| *** stakeda has joined #openstack-cinder | 00:06 | |
| *** salv-orlando has quit IRC | 00:14 | |
| *** wanghao has joined #openstack-cinder | 00:19 | |
| *** wanghao has quit IRC | 00:19 | |
| *** wanghao has joined #openstack-cinder | 00:20 | |
| *** wN has joined #openstack-cinder | 00:22 | |
| *** wN has quit IRC | 00:22 | |
| *** wN has joined #openstack-cinder | 00:22 | |
| *** Apoorva has joined #openstack-cinder | 00:24 | |
| *** wanghao has quit IRC | 00:34 | |
| *** wanghao has joined #openstack-cinder | 00:34 | |
| *** itlinux has quit IRC | 00:37 | |
| *** Apoorva has quit IRC | 00:45 | |
| *** Apoorva has joined #openstack-cinder | 00:46 | |
| *** Apoorva_ has joined #openstack-cinder | 00:48 | |
| *** Apoorva has quit IRC | 00:52 | |
| *** junboli has joined #openstack-cinder | 00:52 | |
| *** itlinux has joined #openstack-cinder | 00:56 | |
| *** itlinux has quit IRC | 01:06 | |
| *** Anticimex has quit IRC | 01:14 | |
| *** salv-orlando has joined #openstack-cinder | 01:15 | |
| *** Anticimex has joined #openstack-cinder | 01:15 | |
| *** Apoorva_ has quit IRC | 01:24 | |
| *** dhinesh has joined #openstack-cinder | 01:27 | |
| *** itlinux has joined #openstack-cinder | 01:30 | |
| *** abishop has quit IRC | 01:30 | |
| *** moshele has joined #openstack-cinder | 01:34 | |
| *** moshele has quit IRC | 01:39 | |
| *** dhinesh has quit IRC | 01:40 | |
| *** dhinesh has joined #openstack-cinder | 01:41 | |
| *** lbragstad has joined #openstack-cinder | 01:45 | |
| *** salv-orlando has quit IRC | 01:49 | |
| *** zhaochao has joined #openstack-cinder | 01:49 | |
| *** zengyingzhe has quit IRC | 01:50 | |
| *** gcb has joined #openstack-cinder | 01:52 | |
| tommylikehu | winston-d: pong, sure! will ping you when I upload the patch:) | 02:02 |
|---|---|---|
| *** moshele has joined #openstack-cinder | 02:04 | |
| *** armax has joined #openstack-cinder | 02:07 | |
| *** zengyingzhe has joined #openstack-cinder | 02:10 | |
| *** dalgaaf has quit IRC | 02:33 | |
| *** dalgaaf has joined #openstack-cinder | 02:33 | |
| *** armax has quit IRC | 02:42 | |
| openstackgerrit | Xiaojun Liao proposed openstack/cinder master: Fix v3 api-ref for restoring backup are wrongly described https://review.openstack.org/520375 | 02:44 |
| *** salv-orlando has joined #openstack-cinder | 02:46 | |
| *** moshele has quit IRC | 03:18 | |
| *** akekane has joined #openstack-cinder | 03:29 | |
| *** akekane is now known as abhishekk | 03:30 | |
| *** dave-mccowan has quit IRC | 03:39 | |
| *** armax has joined #openstack-cinder | 03:48 | |
| *** rcernin has quit IRC | 03:50 | |
| *** rcernin has joined #openstack-cinder | 03:52 | |
| *** psachin has joined #openstack-cinder | 04:03 | |
| *** links has joined #openstack-cinder | 04:13 | |
| *** abhitechie has joined #openstack-cinder | 04:22 | |
| *** yangyapeng has joined #openstack-cinder | 04:26 | |
| *** gmann is now known as gmann_lunch | 04:35 | |
| *** pgadiya has joined #openstack-cinder | 05:13 | |
| *** pgadiya has quit IRC | 05:13 | |
| *** sticker has quit IRC | 05:35 | |
| *** eharney has quit IRC | 05:35 | |
| *** rcernin has quit IRC | 05:38 | |
| *** moshele has joined #openstack-cinder | 05:51 | |
| openstackgerrit | Xiaojun Liao proposed openstack/cinder master: Fix v3 api-ref for restoring backup are wrongly described https://review.openstack.org/520375 | 05:51 |
| *** kwathore_ has joined #openstack-cinder | 05:53 | |
| *** kwathore has quit IRC | 05:53 | |
| *** maestropandy has joined #openstack-cinder | 05:56 | |
| *** threestrands has quit IRC | 06:03 | |
| *** wanghao_ has joined #openstack-cinder | 06:13 | |
| *** wanghao has quit IRC | 06:17 | |
| *** eharney has joined #openstack-cinder | 06:19 | |
| *** gouthamr has quit IRC | 06:20 | |
| *** msno has joined #openstack-cinder | 06:23 | |
| msno | can someone help me in https://ask.openstack.org/en/question/111203/cinder-backup-create-failing-in-openstack-liberty/?answer=111211#post-id-111211 | 06:24 |
| *** lpetrut has joined #openstack-cinder | 06:27 | |
| *** youzp has joined #openstack-cinder | 06:37 | |
| *** salv-orlando has quit IRC | 06:37 | |
| *** salv-orlando has joined #openstack-cinder | 06:38 | |
| *** hoonetorg has quit IRC | 06:41 | |
| *** hoonetorg has joined #openstack-cinder | 06:54 | |
| *** e0ne has joined #openstack-cinder | 06:55 | |
| *** gouthamr has joined #openstack-cinder | 06:55 | |
| *** dhinesh has quit IRC | 07:00 | |
| *** maestropandy has quit IRC | 07:01 | |
| openstackgerrit | TommyLike proposed openstack/cinder master: Handle quota in 2 phases when managing resource. https://review.openstack.org/521044 | 07:08 |
| *** lpetrut has quit IRC | 07:08 | |
| *** lpetrut has joined #openstack-cinder | 07:08 | |
| openstackgerrit | TommyLike proposed openstack/cinder master: Handle quota in 2 phases when managing resource https://review.openstack.org/521044 | 07:10 |
| *** lpetrut has quit IRC | 07:14 | |
| *** itlinux has quit IRC | 07:15 | |
| *** youzp has quit IRC | 07:21 | |
| *** youzp has joined #openstack-cinder | 07:22 | |
| *** lpetrut has joined #openstack-cinder | 07:22 | |
| *** gcb has quit IRC | 07:25 | |
| *** gcb has joined #openstack-cinder | 07:26 | |
| *** abhitechie has quit IRC | 07:32 | |
| *** abhitechie has joined #openstack-cinder | 07:32 | |
| *** lkuchlan has joined #openstack-cinder | 07:34 | |
| *** nsingh has quit IRC | 07:35 | |
| *** nsingh has joined #openstack-cinder | 07:35 | |
| *** abhitechie has quit IRC | 07:37 | |
| *** abhishekk is now known as abhishekk|lunch | 07:37 | |
| *** belmoreira has joined #openstack-cinder | 07:38 | |
| *** AlexeyAbashkin has joined #openstack-cinder | 07:47 | |
| *** maestropandy has joined #openstack-cinder | 07:48 | |
| *** e0ne has quit IRC | 07:48 | |
| *** youzp_ has joined #openstack-cinder | 07:49 | |
| *** youzp_ has quit IRC | 07:50 | |
| *** youzp has quit IRC | 07:50 | |
| *** youzp has joined #openstack-cinder | 07:50 | |
| *** lpetrut has quit IRC | 07:53 | |
| *** maestropandy has quit IRC | 07:58 | |
| *** haobing has joined #openstack-cinder | 07:59 | |
| *** dhinesh has joined #openstack-cinder | 08:01 | |
| *** youzp_ has joined #openstack-cinder | 08:03 | |
| *** pcaruana has joined #openstack-cinder | 08:04 | |
| *** niraj_singh has joined #openstack-cinder | 08:05 | |
| *** nsingh has quit IRC | 08:05 | |
| *** dhinesh has quit IRC | 08:06 | |
| *** youzp has quit IRC | 08:08 | |
| *** kwathore_ has quit IRC | 08:08 | |
| *** josecastroleon has joined #openstack-cinder | 08:08 | |
| *** kwathore has joined #openstack-cinder | 08:08 | |
| *** youzp_ has quit IRC | 08:10 | |
| *** alexchadin has joined #openstack-cinder | 08:11 | |
| *** alexchadin has quit IRC | 08:16 | |
| *** alexchadin has joined #openstack-cinder | 08:17 | |
| *** alexchadin has quit IRC | 08:19 | |
| *** danpawlik has quit IRC | 08:19 | |
| *** alexchadin has joined #openstack-cinder | 08:19 | |
| *** gouthamr has quit IRC | 08:25 | |
| *** gkadam has joined #openstack-cinder | 08:26 | |
| *** ircuser-1 has joined #openstack-cinder | 08:28 | |
| *** amoralej|off is now known as amoralej | 08:38 | |
| *** abhishekk|lunch is now known as abhishekk | 08:39 | |
| *** maestropandy has joined #openstack-cinder | 08:39 | |
| *** salv-orlando has quit IRC | 08:42 | |
| *** gmann_lunch is now known as gmann | 08:42 | |
| *** salv-orlando has joined #openstack-cinder | 08:43 | |
| *** salv-orlando has quit IRC | 08:47 | |
| openstackgerrit | TommyLike proposed openstack/cinder master: Try Fetching value from 'cluster' when raise exception https://review.openstack.org/521459 | 08:48 |
| openstackgerrit | Merged openstack/cinder master: Remove setting of version/release from releasenotes https://review.openstack.org/520638 | 08:53 |
| *** belmoreira has quit IRC | 08:58 | |
| *** salv-orlando has joined #openstack-cinder | 08:58 | |
| *** e0ne has joined #openstack-cinder | 09:05 | |
| *** spectr has joined #openstack-cinder | 09:07 | |
| *** spectr has quit IRC | 09:07 | |
| *** spectr has joined #openstack-cinder | 09:11 | |
| *** spectr has quit IRC | 09:12 | |
| *** hoonetorg has quit IRC | 09:13 | |
| *** wanghao_ has quit IRC | 09:14 | |
| openstackgerrit | Merged openstack/cinder master: NetApp E-series: Fix provisioned_capacity_gb https://review.openstack.org/518406 | 09:17 |
| *** danpawlik has joined #openstack-cinder | 09:18 | |
| *** wanghao has joined #openstack-cinder | 09:24 | |
| *** wanghao_ has joined #openstack-cinder | 09:25 | |
| *** wanghao_ has quit IRC | 09:28 | |
| *** wanghao_ has joined #openstack-cinder | 09:28 | |
| *** wanghao_ has quit IRC | 09:28 | |
| *** wanghao has quit IRC | 09:28 | |
| *** wanghao has joined #openstack-cinder | 09:29 | |
| *** wanghao has quit IRC | 09:29 | |
| *** wanghao has joined #openstack-cinder | 09:29 | |
| *** wanghao has quit IRC | 09:30 | |
| *** wanghao has joined #openstack-cinder | 09:30 | |
| *** wanghao has quit IRC | 09:31 | |
| *** wanghao has joined #openstack-cinder | 09:31 | |
| *** lkuchlan has quit IRC | 09:32 | |
| *** wanghao has quit IRC | 09:32 | |
| *** wanghao has joined #openstack-cinder | 09:32 | |
| *** wanghao has quit IRC | 09:33 | |
| *** wanghao has joined #openstack-cinder | 09:33 | |
| openstackgerrit | Merged openstack/os-brick master: Remove setting of version/release from releasenotes https://review.openstack.org/520647 | 09:33 |
| *** wanghao has quit IRC | 09:33 | |
| *** wanghao has joined #openstack-cinder | 09:34 | |
| *** junboli has quit IRC | 09:34 | |
| *** wanghao has quit IRC | 09:34 | |
| *** gmann is now known as gmann_afk | 09:37 | |
| *** lkuchlan has joined #openstack-cinder | 09:44 | |
| *** maestropandy has quit IRC | 09:44 | |
| *** maestropandy has joined #openstack-cinder | 09:46 | |
| *** lpetrut has joined #openstack-cinder | 09:50 | |
| *** e0ne has quit IRC | 09:52 | |
| *** belmoreira has joined #openstack-cinder | 09:52 | |
| *** lkuchlan_ has joined #openstack-cinder | 09:53 | |
| *** e0ne has joined #openstack-cinder | 09:54 | |
| *** lkuchlan has quit IRC | 09:56 | |
| *** lpetrut has quit IRC | 09:56 | |
| *** kwathore has quit IRC | 10:05 | |
| *** maestropandy has quit IRC | 10:08 | |
| *** lpetrut has joined #openstack-cinder | 10:13 | |
| openstackgerrit | Merged openstack/cinder master: SMBFS: Enable reverting snapshots https://review.openstack.org/512545 | 10:14 |
| *** kwathore has joined #openstack-cinder | 10:15 | |
| *** maestropandy has joined #openstack-cinder | 10:37 | |
| *** maestropandy has quit IRC | 10:47 | |
| openstackgerrit | Lucian Petrut proposed openstack/cinder master: SMBFS: manageable volumes https://review.openstack.org/521158 | 10:50 |
| *** abhishekk has quit IRC | 10:53 | |
| openstackgerrit | lihaijing proposed openstack/cinder master: Add unmanage_snapshot method for RBD driver https://review.openstack.org/521486 | 11:00 |
| *** lkuchlan_ has quit IRC | 11:10 | |
| *** moshele has quit IRC | 11:13 | |
| *** haobing has quit IRC | 11:15 | |
| *** haobing has joined #openstack-cinder | 11:16 | |
| *** belmoreira has quit IRC | 11:21 | |
| *** gcb has quit IRC | 11:24 | |
| *** alexchadin has quit IRC | 11:28 | |
| openstackgerrit | Ciara Stacke proposed openstack/cinder master: VMAX driver - Fix error handling and checks for generic volume groups https://review.openstack.org/521080 | 11:30 |
| *** openstackgerrit has quit IRC | 11:32 | |
| *** maestropandy has joined #openstack-cinder | 11:38 | |
| *** maestropandy has left #openstack-cinder | 12:00 | |
| *** moshele has joined #openstack-cinder | 12:00 | |
| *** belmoreira has joined #openstack-cinder | 12:06 | |
| *** lkuchlan_ has joined #openstack-cinder | 12:24 | |
| *** sdague has joined #openstack-cinder | 12:28 | |
| *** links has quit IRC | 12:31 | |
| *** takedakn has joined #openstack-cinder | 12:50 | |
| *** haobing has quit IRC | 12:52 | |
| *** josecastroleon has quit IRC | 12:59 | |
| *** salv-orlando has quit IRC | 12:59 | |
| *** gcb has joined #openstack-cinder | 12:59 | |
| *** salv-orlando has joined #openstack-cinder | 13:00 | |
| *** salv-orlando has quit IRC | 13:04 | |
| *** abishop has joined #openstack-cinder | 13:07 | |
| *** alexchadin has joined #openstack-cinder | 13:13 | |
| *** openstackgerrit has joined #openstack-cinder | 13:15 | |
| openstackgerrit | zhangchangfeng proposed openstack/cinder master: [WIP] Fix wrong quota deduction when concurrent deletion for volumes which host is None. https://review.openstack.org/521364 | 13:15 |
| *** lhx_ has joined #openstack-cinder | 13:15 | |
| *** josecastroleon has joined #openstack-cinder | 13:16 | |
| *** links has joined #openstack-cinder | 13:18 | |
| *** liverpooler has joined #openstack-cinder | 13:22 | |
| *** markvoelker has quit IRC | 13:23 | |
| *** markvoelker has joined #openstack-cinder | 13:24 | |
| *** gnufied has joined #openstack-cinder | 13:24 | |
| *** dhinesh has joined #openstack-cinder | 13:28 | |
| *** dpawlik has joined #openstack-cinder | 13:29 | |
| *** danpawlik has quit IRC | 13:30 | |
| *** takedakn has quit IRC | 13:31 | |
| *** dhinesh has quit IRC | 13:33 | |
| *** dave-mccowan has joined #openstack-cinder | 13:36 | |
| *** dave-mcc_ has joined #openstack-cinder | 13:46 | |
| *** dave-mccowan has quit IRC | 13:48 | |
| *** josecastroleon has quit IRC | 13:56 | |
| *** yangyapeng has quit IRC | 13:59 | |
| *** salv-orlando has joined #openstack-cinder | 13:59 | |
| *** eharney has quit IRC | 14:00 | |
| *** lkuchlan_ has quit IRC | 14:04 | |
| *** ik__ has joined #openstack-cinder | 14:06 | |
| ik__ | How do we generally debug cinder services? | 14:06 |
| *** schmots has joined #openstack-cinder | 14:06 | |
| ik__ | I was trying to debug through Pycharm's remote debugger, but it fails to start. Not sure what I'm missing. | 14:07 |
| smcginnis | ik__: You can use pdb. Or print statements can go a long way. | 14:07 |
| smcginnis | ik__: I haven't used the pycharm remote debugging myself, but I know others have successfully. | 14:08 |
| ik__ | smcginnis: Hmm. Old school until others offer their help. | 14:08 |
| smcginnis | ;) | 14:09 |
| ik__ | smcginnis: Any idea if they had to do any changes with the way service starts? :) | 14:09 |
| e0ne | ik__: I didn't use pycharm. rpd/rpdb works good for me | 14:09 |
| ik__ | e0ne: rpdb, remote pdb? which ide do you use at the other end? any project configuration I can quickly put to use? | 14:13 |
| openstackgerrit | Lucian Petrut proposed openstack/cinder master: SMBFS: manageable volumes https://review.openstack.org/521158 | 14:13 |
| e0ne | ik__: it's like pdb, but works for (remote) services too. I use vim and submlime only | 14:13 |
| *** mriedem has joined #openstack-cinder | 14:14 | |
| *** josecastroleon has joined #openstack-cinder | 14:14 | |
| ik__ | e0ne: Ah, okay. I've sublime too. Let me checkout rpdb, but the question is still open if anyone can help with remote debugging through pycharm. | 14:15 |
| *** liverpoo1er has joined #openstack-cinder | 14:16 | |
| *** liverpoo1er has quit IRC | 14:16 | |
| *** shvepsy_ has joined #openstack-cinder | 14:17 | |
| *** bkopilov has quit IRC | 14:19 | |
| *** shvepsy has quit IRC | 14:19 | |
| *** bkopilov has joined #openstack-cinder | 14:22 | |
| *** yangyapeng has joined #openstack-cinder | 14:24 | |
| *** yangyapeng has quit IRC | 14:29 | |
| *** gouthamr has joined #openstack-cinder | 14:32 | |
| *** josecastroleon has quit IRC | 14:32 | |
| *** alexchadin has quit IRC | 14:32 | |
| *** belmoreira has quit IRC | 14:34 | |
| *** dhinesh has joined #openstack-cinder | 14:35 | |
| *** ik__ has quit IRC | 14:41 | |
| *** yangyapeng has joined #openstack-cinder | 14:42 | |
| *** yangyapeng has quit IRC | 14:46 | |
| *** dhinesh has quit IRC | 14:48 | |
| *** bkopilov has quit IRC | 14:54 | |
| *** eharney has joined #openstack-cinder | 14:55 | |
| *** KeithMnemonic has joined #openstack-cinder | 14:57 | |
| openstackgerrit | Sean McGinnis proposed openstack/cinder master: Suppress UT log messages https://review.openstack.org/521555 | 14:59 |
| *** lkuchlan_ has joined #openstack-cinder | 15:00 | |
| *** links has quit IRC | 15:01 | |
| *** yangyapeng has joined #openstack-cinder | 15:04 | |
| *** dave-mcc_ is now known as dave-mccowan | 15:06 | |
| *** gcb has quit IRC | 15:08 | |
| *** yangyapeng has quit IRC | 15:08 | |
| openstackgerrit | TommyLike proposed openstack/cinder master: Try Fetching value from 'cluster' when raise exception https://review.openstack.org/521459 | 15:09 |
| *** links has joined #openstack-cinder | 15:13 | |
| openstackgerrit | Sean McGinnis proposed openstack/cinder master: Suppress UT log messages https://review.openstack.org/521555 | 15:16 |
| *** moshele has quit IRC | 15:18 | |
| *** yangyapeng has joined #openstack-cinder | 15:24 | |
| *** abhitechie has joined #openstack-cinder | 15:26 | |
| *** links has quit IRC | 15:27 | |
| *** yangyapeng has quit IRC | 15:28 | |
| *** edmondsw has joined #openstack-cinder | 15:47 | |
| *** schmots_ has joined #openstack-cinder | 15:48 | |
| *** dhinesh has joined #openstack-cinder | 15:49 | |
| *** schmots has quit IRC | 15:50 | |
| *** schmots_ is now known as schmots | 15:50 | |
| *** mvk has quit IRC | 15:51 | |
| *** dhinesh has quit IRC | 15:54 | |
| *** hoonetorg has joined #openstack-cinder | 15:59 | |
| *** itlinux has joined #openstack-cinder | 15:59 | |
| *** jmlowe has joined #openstack-cinder | 16:00 | |
| smcginnis | mtreinish: That seems very odd to me that you don't get the failure. | 16:09 |
| smcginnis | mtreinish: jgriffith and I both were seeing the failure due to large amounts of log output during the test. | 16:09 |
| smcginnis | mtreinish: And removing that output would then get it to pass. | 16:09 |
| mtreinish | smcginnis: I got it to fail by looping the test with --until-failure | 16:11 |
| mtreinish | but it wasn't a 100% thing | 16:11 |
| smcginnis | mtreinish: Mine was 100% on that. | 16:11 |
| smcginnis | mtreinish: How much ram do you have? | 16:11 |
| mtreinish | in my laptop 16GB | 16:12 |
| smcginnis | OK, me too. Was just a hunch. | 16:12 |
| mtreinish | it could be because I'm using py36 instead of 35 (I don't have 35 installed locally) | 16:13 |
| smcginnis | mtreinish: And wouldn't you know, now I'm not getting the output and it's passing for me. | 16:15 |
| smcginnis | mtreinish: But that is really odd, since I should be getting some output from those log.errors. | 16:15 |
| *** josecastroleon has joined #openstack-cinder | 16:23 | |
| mtreinish | smcginnis: I'm on my desktop now and running it with --until-failure and can't get it to fail at all (it's at 110 iterations already) | 16:23 |
| smcginnis | mtreinish: Probably no point in looping until failure. It should fail immediately. | 16:24 |
| smcginnis | mtreinish: Trying to understand now why it's no longer capturing that log output. | 16:24 |
| *** AlexeyAbashkin has quit IRC | 16:25 | |
| mtreinish | this bug is evil... | 16:25 |
| smcginnis | We've seen some wonkiness with self.flags(fatal_exception_format_errors=False) | 16:25 |
| smcginnis | It appears that maybe it's actually working now, where it wasn't last week. Maybe? | 16:25 |
| *** lkuchlan_ has quit IRC | 16:26 | |
| smcginnis | No, that's not it... | 16:26 |
| eharney | smcginnis: i can't get your test case on 521201 to produce a failure on my machine with the single test | 16:28 |
| *** josecastroleon has quit IRC | 16:28 | |
| smcginnis | eharney: I can't either now. And I'm not getting the log output that the test should be generating (which is what causes the failure). | 16:29 |
| smcginnis | So somehow now, LOG.exception is getting magically suppressed, so it doesn't accumulate the test log output and doesn't fail. | 16:30 |
| eharney | smcginnis: it just failed for me running ExceptionTestCase instead of running it with just that one test name | 16:34 |
| smcginnis | eharney: Did you get a lot of captured log output from missing formatting args? | 16:35 |
| *** pcaruana has quit IRC | 16:35 | |
| *** josecastroleon has joined #openstack-cinder | 16:36 | |
| *** moshele has joined #openstack-cinder | 16:36 | |
| *** salv-orlando has quit IRC | 16:38 | |
| *** salv-orlando has joined #openstack-cinder | 16:39 | |
| *** moshele has quit IRC | 16:41 | |
| *** psachin has quit IRC | 16:43 | |
| *** salv-orlando has quit IRC | 16:43 | |
| *** Apoorva has joined #openstack-cinder | 16:44 | |
| openstackgerrit | Lucian Petrut proposed openstack/cinder master: SMBFS: manageable volumes https://review.openstack.org/521158 | 16:48 |
| *** Apoorva has quit IRC | 16:51 | |
| *** ntpttr_laptop has joined #openstack-cinder | 16:51 | |
| *** itlinux has quit IRC | 16:53 | |
| *** Apoorva has joined #openstack-cinder | 16:53 | |
| *** itlinux has joined #openstack-cinder | 16:54 | |
| *** armax has quit IRC | 16:59 | |
| *** Apoorva has quit IRC | 17:15 | |
| *** hemna_ has quit IRC | 17:17 | |
| *** gkadam has quit IRC | 17:17 | |
| *** hemna_ has joined #openstack-cinder | 17:17 | |
| *** lhx_ has quit IRC | 17:19 | |
| *** lpetrut has quit IRC | 17:23 | |
| *** dhinesh has joined #openstack-cinder | 17:28 | |
| eharney | smcginnis: yes | 17:33 |
| *** armax has joined #openstack-cinder | 17:36 | |
| *** edmondsw has quit IRC | 17:37 | |
| *** edmondsw has joined #openstack-cinder | 17:38 | |
| eharney | tommylikehu: do you recall the reason for this change? https://review.openstack.org/#/c/506976/8/cinder/test.py | 17:39 |
| *** bkopilov has joined #openstack-cinder | 17:39 | |
| *** salv-orlando has joined #openstack-cinder | 17:39 | |
| eharney | looks odd to no longer be setting cleanup for overridden conf values... which is used for fatal_exception_format_errors ^ | 17:39 |
| eharney | smcginnis: that may explain at least part of the fatal_exception_format_errors wonkiness | 17:41 |
| smcginnis | eharney: Yeah, didn't notice at the time reviewing that patch, but that doesn't seem right. | 17:41 |
| *** edmondsw has quit IRC | 17:42 | |
| *** moshele has joined #openstack-cinder | 17:42 | |
| *** ntpttr_laptop has quit IRC | 17:42 | |
| *** ntpttr_laptop__ has joined #openstack-cinder | 17:42 | |
| smcginnis | eharney: Should be easy enough to change that back to self.override_config | 17:42 |
| *** yangyapeng has joined #openstack-cinder | 17:43 | |
| *** Apoorva has joined #openstack-cinder | 17:44 | |
| *** salv-orlando has quit IRC | 17:45 | |
| *** Apoorva has quit IRC | 17:45 | |
| *** yangyapeng has quit IRC | 17:50 | |
| *** Apoorva has joined #openstack-cinder | 17:50 | |
| *** msno has quit IRC | 17:51 | |
| *** ntpttr_laptop__ has quit IRC | 17:52 | |
| *** Apoorva_ has joined #openstack-cinder | 17:59 | |
| *** pcaruana has joined #openstack-cinder | 18:00 | |
| openstackgerrit | Goutham Pacha Ravi proposed openstack/cinder master: NetApp ONTAP: Fix delayed snapshot deletion call https://review.openstack.org/521616 | 18:02 |
| *** Apoorva has quit IRC | 18:02 | |
| openstackgerrit | Eric Harney proposed openstack/cinder master: Fix TestCase.flags() https://review.openstack.org/521619 | 18:07 |
| *** yangyapeng has joined #openstack-cinder | 18:12 | |
| *** moshele has quit IRC | 18:14 | |
| *** yangyapeng has quit IRC | 18:17 | |
| *** lpetrut has joined #openstack-cinder | 18:29 | |
| *** yangyapeng has joined #openstack-cinder | 18:31 | |
| *** yangyapeng has quit IRC | 18:34 | |
| *** Apoorva_ has quit IRC | 18:35 | |
| *** Apoorva has joined #openstack-cinder | 18:36 | |
| *** e0ne has quit IRC | 18:36 | |
| *** salv-orlando has joined #openstack-cinder | 18:40 | |
| *** salv-orlando has quit IRC | 18:42 | |
| *** salv-orlando has joined #openstack-cinder | 18:42 | |
| *** imacdonn has joined #openstack-cinder | 18:45 | |
| *** liverpooler has quit IRC | 18:45 | |
| *** lpetrut has quit IRC | 18:51 | |
| *** lpetrut has joined #openstack-cinder | 18:55 | |
| *** yangyapeng has joined #openstack-cinder | 18:57 | |
| *** lkuchlan_ has joined #openstack-cinder | 19:00 | |
| *** lpetrut has quit IRC | 19:01 | |
| *** yangyapeng has quit IRC | 19:01 | |
| *** lpetrut has joined #openstack-cinder | 19:01 | |
| *** itlinux has quit IRC | 19:03 | |
| *** yangyapeng has joined #openstack-cinder | 19:15 | |
| *** yangyapeng has quit IRC | 19:20 | |
| *** schmots_ has joined #openstack-cinder | 19:23 | |
| *** schmots has quit IRC | 19:23 | |
| *** schmots_ is now known as schmots | 19:23 | |
| *** lkuchlan_ has quit IRC | 19:28 | |
| *** itlinux has joined #openstack-cinder | 19:31 | |
| *** schmots_ has joined #openstack-cinder | 19:32 | |
| *** schmots has quit IRC | 19:33 | |
| *** schmots_ is now known as schmots | 19:33 | |
| *** lkuchlan_ has joined #openstack-cinder | 19:34 | |
| *** yangyapeng has joined #openstack-cinder | 19:36 | |
| openstackgerrit | Merged openstack/cinder master: Suppress UT log messages https://review.openstack.org/521555 | 19:36 |
| openstackgerrit | Merged openstack/cinder master: Inspur Cinder iSCSI driver https://review.openstack.org/502455 | 19:37 |
| smcginnis | Let the rechecks... begin. | 19:37 |
| *** jmlowe has quit IRC | 19:37 | |
| *** yangyapeng has quit IRC | 19:41 | |
| jungleboyj | smcginnis: Wheee! | 19:49 |
| *** amoralej is now known as amoralej|off | 19:50 | |
| *** yangyapeng has joined #openstack-cinder | 19:57 | |
| *** lkuchlan_ has quit IRC | 20:01 | |
| *** yangyapeng has quit IRC | 20:01 | |
| ildikov | smcginnis: we have a very quick Upstream Institute meeting on #openstack-meeting-3 if you're interested :) | 20:09 |
| *** e0ne has joined #openstack-cinder | 20:11 | |
| *** yangyapeng has joined #openstack-cinder | 20:17 | |
| *** yangyapeng has quit IRC | 20:22 | |
| *** salv-orlando has quit IRC | 20:24 | |
| *** salv-orlando has joined #openstack-cinder | 20:25 | |
| *** salv-orlando has quit IRC | 20:28 | |
| *** edmondsw has joined #openstack-cinder | 20:29 | |
| *** schmots has quit IRC | 20:30 | |
| *** schmots has joined #openstack-cinder | 20:30 | |
| *** edmondsw has quit IRC | 20:33 | |
| *** liverpooler has joined #openstack-cinder | 20:33 | |
| *** salv-orlando has joined #openstack-cinder | 20:36 | |
| *** pcaruana has quit IRC | 20:39 | |
| *** yangyapeng has joined #openstack-cinder | 20:41 | |
| *** yangyapeng has quit IRC | 20:45 | |
| *** schmots has quit IRC | 20:48 | |
| openstackgerrit | Sean McGinnis proposed openstack/cinder master: Move legacy-cinder-tox-* jobs in-tree https://review.openstack.org/521654 | 20:49 |
| *** schmots has joined #openstack-cinder | 20:51 | |
| openstackgerrit | Sean McGinnis proposed openstack/cinder master: Move legacy-cinder-tox-* jobs in-tree https://review.openstack.org/521654 | 20:53 |
| *** felipemonteiro has joined #openstack-cinder | 20:57 | |
| openstackgerrit | Sean McGinnis proposed openstack/cinder master: Move legacy-cinder-tox-* jobs in-tree https://review.openstack.org/521654 | 20:58 |
| abishop | smcginnis: do you think there's any hope for https://review.openstack.org/496858 | 20:58 |
| abishop | smcginnis: I gather it's needed for https://review.openstack.org/485232 | 20:59 |
| smcginnis | abishop: I do think we still need it. We are testing against v2 right now, and that's deprecated. We really need to get to v3. | 20:59 |
| smcginnis | abishop: I don't think it's necessarily required for the v1 endpoint removal. IIRC, there are just a few spots that need to be updated for that one. | 20:59 |
| abishop | smcginnis: I couldn't follow the details, but somehow it's related to removing v1 endpoint ? | 20:59 |
| smcginnis | abishop: Not really. | 21:00 |
| smcginnis | abishop: Just switching from running tests against v2 by default to running against v3. | 21:00 |
| abishop | smcginnis: looks like I need to study some details. I'm on the hook to do some TripleO cleanup in this area | 21:01 |
| smcginnis | abishop: Cool. If I can help, let me know. | 21:01 |
| abishop | will do, thx | 21:02 |
| smcginnis | abishop: And if you can pressure qa to get that through, I would love some help there. ;) | 21:02 |
| *** Apoorva_ has joined #openstack-cinder | 21:02 | |
| *** Apoorva has quit IRC | 21:06 | |
| ildikov | smcginnis: jungleboyj: so I'm trying to figure out how to do the check in Nova not to allow having the same volume attached to the same instance on the same host twice | 21:07 |
| ildikov | smcginnis: jungleboyj: so I was looking into the volume attachment details and found an attached_host field | 21:08 |
| ildikov | smcginnis: jungleboyj: and was wondering what that is exactly? :) | 21:09 |
| jungleboyj | ildikov: Good question. | 21:09 |
| jungleboyj | It is used in 'migrate_volume_completion' | 21:12 |
| *** e0ne has quit IRC | 21:12 | |
| *** erlon has joined #openstack-cinder | 21:12 | |
| *** AlexeyAbashkin has joined #openstack-cinder | 21:12 | |
| mtreinish | smcginnis: I just added another comment https://github.com/mtreinish/stestr/pull/127 with comparisons of that section to testr | 21:13 |
| jungleboyj | jgriffith: ^^ | 21:13 |
| mtreinish | I did try changing the character encoding to latin-1 (from utf8) in case it wasn't able to handle some unicode character in the logs (which seemed unlikely) | 21:13 |
| mtreinish | but that didn't seem to fix it either | 21:13 |
| jungleboyj | ildikov: Is your question if that can be used to determine if the host is already attached? | 21:15 |
| *** AlexeyAbashkin has quit IRC | 21:17 | |
| *** dave-mccowan has quit IRC | 21:18 | |
| ildikov | jungleboyj: so we don't want to allow the same volume attached to the same in stance on the same host twice | 21:20 |
| jungleboyj | ildikov: Right. | 21:20 |
| ildikov | jungleboyj: and as we always say that Cinder should be the source of truth, I try to find a way to figure out the host for the instance the volume is already attached to | 21:20 |
| jungleboyj | Right. | 21:21 |
| *** markstur has joined #openstack-cinder | 21:22 | |
| ildikov | and the only thing I saw was that fiels | 21:23 |
| jungleboyj | Yeah, looking here. | 21:23 |
| ildikov | but wasn't sure whether that's a uuid, a hostname, if it's the instance's host even | 21:23 |
| smcginnis | mtreinish: Maybe that's not the right place? | 21:23 |
| ildikov | the other option is the BDM in Nova of course, but the flows are a bit messy there... | 21:23 |
| mtreinish | smcginnis: I replaced all six.binary_type() casts and encode('utf8') calls to be encode('latin-1') iirc | 21:24 |
| mtreinish | I'll revive the patch and test it again | 21:25 |
| smcginnis | mtreinish: FYI, since merging a patch to limit the amount of logger output captured per test, we are now getting good results running py35 unit tests. | 21:26 |
| mtreinish | smcginnis: ok cool, but it still doesn't help me understand what's going on :) | 21:27 |
| mtreinish | and this is really bothering me that I can't figure out why it only fails on cinder logging and py3 | 21:27 |
| mtreinish | and it's not hitting the attachment size limit in subunit | 21:27 |
| smcginnis | mtreinish: Yeah, this one has been really frustrating. | 21:27 |
| smcginnis | mtreinish: Must be large enough to trigger some condition, but small enough to be under the attachment size limit. | 21:28 |
| mtreinish | smcginnis: sure, but you'd think if I injected random strings in an attachment right under the size limit it would trip it | 21:29 |
| mtreinish | but I couldn't reproduce that by just printing random data in a loop | 21:29 |
| jungleboyj | ildikov: So, I just was playing with it on the system I am on and it shows the name of the host where the instance is running. | 21:29 |
| openstackgerrit | Sean McGinnis proposed openstack/cinder master: DNM: Test for in-tree compliance test https://review.openstack.org/521661 | 21:29 |
| jungleboyj | ildikov: Oh wait. | 21:30 |
| jungleboyj | Interesting 'attached_host' is NULL if there is an attached_instance it appears. | 21:31 |
| jgriffith | jungleboyj: ildikov https://github.com/openstack/cinder/blob/master/cinder/db/sqlalchemy/models.py#L915 | 21:31 |
| jgriffith | That exists for every attachment, and includes the connector for each attachment | 21:32 |
| jgriffith | and the connector requires things like instance and host | 21:32 |
| openstackgerrit | Louie Kwan proposed openstack/cinder master: Allow purging of records less than 1 day old https://review.openstack.org/511567 | 21:32 |
| jgriffith | attached_host in the main volumeattachment column is also a *thing* | 21:33 |
| mtreinish | smcginnis: fwiw, this is what I was testing: http://paste.openstack.org/show/626864/ (and I just ran it locally again and it failed) | 21:33 |
| jgriffith | as is InstanceUUID | 21:33 |
| jgriffith | so I'm unclear what the question is? | 21:33 |
| mtreinish | smcginnis: this totally feels like a mutliprocess failure to me. Like multiple output streams interfering with each other. | 21:33 |
| mtreinish | but the simple reproduce cases basically ruled that out | 21:34 |
| jungleboyj | Well, my question at the moment is why attachment_host is null when instance_uuid is set. | 21:35 |
| jungleboyj | Why are both not set? | 21:35 |
| jungleboyj | jgriffith: The main question goes back to ildikov 's question of how to determine if a volume is already attached once to an instance on the same host and if attached_host can help. | 21:36 |
| smcginnis | mtreinish: Yeah, I was hitting it running a single test. | 21:36 |
| jgriffith | jungleboyj: attached_host and instance_uuid are the crazy old db-api/object routines db:volume_attached() | 21:37 |
| smcginnis | mtreinish: Some conversion issue somewhere in the flow still seems likely. The output is always truncated, and it's always b'' even though other output is not. | 21:37 |
| jgriffith | it's an update mechanism | 21:37 |
| jgriffith | jungleboyj: I believe they are both set when an actual attachment is made/connected. | 21:38 |
| jungleboyj | jgriffith: Hmmm, I am probably using the old attachment flow here. | 21:38 |
| jgriffith | well, you definitely are | 21:39 |
| jgriffith | but regardless, the old flow should be setting that | 21:39 |
| jungleboyj | jgriffith: Yes. | 21:39 |
| jgriffith | if not, skip it and in the new code use the new attachmentspecs data to retrieve the connector that was used | 21:39 |
| jgriffith | jungleboyj: looks something like this: http://paste.openstack.org/show/626866/ | 21:40 |
| jgriffith | if you want a "real" full view you should run it locally and dump it to the logs :) | 21:41 |
| jungleboyj | jgriffith: Oh duh! We have to know the host from the connector. That is how the storage knows who to set up the target for. | 21:41 |
| jungleboyj | ildikov: ^^^ | 21:41 |
| jgriffith | jungleboyj: right | 21:42 |
| jungleboyj | I have only looked at that at least once today. | 21:42 |
| * jungleboyj is on cold medicine. | 21:42 | |
| jgriffith | jungleboyj: mmmmmm... drugs man, drugs... far out | 21:43 |
| jungleboyj | Yeah ... *cough cough* They are only working so well. | 21:43 |
| jgriffith | I can spin up devstack and look at some of this in a little bit here | 21:43 |
| jgriffith | I don't have a full running devstack at the moment, all standalone stuff | 21:44 |
| mriedem | jungleboyj: i suppose we won't get an os-brick release until next week aye? | 21:44 |
| jungleboyj | mriedem: Well, I can spin one whenever we need one. Though, I am thinking waiting until next week is probably best. :-) | 21:44 |
| *** felipemonteiro has quit IRC | 21:44 | |
| *** eharney has quit IRC | 21:45 | |
| smcginnis | jungleboyj, mriedem: I think we just had a requirements update I wanted to see go through. | 21:45 |
| smcginnis | We could probably get one going. | 21:45 |
| * smcginnis checks the queue again | 21:45 | |
| mriedem | there is a nova libvirt volume driver change depending on a new brick connector | 21:45 |
| *** linkmark has joined #openstack-cinder | 21:45 | |
| jungleboyj | mriedem: Oh yeah I saw a note on that. | 21:46 |
| smcginnis | A cinder volume driver too. | 21:46 |
| jgriffith | jungleboyj: ildikov By the way.. https://github.com/openstack/cinder/blob/master/doc/source/contributor/attach_detach_conventions.rst | 21:46 |
| jgriffith | and https://github.com/openstack/cinder/blob/master/doc/source/contributor/attach_detach_conventions_v2.rst | 21:46 |
| jgriffith | might be useful as well | 21:46 |
| jgriffith | hmm... although | 21:47 |
| jgriffith | there's a bunch of stuff that's been removed there? | 21:47 |
| *** rcernin has joined #openstack-cinder | 21:48 | |
| Swanson | Charles Manson is dead. That just leaves Paul and Ringo. | 21:48 |
| Swanson | Carry on. | 21:48 |
| jgriffith | There used to be a pretty comprehensive doc in tree that walked through attach/detach and included all of the inputs and responses | 21:48 |
| jgriffith | Swanson: he's not really dead ya know | 21:48 |
| smcginnis | mriedem: I'm not familiar with these nova config settings. Does this look OK to you? https://review.openstack.org/#/c/510776/ | 21:49 |
| jgriffith | Swanson: it's a conspiracy | 21:49 |
| smcginnis | Hah | 21:49 |
| jgriffith | He's running for office in 2020 | 21:49 |
| Swanson | Very surprised he didn't get the DHS spot before he died. | 21:50 |
| jungleboyj | jgriffith: hemna had that I thought. | 21:50 |
| mtreinish | smcginnis: I'll keep playing with this in a bit (need to change locations now) https://github.com/mtreinish/stestr/commit/a0ef15a27d3202d60c32d465d7c24e50e69f82a0 | 21:51 |
| ildikov | jungleboyj: jgriffith: sorry, I'm back | 21:51 |
| mtreinish | smcginnis: but in my local testing it didn't make any difference | 21:51 |
| jgriffith | jungleboyj: https://j-griffith.github.io/articles/2015-07/volume-attach-code-flow-in-cinder | 21:51 |
| mriedem | smcginnis: i'm not familiar with the libvirt.iscsi_iface option either, but i see our help text on that option doesn't mention iser or tcp | 21:51 |
| jungleboyj | ildikov: Welcome. I think jgriffith answered your question. :-) | 21:52 |
| mriedem | ooo we have a doc | 21:52 |
| jgriffith | jungleboyj: I also merged that into tree at one point but looks like it's gone with the doc migration stuff | 21:52 |
| smcginnis | brb | 21:52 |
| mriedem | https://docs.openstack.org/nova/pike/admin/configuration/iscsi-offload.html | 21:52 |
| *** schmots has quit IRC | 21:54 | |
| jungleboyj | smcginnis: Looks like we have one os-brick patch working through the gate and one that hit a gate failure. | 21:55 |
| mriedem | smcginnis: commented inline | 21:56 |
| ildikov | jungleboyj: jgriffith: my bad, too many moving parts here... | 21:56 |
| jungleboyj | ildikov: Why your bad? | 21:57 |
| ildikov | jungleboyj: jgriffith: I'm also getting a little confused on what gets returned with the volume info now that we have the two attachment flows, but I guess that didn't fundamentally changed? | 21:57 |
| jgriffith | ildikov: well, it was vastly different initially | 21:58 |
| jgriffith | oh wait.. volume info? | 21:58 |
| ildikov | jgriffith: there's attachment info in the volume info | 21:58 |
| ildikov | jgriffith: or well, used to be at least :) | 21:59 |
| jgriffith | ildikov: maybe start at the beginning.. what are you trying to do? | 21:59 |
| ildikov | jgriffith: trying to add the check in Nova to prevent attaching the smae volume to the same instance on the same host twice with the principle of Cinder is the ultimate source of truth | 22:00 |
| jgriffith | ildikov: using the new attachment api's? | 22:01 |
| *** itlinux has quit IRC | 22:01 | |
| ildikov | jgriffith: I thought to save a few calls to Cinder if I have the info in the volume info available too | 22:02 |
| ildikov | jgriffith: as we do a volume get before attach in the Nova API | 22:02 |
| jgriffith | ildikov: can't really do that from the volume | 22:04 |
| *** itlinux has joined #openstack-cinder | 22:04 | |
| jgriffith | ildikov: you could look at attach_status, if already attached then get attachments-for-volume and go from there | 22:04 |
| jgriffith | or even filter on host and instance and if nothing comes back then proceed | 22:04 |
| jgriffith | but the attachment details are decoupled from the volume | 22:04 |
| *** Apoorva_ has quit IRC | 22:06 | |
| ildikov | jgriffith: I would just get the attachments and go through them to see whether the host appears in any of them | 22:07 |
| jgriffith | ildikov: you don't need to do that though, no need to iterate when the db query will get what you want anyway | 22:09 |
| jgriffith | ildikov: also if it's helpful there's this data too: https://gist.github.com/j-griffith/fe0cbe90d7fd8c067d0568b8e9313d57 | 22:09 |
| smcginnis | mriedem: Thanks for taking a look. | 22:09 |
| jgriffith | ildikov: I mean, it depends on what you mean by "just get the attachments and go through them" | 22:09 |
| jgriffith | ildikov: if you mean for that volume, that's not so bad | 22:10 |
| smcginnis | jungleboyj: The gate failure one doesn't look critical, so we could wait to pick that one up in the next rev. | 22:10 |
| jungleboyj | Ok. | 22:10 |
| jgriffith | ildikov: but you could just specify exactly what you want with filters on things like volume and host which makes it pretty simple I think | 22:10 |
| smcginnis | jungleboyj: Oh, actually that's a backport to stable/ocata, so doesn't even matter. | 22:10 |
| ildikov | jgriffith: if I can get a list of attachments for a given volume that contains the info I can get the host from, that's what I would need | 22:11 |
| ildikov | I guess :) | 22:11 |
| jungleboyj | smcginnis: Right you are. So just wait for ignoreskipactivation ? | 22:12 |
| openstackgerrit | Merged openstack/cinder master: [Trivial] set min max_over_subscription_ratio to 1 https://review.openstack.org/520049 | 22:13 |
| openstackgerrit | Merged openstack/cinder master: Fix qos-spec-v2-qos-specs error https://review.openstack.org/521348 | 22:13 |
| openstackgerrit | Merged openstack/cinder master: Add v3 api-ref for showing encryption specs item https://review.openstack.org/521239 | 22:13 |
| jgriffith | ildikov: there's probably not a filter option in the api yet for attachments, but you can do volume-id like you said; or you can just look at the bdm's for the host and avoid the call otu to Cinder altogether | 22:14 |
| jgriffith | seems like there are multiple ways to do it, not sure what might be best | 22:14 |
| *** mvk has joined #openstack-cinder | 22:14 | |
| * jgriffith looking at api/v3/attachments.py | 22:15 | |
| jgriffith | Yeah, not sure you're going to have what you want directly exposed there other than generic list and show | 22:16 |
| ildikov | jgriffith: I thought to add the check to check_attach_and_reserve_volume in Nova, but by then we create a new BDM, which might make things messy on that front... | 22:16 |
| jgriffith | ildikov: you could just check nova BDM's for the host you're on *before* the new attach | 22:16 |
| *** e0ne has joined #openstack-cinder | 22:16 | |
| jgriffith | if one exists on that host for that volume already then you refuse to proceed | 22:17 |
| jgriffith | otherwise, continue on | 22:17 |
| ildikov | jgriffith: yeah, I guess I'll need to do that | 22:17 |
| *** edmondsw has joined #openstack-cinder | 22:17 | |
| jungleboyj | jgriffith: That makes sense. | 22:17 |
| jgriffith | I still think we should just enforce it on the Cinder side though. | 22:17 |
| jgriffith | I mean... the behavior currently is a bug, and has to be fixed | 22:18 |
| ildikov | jgriffith: well, we only have connection_info in the BDM | 22:18 |
| jgriffith | I just can't fix it until I know what is happening on the other side :) | 22:18 |
| jgriffith | ildikov: that's all you need though no? | 22:18 |
| ildikov | jgriffith: so I guess I can figure it out from the Instance... | 22:18 |
| jgriffith | ildikov: you don't need that much info.. just a volume-id and the nova host (which you're on so you already have that) | 22:19 |
| ildikov | jgriffith: yeah, I would be happy if it could be fixed on the Cinder side :) | 22:19 |
| ildikov | jgriffith: when the request comes in to the Nova API, I think I only have the info of the Instance | 22:19 |
| jgriffith | only if it's boot from volume | 22:20 |
| ildikov | jgriffith: as if we go down to the compute we might as well have Cinder to tell we're stupid | 22:20 |
| jgriffith | if it's not BFV you're in the manager layer anyway and the instance has been scheduled | 22:21 |
| *** rtjure has quit IRC | 22:21 | |
| ildikov | yeah, running actually | 22:21 |
| jgriffith | and even BFV, finalization of the attachment/connection to cinder doesn't happen til after scheduling the instance either | 22:21 |
| ildikov | I guess I should've just asked a BDM advice from mriedem :) | 22:21 |
| *** edmondsw has quit IRC | 22:22 | |
| jgriffith | ildikov: what.. my advice isn't good enough? :) | 22:22 |
| ildikov | and get a lecture on how simple it is before I got into my complicated thoughts, lol | 22:22 |
| jungleboyj | jgriffith: Is better than mine. :-) | 22:22 |
| ildikov | jgriffith: I just meant to save you and everyone here from a Nova BDM discussion | 22:23 |
| *** rtjure has joined #openstack-cinder | 22:23 | |
| ildikov | jgriffith: I always like your advices :) | 22:23 |
| *** yangyapeng has joined #openstack-cinder | 22:23 | |
| jungleboyj | Group hug! | 22:23 |
| *** liverpooler has quit IRC | 22:24 | |
| jgriffith | ildikov: :) I was just teasing ya | 22:24 |
| jgriffith | besides, mriedem has enough on his plate, like reviewing my patch updates :) | 22:25 |
| ildikov | jgriffith: so, when we do calls like chakc_attach_and_reserve_volume we don't call down to the manager yet | 22:25 |
| jgriffith | see what I did there.... tricky :) | 22:25 |
| ildikov | jgriffith: fingers crossed it works ;) | 22:25 |
| *** abishop has quit IRC | 22:27 | |
| *** yangyapeng has quit IRC | 22:28 | |
| ildikov | jgriffith: I'm still confused wit the BDM though :) | 22:31 |
| *** e0ne has quit IRC | 22:32 | |
| *** e0ne has joined #openstack-cinder | 22:33 | |
| ildikov | jgriffith: hmm, ok, I guess I could just say that if the BDM for the volume and instance exists already then fail as we only attach the same volume to the same instance twice for live_migrate and that's a different path anyway | 22:36 |
| ildikov | jgriffith: and then someone will eat me alive in the review for this, but oh well, I'm a big girl, I can take it... | 22:36 |
| jgriffith | ildikov: I thought the scenario you described there was exactly what we talked about before | 22:37 |
| jgriffith | seems like a good idea to me | 22:37 |
| ildikov | jgriffith: the only problem I guess is that in Nova there are ghost BDM's every now and then IIUC, which means I might fail the call over an invalid BDM | 22:38 |
| ildikov | jgriffith: shouldn't be my problem for now though :) | 22:38 |
| ildikov | jgriffith: and things like swap_volume can still give me a headache... | 22:39 |
| *** Apoorva has joined #openstack-cinder | 22:39 | |
| jgriffith | ildikov: they shouldn't, because it's a new volume | 22:39 |
| *** Apoorva has quit IRC | 22:40 | |
| jgriffith | at least in this context it shouldn't | 22:40 |
| jgriffith | it *should* hover give you a headache because it's sort of nutty | 22:40 |
| *** Apoorva has joined #openstack-cinder | 22:40 | |
| ildikov | new volume? | 22:40 |
| *** yangyapeng has joined #openstack-cinder | 22:44 | |
| *** e0ne has quit IRC | 22:49 | |
| *** yangyapeng has quit IRC | 22:49 | |
| *** itlinux has quit IRC | 22:49 | |
| *** itlinux has joined #openstack-cinder | 22:50 | |
| *** lbragstad has quit IRC | 22:51 | |
| ildikov | hmm, one more thing | 22:55 |
| ildikov | if I just do a BDM check before everything else, I cannot easily differentiate between the old new flows | 22:55 |
| ildikov | and we need this only for the new flow | 22:56 |
| ildikov | or maybe it's the same error in both cases anyway and that'll hide it well enough | 22:56 |
| jgriffith | ildikov: hmmm | 22:57 |
| jgriffith | ildikov: well, actually if you find a BDM with the volume-id of interest you can just check it to see if it has an attachment id no? | 22:58 |
| jgriffith | ildikov: so the same way we are doing it everywhere else on the Nova side currently | 22:58 |
| ildikov | jgriffith: but if we try to attach the volume with the new flow now as the new Cinder version is available and I let this go because the old flow is handling it in Cinder, then I'm screwed, so I guess if I check something that would be the Cinder microversion | 22:59 |
| jgriffith | ildikov: ? | 22:59 |
| jgriffith | that made no sense to me at all | 23:00 |
| ildikov | so the check is at places like a normal attach_volume right now | 23:00 |
| ildikov | where we decide on how to attach a volume based on the available Cinder microversion | 23:00 |
| jgriffith | ildikov: you mean if they like did an upgrade while a volume was attached? Then fine, check it and behave the same regardless | 23:00 |
| jgriffith | we're not supporting multi-attach in the old API anyway so it's sort of easy to just punch out before even getting that far I'd think | 23:01 |
| ildikov | the more I think about it the more it gets a microversion check whether or not to check the BDM in Nova | 23:02 |
| jgriffith | but that's where we started I think | 23:02 |
| jgriffith | and that got nixed | 23:02 |
| ildikov | I can upload what I've just done for now and you can tell how many places it's fundamentally wrong | 23:03 |
| jgriffith | ildikov: maybe it's not wrong at all... maybe it's perfect? | 23:03 |
| *** lbragstad has joined #openstack-cinder | 23:04 | |
| ildikov | jgriffith: yeah, it's already Tuesday here, so maybe :) | 23:05 |
| ildikov | jgriffith: https://review.openstack.org/#/c/330285/148..149/nova/compute/api.py | 23:06 |
| ildikov | jgriffith: need to test it still whether it does what I want, but something like this was the idea | 23:07 |
| ildikov | the problem with the old flow is that if we're changing an error message with this, that's a problem... | 23:07 |
| jgriffith | that looks sane to me | 23:08 |
| jgriffith | if you're worried about that then add a check for bdm.attachment_id around line #3530 | 23:09 |
| ildikov | ok, I will ask Matt to take a look too then and if he thinks the same I will fix/add tests | 23:09 |
| jgriffith | I don't know.. not sure I follow the problem exactly but yeah, the Nova cores will certainly be better resources than I for this | 23:09 |
| *** sdague has quit IRC | 23:09 | |
| ildikov | in this case the old attachment doesn't determine which flow we plan to attach the volume | 23:10 |
| ildikov | jgriffith: no offense really, I just remember a couple of cases we thought things are sane, but they turned out not be when it came to reviews... | 23:10 |
| ildikov | so I'm not putting too much effort into anything now before validating the idea... :) | 23:11 |
| ildikov | call me lazy, I probably am, but still :) | 23:12 |
| ildikov | jgriffith: ok, my soul is saved, and we're both sane today :) | 23:19 |
| *** lpetrut has quit IRC | 23:32 | |
| *** mriedem is now known as mriedem_away | 23:36 | |
| *** linkmark has quit IRC | 23:38 | |
| *** armax has quit IRC | 23:59 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!