | |
| openstackgerrit | caixiaoyu proposed openstack/cinder master: Add test case: clone volume with bad size https://review.openstack.org/618522 | 08:41 |
| openstackgerrit | wangwei proposed openstack/cinder master: Fix the problem of the scheduler status https://review.openstack.org/616440 | 08:53 |
| *** lixiaoy1 has quit IRC | 08:57 | |
| *** arne_wiebalck_ has joined #openstack-cinder | 09:08 | |
| *** arne_wiebalck_ has quit IRC | 09:09 | |
| *** arne_wiebalck_ has joined #openstack-cinder | 09:11 | |
| *** dcdamien has quit IRC | 09:20 | |
| openstackgerrit | Rajat Dhasmana proposed openstack/cinder-tempest-plugin master: DNM: Test jobs running on bionic instead of xenial https://review.openstack.org/620542 | 09:22 |
| *** luizbag has joined #openstack-cinder | 09:24 | |
| *** whoami-rajat has joined #openstack-cinder | 09:28 | |
| whoami-rajat | smcginnis: Hi Sean, are all cinder repo's (cinder-tempest-plugin, os-brick ...) jobs running on bionic? | 09:30 |
| smcginnis | whoami-rajat: I think we only have cinder running with a bionic job at this point. | 09:30 |
| smcginnis | whoami-rajat: I don't believe depending on devstack's job changes will do anything for our repos to test. | 09:34 |
| smcginnis | whoami-rajat: whoami-rajat The openstack-python36-jobs template should be added for that. | 09:35 |
| smcginnis | whoami-rajat: That causes openstack-tox-py36 to be run, which is defined to use bionic. | 09:35 |
| smcginnis | http://git.openstack.org/cgit/openstack/cinder/tree/.zuul.yaml | 09:35 |
| smcginnis | http://git.openstack.org/cgit/openstack-infra/openstack-zuul-jobs/tree/zuul.d/jobs.yaml#n391 | 09:35 |
| smcginnis | whoami-rajat: Oh, it does appear all Cinder repos have been updated to include that. So yes, all cinder repos have bionic jobs. | 09:37 |
| *** caixiaoyu_2 has quit IRC | 09:37 | |
| whoami-rajat | smcginnis: so the DNM patches aren't needed anymore? | 09:38 |
| smcginnis | Correct | 09:38 |
| whoami-rajat | smcginnis: ok thanks sean. | 09:38 |
| *** caixiaoyu has joined #openstack-cinder | 09:39 | |
| *** bhagyashris has quit IRC | 09:45 | |
| *** caixiaoyu_1 has joined #openstack-cinder | 09:55 | |
| openstackgerrit | Rajat Dhasmana proposed openstack/python-brick-cinderclient-ext master: DNM: Test jobs running on bionic instead of xenial https://review.openstack.org/620546 | 09:56 |
| *** ganso has joined #openstack-cinder | 09:58 | |
| *** caixiaoyu has quit IRC | 09:59 | |
| *** arne_wiebalck_ has quit IRC | 10:04 | |
| *** dcdamien has joined #openstack-cinder | 10:05 | |
| *** Liang__ has joined #openstack-cinder | 10:10 | |
| *** tobias-urdin has quit IRC | 10:10 | |
| *** e0ne has joined #openstack-cinder | 10:14 | |
| openstackgerrit | Maxime Guyot proposed openstack/cinder master: Add rbd_data_pool https://review.openstack.org/539566 | 10:15 |
| *** tobias-urdin has joined #openstack-cinder | 10:15 | |
| openstackgerrit | Rajat Dhasmana proposed openstack/python-brick-cinderclient-ext master: DNM: Test jobs running on bionic instead of xenial https://review.openstack.org/620546 | 10:16 |
| openstackgerrit | Rajat Dhasmana proposed openstack/cinder-tempest-plugin master: DNM: Test jobs running on bionic instead of xenial https://review.openstack.org/620542 | 10:18 |
| smcginnis | whoami-rajat: We already have jobs running on bionic. Why do we need all those DNM patches? | 10:23 |
| whoami-rajat | smcginnis: we've some jobs that are still running on xenial as can be seen here cinder-tempest-plugin-lvm-lio for cinder-tempest-plugin | 10:26 |
| whoami-rajat | http://logs.openstack.org/64/616564/1/check/cinder-tempest-plugin-lvm-lio/3879dde/job-output.txt.gz#_2018-11-08_14_50_32_599975 | 10:26 |
| whoami-rajat | smcginnis: and brick-cinderclient-dsvm-functional for python-brick-cinderclient-ext | 10:28 |
| whoami-rajat | http://logs.openstack.org/68/620468/1/check/brick-cinderclient-dsvm-functional/301c4d2/job-output.txt.gz#_2018-11-28_06_53_10_592345 | 10:28 |
| smcginnis | whoami-rajat: Yes, but we have jobs that also run on bionic, so there shouldn't be any need to test that. | 10:28 |
| *** erlon has joined #openstack-cinder | 10:29 | |
| smcginnis | And I still don't understand what is expected to be tested by depending on a patch that just changes what jobs are run against devstack patches. | 10:29 |
| whoami-rajat | smcginnis: The devstack and tempest jobs are being migrated to bionic so all modules are being checked for stability with the release. maybe the description is more clear here http://lists.openstack.org/pipermail/openstack-discuss/2018-November/000168.html | 10:33 |
| *** jistr is now known as jistr|mtg | 10:34 | |
| openstackgerrit | Lee Yarwood proposed openstack/cinder stable/rocky: lvm: Avoid premature calls to terminate_connection for muiltiattach vols https://review.openstack.org/618472 | 10:35 |
| openstackgerrit | Lee Yarwood proposed openstack/cinder stable/queens: lvm: Avoid premature calls to terminate_connection for muiltiattach vols https://review.openstack.org/618473 | 10:36 |
| smcginnis | whoami-rajat: Yeah, looks not needed and the comment on one of those patches from Ghanshyam seems to confirm that. | 10:36 |
| *** caixiaoyu_1 has quit IRC | 10:46 | |
| *** Liang__ has quit IRC | 10:48 | |
| *** caixiaoyu_1 has joined #openstack-cinder | 10:58 | |
| *** Dinesh_Bhor has quit IRC | 11:42 | |
| *** dave-mccowan has joined #openstack-cinder | 11:44 | |
| *** tpsilva has joined #openstack-cinder | 11:45 | |
| openstackgerrit | Erlon R. Cruz proposed openstack/cinder-tempest-plugin master: Adds .idea files to .gitignore https://review.openstack.org/620575 | 11:46 |
| openstackgerrit | Erlon R. Cruz proposed openstack/cinder-tempest-plugin master: Fix volume revert to snapshot tests https://review.openstack.org/620576 | 11:46 |
| openstackgerrit | Erlon R. Cruz proposed openstack/cinder-tempest-plugin master: Adds new scenario test for revert to snapshot https://review.openstack.org/620577 | 11:46 |
| *** e0ne has quit IRC | 12:09 | |
| openstackgerrit | caixiaoyu proposed openstack/cinder master: Add test case: extend non-exist volume https://review.openstack.org/620584 | 12:14 |
| gmann | smcginnis: whoami-rajat actually those DNM patches will test few specific jobs which cinder has dependency on other repo etc. for example cinder-rally-task. this job is different than other jobs running on cinder by doing the checkout of openstack/rally-openstack master. So 620501 will comfirm when QA team will migrate their base jobs to bionic, cinder-rally-task will not cause any issue. because there is chance that | 12:20 |
| gmann | openstack/rally-openstack master might have code which does not work on Bionic. so it is better to check cross dependency also to avoid any gate block. | 12:20 |
| gmann | and similarly you can take decision on other repo jobs also if they are cross repo dependency and running on xenial as of now. those DNM patch can give OK to qa team to go ahead and merge the base job on bionic. | 12:21 |
| gmann | if you feel such specific jobs does not make much difference then, we can skip the testing patches | 12:23 |
| *** brinzh has quit IRC | 12:34 | |
| *** takamatsu has quit IRC | 12:37 | |
| *** takamatsu has joined #openstack-cinder | 12:43 | |
| *** caixiaoyu_1 has quit IRC | 12:55 | |
| *** arne_wiebalck_ has joined #openstack-cinder | 12:58 | |
| *** rosmaita has joined #openstack-cinder | 13:02 | |
| *** e0ne has joined #openstack-cinder | 13:05 | |
| *** Liang__ has joined #openstack-cinder | 13:13 | |
| *** abishop has joined #openstack-cinder | 13:21 | |
| *** jistr|mtg is now known as jistr | 13:21 | |
| e0ne | jungleboyj, smcginnis: hi. I've got a question about https://bugs.launchpad.net/cinder/+bug/1802070 | 13:23 |
| openstack | Launchpad bug 1802070 in openstack-ansible "Add resource_filter field in cinder.conf for non-admin user can retrieve volume & snapshot volume" [Undecided,In progress] - Assigned to KimMinsik (for-beatitudo) | 13:23 |
| e0ne | jungleboyj, smcginnis: I suppose, the expected behavior is to reject all filers, isn't it? | 13:24 |
| *** takamatsu has quit IRC | 13:29 | |
| *** takamatsu has joined #openstack-cinder | 13:30 | |
| tobias-urdin | ^ i did hit that as well for rocky, pushed patches to RDO to include the resource_filters.json in /etc/cinder and notified the ubuntu team | 13:31 |
| tobias-urdin | dont think i ever filed a bug thought but i talked about it here on irc | 13:31 |
| *** Dinesh_Bhor has joined #openstack-cinder | 13:43 | |
| openstackgerrit | Liang Fang proposed openstack/cinder master: [WIP] Handle image download and conversion errors https://review.openstack.org/618976 | 13:44 |
| *** takamatsu has quit IRC | 13:44 | |
| *** takamatsu has joined #openstack-cinder | 13:48 | |
| *** eharney has joined #openstack-cinder | 13:56 | |
| *** Dinesh_Bhor has quit IRC | 14:10 | |
| *** arne_wiebalck_ has quit IRC | 14:13 | |
| *** mriedem has joined #openstack-cinder | 14:23 | |
| *** udesale has quit IRC | 14:24 | |
| *** udesale has joined #openstack-cinder | 14:25 | |
| *** MrRon has joined #openstack-cinder | 14:29 | |
| MrRon | Hello there. May I ask a question about Cinder database schema as I'm debugging a problem on my side (I'm still quite new to OpenStack). | 14:31 |
| MrRon | In the End I just want to figure out which file in the git repo of cinder is responsible to create the database and the tables :) | 14:32 |
| MrRon | Is it models.py ? | 14:32 |
| eharney | MrRon: they're created by the numbered files in cinder/db/sqlalchemy/migrate_repo/versions/ | 14:33 |
| MrRon | eharney: Ah ok. So there is no single file which creates the db schema at installation time? I thought the versions dir is used for migrations in case of fixes. | 14:35 |
| eharney | MrRon: not really, 085 does a bulk of the creation of tables etc, then later numbers make changes on top of that -- they all get applied at installation time | 14:36 |
| MrRon | eharney: Thanks a lot! Now I understand better! | 14:40 |
| *** dklyle has quit IRC | 14:47 | |
| openstackgerrit | Brian Rosmaita proposed openstack/cinder master: Correct default policy file https://review.openstack.org/620450 | 14:51 |
| *** MrRon has quit IRC | 14:58 | |
| _hemna | e0ne: did someone delete the resource_filters.json? | 14:58 |
| _hemna | regarding that bug | 14:58 |
| e0ne | _hemna: looks like it was not copied originally | 14:59 |
| *** dustins has joined #openstack-cinder | 15:04 | |
| *** dpawlik has quit IRC | 15:16 | |
| *** hedvig_01 has joined #openstack-cinder | 15:28 | |
| *** gkadam has quit IRC | 15:35 | |
| openstackgerrit | Ivan Kolodyazhny proposed openstack/cinder master: Remove deprecated query_volume_filters config option https://review.openstack.org/620632 | 15:36 |
| *** udesale has quit IRC | 15:43 | |
| *** udesale has joined #openstack-cinder | 15:43 | |
| *** dustins_ has joined #openstack-cinder | 15:48 | |
| *** xyang has joined #openstack-cinder | 15:50 | |
| *** dustins has quit IRC | 15:50 | |
| openstackgerrit | Ivan Kolodyazhny proposed openstack/cinder master: Remove deprecated query_volume_filters config option https://review.openstack.org/620632 | 15:53 |
| *** hedvig_01 has quit IRC | 15:54 | |
| *** Luzi has quit IRC | 16:15 | |
| *** alkhodos has joined #openstack-cinder | 16:16 | |
| *** Liang__ has quit IRC | 16:18 | |
| *** dpawlik has joined #openstack-cinder | 16:24 | |
| alkhodos | Hi all! Can someone help me understand why this patch https://review.openstack.org/#/c/586239/ fails openstack-tox-docs test? | 16:27 |
| smcginnis | alkhodos: Did you take a look at the logs? | 16:28 |
| alkhodos | smcginnis: yes, but I struggle to understand the reason | 16:28 |
| smcginnis | http://logs.openstack.org/39/586239/9/check/openstack-tox-docs/08cafa6/job-output.txt.gz#_2018-11-08_10_24_56_048266 | 16:30 |
| alkhodos | smcginnis: but this file was not touched by the patch... | 16:31 |
| alkhodos | smcginnis: am I missing something obvious? | 16:33 |
| smcginnis | alkhodos: I think it's because you removed the code-block directive in nfs.py. Was there a reason for doing that? | 16:33 |
| *** udesale has quit IRC | 16:34 | |
| alkhodos | smcginnis: you're probably right, I'll try putting it back there, thanks | 16:36 |
| smcginnis | alkhodos: Make sure you run tox -e docs locally. That should catch it. | 16:38 |
| alkhodos | smcginnis: I will | 16:38 |
| *** dpawlik has quit IRC | 16:42 | |
| openstackgerrit | Jon Bernard proposed openstack/cinder master: RBD: add support for multiattach https://review.openstack.org/595827 | 16:54 |
| alkhodos | smcginnis: I've just tried tox -e docs without re-adding the code-block directive and I don't have any errors locally? Is that the right test suite? | 16:55 |
| smcginnis | Hmm, should be. I can try running that locally after the meeting is over. | 16:55 |
| alkhodos | smcginnis: I'd really appreciate that, don't want to spam patchsets before finding the issue | 16:56 |
| *** lbragstad has quit IRC | 17:02 | |
| *** e0ne has quit IRC | 17:02 | |
| *** dpawlik has joined #openstack-cinder | 17:03 | |
| *** pcaruana has quit IRC | 17:03 | |
| *** lbragstad has joined #openstack-cinder | 17:05 | |
| *** dpawlik has quit IRC | 17:07 | |
| openstackgerrit | Helen Walsh proposed openstack/cinder master: PowerMax driver - rebranding VMAX to PowerMax https://review.openstack.org/620659 | 17:09 |
| *** KeithMnemonic has quit IRC | 17:14 | |
| *** jackivanov has quit IRC | 17:28 | |
| openstackgerrit | Helen Walsh proposed openstack/cinder master: PowerMax driver - rebranding VMAX to PowerMax https://review.openstack.org/620659 | 17:37 |
| openstackgerrit | Sean McGinnis proposed openstack/cinder stable/queens: Remove cinder-tox-compliance job https://review.openstack.org/620665 | 17:45 |
| *** abhishekk has joined #openstack-cinder | 17:53 | |
| openstackgerrit | Helen Walsh proposed openstack/cinder master: PowerMax driver - rebranding VMAX to PowerMax https://review.openstack.org/620659 | 17:55 |
| *** hedvig_01 has joined #openstack-cinder | 18:02 | |
| *** mszwed has quit IRC | 18:02 | |
| hedvig_01 | Hi, we got a -1 from zuul and it looks like setup error. I did a recheck last night and I still don't see us in the pipeline. Has it been disabled? https://review.openstack.org/#/c/433341/ | 18:03 |
| smcginnis | hedvig_01: I don't see any rechecks on there. | 18:04 |
| hedvig_01 | I replied recheck as a comment | 18:04 |
| smcginnis | Oh, nevermind, I see it. | 18:04 |
| hedvig_01 | okay | 18:04 |
| hedvig_01 | We are not in the pipeline though. I checked the zuul status link | 18:05 |
| *** dcdamien has quit IRC | 18:05 | |
| smcginnis | I wonder if they had to restart zuul last night or something. Try leaving a recheck comment again. | 18:05 |
| *** rtjure has joined #openstack-cinder | 18:13 | |
| hedvig_01 | okay thank you | 18:14 |
| openstackgerrit | Gorka Eguileor proposed openstack/cinder master: Import the cinderlib code https://review.openstack.org/620669 | 18:17 |
| openstackgerrit | Gorka Eguileor proposed openstack/cinder master: Add cinderlib docs https://review.openstack.org/620670 | 18:17 |
| openstackgerrit | Gorka Eguileor proposed openstack/cinder master: Add cinderlib functional tests https://review.openstack.org/620671 | 18:17 |
| *** geguileo has joined #openstack-cinder | 18:18 | |
| *** abhishekk has quit IRC | 18:18 | |
| *** chhagarw has quit IRC | 18:41 | |
| *** preece has joined #openstack-cinder | 19:03 | |
| *** dklyle has joined #openstack-cinder | 19:03 | |
| *** whoami-rajat has quit IRC | 19:07 | |
| *** e0ne has joined #openstack-cinder | 19:07 | |
| mriedem | so uh, you can't attach a server to a volume in 'error' status right? | 19:08 |
| *** lbragstad has quit IRC | 19:08 | |
| mriedem | but what happens if you update a volume attachment with a connector to a volume in error status? will that be allowed and the volume status will change to 'attaching'? | 19:08 |
| mriedem | hard to tell from the code, but that seems to be what happens in the cinder db api code | 19:09 |
| mriedem | in the volume_attached() method | 19:09 |
| jungleboyj | mriedem: Hmmm, I don't think we would want to allow attachment of a volume in error status regardless. | 19:10 |
| mriedem | well, i'm thinking through the rebuild/reimage flow, | 19:10 |
| jungleboyj | Ok ... | 19:10 |
| mriedem | the server is going to hold an empty volume attachment record to the volume while it's being reimaged so we don't lose the volume | 19:10 |
| mriedem | to another server | 19:10 |
| mriedem | if the reimage fails and the volume goes to error status, | 19:10 |
| mriedem | i was wondering if nova should update the empty attachment to have the host connector again, but i think that will put the volume in 'attaching' status rather than error | 19:11 |
| mriedem | which would be bad probably | 19:11 |
| *** lbragstad has joined #openstack-cinder | 19:12 | |
| *** lbragstad has quit IRC | 19:12 | |
| mriedem | also, | 19:13 |
| *** rosmaita has left #openstack-cinder | 19:14 | |
| *** lbragstad has joined #openstack-cinder | 19:14 | |
| mriedem | it looks like if this happens, the admin fixes the backend and then wants to reset the volume status to 'reserved' using the os-reset_status volume action API, they can't do that | 19:14 |
| *** lbragstad has quit IRC | 19:14 | |
| mriedem | because the 'reserved' status is not in this list https://github.com/openstack/cinder/blob/master/cinder/objects/fields.py#L168 | 19:14 |
| mriedem | but 'reserved' is definitely a valid volume status right? | 19:14 |
| *** pck has quit IRC | 19:14 | |
| *** luizbag has quit IRC | 19:14 | |
| mriedem | it's in the api-ref at least https://developer.openstack.org/api-ref/block-storage/v3/?expanded=reset-a-volume-s-statuses-detail#volumes-volumes | 19:14 |
| mriedem | and this https://developer.openstack.org/api-ref/block-storage/v3/?expanded=reset-a-volume-s-statuses-detail#reserve-volume | 19:14 |
| *** e0ne has quit IRC | 19:15 | |
| *** erlon has quit IRC | 19:15 | |
| *** lbragstad has joined #openstack-cinder | 19:16 | |
| *** lbragstad has quit IRC | 19:16 | |
| jungleboyj | Hmmm, that is interesting that we can't change the status of a reserved volume. Seems like that should be valid. | 19:17 |
| *** e0ne has joined #openstack-cinder | 19:17 | |
| *** e0ne has quit IRC | 19:18 | |
| jungleboyj | But if it is listed in the API it shoudl work. Right? | 19:19 |
| jgriffith | mriedem: I odn't remember *for sure* but it shouldn't toggle the volume status, on the attach-status | 19:20 |
| jgriffith | which is "good" | 19:20 |
| jgriffith | The API shouldn't allow subsequent attach calls based on the status check | 19:20 |
| jgriffith | mriedem: I can test that out, see if reserve/update attachment works when voume status=error, but for some reason I thougt we checked that at the API layer | 19:21 |
| mriedem | jungleboyj: the os-reset_status API reference just says you can update the status, not what the valid values are | 19:23 |
| mriedem | but looking at https://developer.openstack.org/api-ref/block-storage/v3/?expanded=reset-a-volume-s-statuses-detail,reserve-volume-detail#volumes-volumes | 19:23 |
| mriedem | er https://developer.openstack.org/api-ref/block-storage/v3/#volumes-volumes | 19:24 |
| mriedem | i would think any of those are fair game | 19:24 |
| mriedem | jgriffith: maybe attachment_create but i'm not seeing volume status checks in attachment_update | 19:24 |
| jungleboyj | mriedem: That would be my expectation as well but I don't know the API that well. :-( | 19:24 |
| jgriffith | https://github.com/openstack/cinder/blob/master/cinder/volume/api.py#L683 | 19:24 |
| jgriffith | Oh.. hmm | 19:25 |
| mriedem | yeah reserve_volume is os-reserve actoin, which is legacy | 19:25 |
| mriedem | https://github.com/openstack/cinder/blob/master/cinder/volume/api.py#L2069 | 19:25 |
| mriedem | is attachment_create | 19:25 |
| mriedem | but i think, from what i'm looking at in code, that one could update an attachment on an error'ed volume and it will change the volume status to 'attaching' | 19:26 |
| mriedem | https://github.com/openstack/cinder/blob/master/cinder/db/sqlalchemy/api.py#L1581 | 19:26 |
| jgriffith | mriedem: I think you may be correct :( | 19:27 |
| mriedem | get out the party hats people | 19:27 |
| jgriffith | LOL | 19:27 |
| *** dklyle has quit IRC | 19:27 | |
| jgriffith | Nahh, that's for when you're actually wrong about something | 19:27 |
| jgriffith | :) | 19:27 |
| jungleboyj | He he. | 19:27 |
| mriedem | jungleboyj: https://bugs.launchpad.net/cinder/+bug/1805692 | 19:29 |
| openstack | Launchpad bug 1805692 in Cinder "Unable to reset volume status to 'reserved'" [Medium,Triaged] | 19:29 |
| jungleboyj | mriedem: Got it. Thanks. | 19:30 |
| *** erlon has joined #openstack-cinder | 19:30 | |
| *** rosmaita has joined #openstack-cinder | 19:46 | |
| *** erlon has quit IRC | 19:52 | |
| *** eharney has quit IRC | 19:59 | |
| *** eharney has joined #openstack-cinder | 20:11 | |
| *** e0ne has joined #openstack-cinder | 20:19 | |
| *** e0ne has quit IRC | 20:20 | |
| imacdonn | eharney: you around? | 20:23 |
| imacdonn | will ramble a bit, for consumption whenever .... regarding cascading delete, following up from yesterday | 20:26 |
| imacdonn | the issue is that, with some backends (e.g. Oracle ZFSSA), a volume created from a snapshot remains dependent on that snapshot ... in fact, in the ZFSSA case, the backend only actually stores deltas since the snapshot | 20:27 |
| imacdonn | cinder, or at least the unit tests, seem to assume that a volume created from a snapshot becomes an independent entity | 20:27 |
| eharney | volumes created from snapshots have to be independent of the source snapshot/volume | 20:28 |
| smcginnis | The storage I used to maintain works that way too, but I could still mark that snapshot as deleted. | 20:28 |
| smcginnis | It just wouldn't go away until any other dependent volumes were also deleted. | 20:28 |
| eharney | right, which is a fine way to handle that | 20:28 |
| imacdonn | so that's where we hit an impasse | 20:28 |
| imacdonn | hmm, mark as deleted ... where ? | 20:29 |
| *** lbragstad has joined #openstack-cinder | 20:30 | |
| smcginnis | In my case, the backend allowed that to happen. eharney, didn't the ceph driver need to do something special to handle this? | 20:30 |
| eharney | i don't think so? | 20:30 |
| eharney | well, the rbd does do some form for marking either volumes or snaps as deleted when they aren't yet, so they can be cleaned up later | 20:31 |
| eharney | but what i'm saying is that from the Cinder API POV, they are independent | 20:31 |
| smcginnis | That must be what I was thinking of. | 20:31 |
| smcginnis | Very similar to how my old storage worked. | 20:31 |
| eharney | the important part is that the semantics of cinder API calls work as expected, the driver can make that happen however is needed | 20:32 |
| smcginnis | ++ | 20:32 |
| imacdonn | is the expected behaviour documented somewhere? | 20:33 |
| imacdonn | it feels wrong to me, so call something deleted when it really still exists | 20:34 |
| imacdonn | s/so/to/ | 20:34 |
| imacdonn | I guess you're saying that "it's dead to cinder" | 20:35 |
| *** openstackgerrit has quit IRC | 20:36 | |
| eharney | yeah, from Cinder's POV it's deleted | 20:36 |
| imacdonn | I guess maybe I can make the driver check, when asked to delete a volume, if it has any snapshots that have dependent children, and only actually delete it if there are none .... bit concerned about creating orphans .. I'll look into it | 20:38 |
| jungleboyj | imacdonn: It is really a question of whether it is managed by Cinder. If we have deleted it ... no. | 20:39 |
| imacdonn | "orphans" isn't the right word ... leaving behind parents that never get cleaned up | 20:39 |
| *** dcdamien has joined #openstack-cinder | 20:39 | |
| jungleboyj | But your concern is right. | 20:39 |
| eharney | well, the driver needs to make sure they get cleaned up at some point | 20:39 |
| imacdonn | I don't know of any way to trigger "at some point" for a volume that cinder has already forgotten | 20:40 |
| *** e0ne has joined #openstack-cinder | 20:40 | |
| imacdonn | so I have a volume "A", that I snapshot'ed, then created "B" and "C" from the snapshot | 20:41 |
| imacdonn | I then delete "A" ... cinder believes it no longer exists, but the backend has to keep it "for now" | 20:41 |
| smcginnis | ZFFSA doesn't have any way to "delete" the snapshot and have it get cleaned up once those other volumes go away? | 20:41 |
| imacdonn | later, I delete "B".... I suppose I have to look up through it's ancestry to determine that it has a parent volume .. then determine if that parent volume has other children, and if not, delete it ... but only if it's been deleted from cinder already | 20:42 |
| imacdonn | smcginnis: not that I know of | 20:42 |
| eharney | the driver would have to associate A with B somehow to keep track of it, when A was deleted | 20:42 |
| imacdonn | the "keeping track of" is the problem .. how's it supposed to do that? It doesn't have a database.... | 20:43 |
| eharney | the RBD driver does this by renaming the object on the backend to have a name that ends with ".deleted", and then knows it can trash those later when they are in the parent chain of another thing being deleted | 20:45 |
| imacdonn | that may be a possibility | 20:45 |
| *** e0ne has quit IRC | 20:49 | |
| imacdonn | I'll look into it .. thanks | 20:49 |
| alkhodos | imacdonn: also, zfs should be able to promote datasets to get ownership of snapshots. So your clone will become the parent of snapshot and the original volume can be deleted after that. | 20:54 |
| imacdonn | alkhodos: pretty sure that's not possible with zvols (cinder volumes are zvols, not datasets) | 20:56 |
| alkhodos | imacdonn: we do that for zvols in NexentaStor (which is ZFS based) | 20:57 |
| alkhodos | imacdonn: keep it mind that it can become complicated with large number of zvols/snapshots | 20:58 |
| imacdonn | alkhodos: yeah, I can see it being complicated w.r.t. space management, at least ... I'm not sure if that ZFS capability is exposed in the ZFS Storage Appliance ... I'm fairly sure it was not at one point in the past, but maybe it got added | 20:59 |
| alkhodos | imacdonn: yes, we had some difficulties with promote at some point too. Problem with renaming for later deletion is that when you have a large tree of zvols, snapshot, then multiple clones from that snapshot, then snapshots again etc, those zvols and snapshots tend to stay there forever | 21:01 |
| imacdonn | alkhodos: yeah, I was pondering that too | 21:02 |
| alkhodos | imacdonn: so we figured that correct promote usage is the only real option, just sharing ;) | 21:03 |
| imacdonn | alkhodos: will take that into consideration - thanks! | 21:03 |
| alkhodos | smcginnis: did you get a chance to run that tox -e docs test for my patch? Sorry for being annoying... | 21:04 |
| alkhodos | imacdonn: sure, np | 21:04 |
| jungleboyj | alkhodos: What was the patch? | 21:05 |
| *** dustins_ has quit IRC | 21:05 | |
| smcginnis | Sorry, been non-stop meetings. I'll pull that down quick now while someone else has the floor. :) | 21:05 |
| alkhodos | jungleboyj: https://review.openstack.org/#/c/586239/ | 21:05 |
| alkhodos | smcginnisL thanks! sorry for the distraction | 21:05 |
| smcginnis | I don't mind being distracted from this meeting. ;) | 21:06 |
| jungleboyj | alkhodos: Trying. | 21:06 |
| smcginnis | I get the error. | 21:07 |
| alkhodos | smcginnis: running on devstack? | 21:07 |
| smcginnis | ? | 21:07 |
| smcginnis | It's a doc build error. tox -e docs | 21:08 |
| alkhodos | I thought I needed to run tox -e docs from cinder location, no?.. | 21:08 |
| smcginnis | I guess it depends what you mean by cinder location. At the root of the cinder repo, you run the command tox -e docs, just like running the unit tests. | 21:10 |
| alkhodos | yep, that's what I did. Very strange, I'll look into it | 21:10 |
| alkhodos | thanks! | 21:10 |
| imacdonn | 2018-11-08 10:24:56.048266 | ubuntu-xenial | /home/zuul/src/git.openstack.org/openstack/cinder/doc/source/drivers.rst:1251:Unexpected indentation. | 21:19 |
| imacdonn | 2018-11-08 10:24:57.012447 | ubuntu-xenial | ERROR: InvocationError for command '/home/zuul/src/git.openstack.org/openstack/cinder/.tox/docs/bin/sphinx-build -W -b html doc/source doc/build/html' (exited with code 2) | 21:19 |
| imacdonn | alkhodos ^^ | 21:19 |
| *** eharney has quit IRC | 21:20 | |
| jungleboyj | I saw that too. | 21:21 |
| imacdonn | although that file is not modified in the change .. hmmm | 21:21 |
| *** e0ne has joined #openstack-cinder | 21:21 | |
| imacdonn | there's also: Error when trying to get requirement for VCS system Command "git config --get-regexp remote\..*\.url" failed with error code 1 in /home/zuul/src/git.openstack.org/openstack/cinder, falling back to uneditable format,Could not determine repository location of /home/zuul/src/git.openstack.org/openstack/cinder | 21:21 |
| *** e0ne has quit IRC | 21:21 | |
| *** ganso has quit IRC | 21:22 | |
| alkhodos | imacdonn: thanks, Sean gave me an idea of what could cause that, it's just that I didn't get the error locally. Re-running it atm to see if proposed fix helps | 21:22 |
| imacdonn | k, good luck :) | 21:23 |
| alkhodos | imacdonn: thx :) | 21:23 |
| *** mchlumsky has quit IRC | 21:25 | |
| *** e0ne has joined #openstack-cinder | 21:26 | |
| *** e0ne has quit IRC | 21:27 | |
| smcginnis | That file is dynamically generated at doc build time to create the driver list: https://docs.openstack.org/cinder/latest/drivers.html | 21:30 |
| smcginnis | Looking at the generated file, it's not happy with the formatting of the list. | 21:30 |
| alkhodos | smcginnis: yep, just found this out | 21:30 |
| jungleboyj | smcginnis: ++ | 21:30 |
| smcginnis | alkhodos: Yep, adding the code-block directives back fixes the issue. | 21:33 |
| *** lemko has quit IRC | 21:35 | |
| *** lbragstad has quit IRC | 21:37 | |
| *** lbragstad has joined #openstack-cinder | 21:37 | |
| *** rcernin has joined #openstack-cinder | 21:50 | |
| *** openstackgerrit has joined #openstack-cinder | 21:55 | |
| openstackgerrit | Alexey Khodos proposed openstack/cinder master: Refactored NexentaStor5 driver https://review.openstack.org/586239 | 21:55 |
| *** rcernin has quit IRC | 22:03 | |
| *** abishop has quit IRC | 22:11 | |
| *** rcernin has joined #openstack-cinder | 22:11 | |
| *** rcernin has quit IRC | 22:11 | |
| *** rcernin has joined #openstack-cinder | 22:12 | |
| *** tpsilva has quit IRC | 22:22 | |
| mriedem | smcginnis: you might enjoy this https://github.com/openstack/nova/commit/eadbacbda628ecc969a980378faf55bc02f514bf | 22:38 |
| smcginnis | mriedem: Woot! | 22:39 |
| mriedem | it's like i traveled back in time | 22:40 |
| smcginnis | mriedem: So it does really only work with the v3 api now? | 22:41 |
| smcginnis | And I could have removed that check for v2 when I removed v1... | 22:42 |
| *** Reepicheep has quit IRC | 22:45 | |
| mriedem | umm, you could push a devstack change quick to test it, | 22:45 |
| mriedem | configure nova.conf with [cinder]catalog_info=volumev2:cinderv2:publicURL | 22:45 |
| *** slaweq has quit IRC | 22:46 | |
| mriedem | i think it would fail though since we check the version and fail if it's not 3 when constructing our internal cinderclient | 22:47 |
| mriedem | and we default to using 3.0 for all requests | 22:48 |
| *** N3l1x has quit IRC | 23:17 | |
| imacdonn | re cascading delete, just noticed this irony, given earlier mention of rbd... | 23:37 |
| imacdonn | @testtools.skipIf(CONF.volume.storage_protocol == 'ceph', | 23:37 |
| imacdonn | 'Skip because of Bug#1677525') | 23:37 |
| imacdonn | def test_volume_from_snapshot_cascade_delete(self): | 23:37 |
| *** erlon has joined #openstack-cinder | 23:38 | |
| *** slaweq has joined #openstack-cinder | 23:55 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!