| *** enriquetaso has quit IRC | 00:04 | |
| *** hoonetorg has quit IRC | 00:10 | |
| *** n-saito has quit IRC | 00:12 | |
| *** zhanglong has joined #openstack-cinder | 00:26 | |
| *** n-saito has joined #openstack-cinder | 00:34 | |
| *** rcernin has quit IRC | 00:42 | |
| *** rcernin has joined #openstack-cinder | 00:46 | |
| *** zhanglong has quit IRC | 01:12 | |
| *** zhanglong has joined #openstack-cinder | 01:13 | |
| openstackgerrit | Walter A. Boring IV (hemna) proposed openstack/cinder master: WIP Add ceph iscsi volume driver https://review.opendev.org/662829 | 01:19 |
|---|---|---|
| *** rcernin has quit IRC | 02:13 | |
| *** rcernin has joined #openstack-cinder | 02:15 | |
| *** rcernin has quit IRC | 02:33 | |
| *** dave-mccowan has quit IRC | 02:41 | |
| *** dave-mccowan has joined #openstack-cinder | 02:49 | |
| *** rcernin has joined #openstack-cinder | 02:55 | |
| *** zhanglong has quit IRC | 03:16 | |
| *** ociuhandu has joined #openstack-cinder | 03:16 | |
| *** zhanglong has joined #openstack-cinder | 03:16 | |
| *** ociuhandu has quit IRC | 03:20 | |
| *** zhanglong has quit IRC | 03:27 | |
| *** zhanglong has joined #openstack-cinder | 03:28 | |
| *** markvoelker has joined #openstack-cinder | 04:03 | |
| *** markvoelker has quit IRC | 04:07 | |
| *** zhanglong has quit IRC | 04:09 | |
| *** zhanglong has joined #openstack-cinder | 04:10 | |
| *** ykarel|away is now known as ykarel | 04:14 | |
| *** dave-mccowan has quit IRC | 04:19 | |
| *** udesale has joined #openstack-cinder | 04:28 | |
| *** abdysn has joined #openstack-cinder | 05:05 | |
| *** abhishekk|away is now known as abhishekk | 05:33 | |
| *** m75abrams has joined #openstack-cinder | 05:38 | |
| *** xinliang has joined #openstack-cinder | 06:36 | |
| *** ianw is now known as ianw_pto | 06:38 | |
| *** trident has quit IRC | 06:39 | |
| *** xinliang has quit IRC | 06:42 | |
| *** trident has joined #openstack-cinder | 06:42 | |
| *** markvoelker has joined #openstack-cinder | 06:43 | |
| *** markvoelker has quit IRC | 06:48 | |
| *** rcernin has quit IRC | 06:59 | |
| openstackgerrit | Keigo Noha proposed openstack/cinder master: Change default glance_num_retries to 3 https://review.opendev.org/739907 | 07:01 |
| *** rcernin has joined #openstack-cinder | 07:06 | |
| *** tosky has joined #openstack-cinder | 07:33 | |
| *** xinliang has joined #openstack-cinder | 07:36 | |
| *** rcernin has quit IRC | 07:41 | |
| *** zhanglong has quit IRC | 08:01 | |
| *** martinkennelly has joined #openstack-cinder | 08:02 | |
| *** martinkennelly has quit IRC | 08:03 | |
| *** ociuhandu has joined #openstack-cinder | 08:10 | |
| *** Lucas_Gray has joined #openstack-cinder | 08:11 | |
| *** zhanglong has joined #openstack-cinder | 08:14 | |
| *** Wryhder has joined #openstack-cinder | 08:16 | |
| *** Lucas_Gray has quit IRC | 08:16 | |
| *** Wryhder is now known as Lucas_Gray | 08:17 | |
| *** m75abrams is now known as mabrams | 08:17 | |
| *** zhanglong has quit IRC | 08:19 | |
| *** zhanglong has joined #openstack-cinder | 08:21 | |
| *** tkajinam has quit IRC | 08:23 | |
| *** zhanglong has quit IRC | 08:33 | |
| *** spatel has joined #openstack-cinder | 08:36 | |
| *** spatel has quit IRC | 08:41 | |
| *** devfaz has joined #openstack-cinder | 08:49 | |
| devfaz | Hi, im just trying to upgrade from pike to queens, and while doing online_data_migrations im just getting "Error attempting to run volume_service_uuids_online_data_migration". Is there a way to get more information about the source of the issue? I already tried "--debug", but didnt add futher output. | 08:50 |
| *** zhanglong has joined #openstack-cinder | 08:55 | |
| *** zhanglong has quit IRC | 09:05 | |
| whoami-rajat | devfaz, hi, did you run db migrations before the running the online migration? | 09:05 |
| *** zhanglong has joined #openstack-cinder | 09:06 | |
| devfaz | yes, but I can repeat it.. one moment. | 09:06 |
| devfaz | you mean "db sync"? | 09:07 |
| devfaz | whoami-rajat: you mean "db sync"? | 09:08 |
| whoami-rajat | devfaz, yep, cinder-manage db sync | 09:09 |
| devfaz | whoami-rajat, yes done twice - didnt change anything. | 09:09 |
| whoami-rajat | devfaz, hmm, can you access sql db? if yes, can you check the volumes table contains a column named 'service_uuid' | 09:10 |
| devfaz | give me a minute.. | 09:10 |
| whoami-rajat | sure | 09:11 |
| devfaz | yes, there is a service_uuid-field. | 09:11 |
| whoami-rajat | the commands are, use cinder; desc volumes; | 09:11 |
| whoami-rajat | ok | 09:11 |
| whoami-rajat | does it has any non null values? | 09:13 |
| whoami-rajat | devfaz, ^ | 09:13 |
| devfaz | whoami-rajat, found https://bugs.launchpad.net/cinder/+bug/1821207 and we have NULL values in the host-field in some volumes. We will try to remove these volumes or do you want to check something before? | 09:15 |
| openstack | Launchpad bug 1821207 in Cinder ""cinder-manage db online_data_migrations" fails if volume as a NULL value in host field" [Undecided,Fix released] - Assigned to Eric Harney (eharney) | 09:15 |
| devfaz | whoami-rajat, and yes - there *are* already non null values in service_uuid. | 09:16 |
| whoami-rajat | devfaz, so the migration ran for some volumes but failed for some | 09:17 |
| whoami-rajat | devfaz, if you could confirm that there is NULL host field in volumes then that would be the issue mentioned in launchpad bug | 09:18 |
| devfaz | whoami-rajat, just did sql query deleted=0 and host is NULL - got ~6 Volumes. Will try to remove these if you dont want to use this for debugging things. | 09:19 |
| whoami-rajat | devfaz, hmm, then there's probably the problem with the code, the 6 volumes should have service_uuid field as NULL | 09:21 |
| openstackgerrit | Rajat Dhasmana proposed openstack/cinder stable/queens: Fix service_uuid migration for volumes with no host https://review.opendev.org/739937 | 09:23 |
| whoami-rajat | devfaz, yeah, that was backported till rocky but not in queens. doing this change should fix your problem ^^ | 09:23 |
| devfaz | whoami-rajat, thx a lot! I will give it a try. | 09:24 |
| whoami-rajat | np | 09:25 |
| *** rcernin has joined #openstack-cinder | 09:26 | |
| devfaz | whoami-rajat, thx a lot - worked! | 09:27 |
| whoami-rajat | devfaz, great! | 09:30 |
| openstackgerrit | wanghao proposed openstack/cinder master: time comparison filter didn't support time zone offset https://review.opendev.org/734415 | 09:36 |
| *** rcernin has quit IRC | 09:41 | |
| *** geguileo has quit IRC | 09:47 | |
| *** geguileo has joined #openstack-cinder | 09:50 | |
| openstackgerrit | haixin proposed openstack/cinder master: Query error state volumes should contain error_managing state https://review.opendev.org/725778 | 09:59 |
| *** e0ne has joined #openstack-cinder | 10:12 | |
| *** zhanglong has quit IRC | 10:14 | |
| *** brinzhang_ has quit IRC | 10:16 | |
| *** zhanglong has joined #openstack-cinder | 10:23 | |
| *** zhanglong has quit IRC | 10:27 | |
| *** zhanglong has joined #openstack-cinder | 10:34 | |
| *** bhagyashris is now known as bhagyashris|brb | 10:42 | |
| *** zhanglong has quit IRC | 10:44 | |
| *** zhanglong has joined #openstack-cinder | 10:45 | |
| *** xinliang has quit IRC | 10:54 | |
| stephenfin | rosmaita: When you're about, could I get your thoughts on https://review.opendev.org/#/c/736758/ ? I'm sure it's bound to be a little contentious but it seems like the easiest-to-debug approach IMO | 10:59 |
| *** bhagyashris|brb is now known as bhagyashris | 10:59 | |
| *** udesale_ has joined #openstack-cinder | 11:03 | |
| tosky | geguileo: ^^ | 11:04 |
| geguileo | lol, my code in cinderlib did that, because that also happens if you are running cinder in a container and the host doesn't have ceph-common installed | 11:06 |
| geguileo | though the code doesn't look complete to me | 11:06 |
| geguileo | and should use privsep | 11:06 |
| geguileo | and remove the link on disconnect | 11:06 |
| *** udesale has quit IRC | 11:06 | |
| openstackgerrit | Lucian Petrut proposed openstack/cinder master: smbfs: pick up remotefs method signature change https://review.opendev.org/739954 | 11:19 |
| *** zhanglong has quit IRC | 11:29 | |
| geguileo | stephenfin: I have reviewed the patch | 11:29 |
| geguileo | I'm OK with creating the symlinks (actually more than happy), but there are cases that are not covered by the patch | 11:30 |
| *** markvoelker has joined #openstack-cinder | 11:34 | |
| *** markvoelker has quit IRC | 11:39 | |
| openstackgerrit | Walter A. Boring IV (hemna) proposed openstack/cinder master: WIP Add ceph iscsi volume driver https://review.opendev.org/662829 | 11:46 |
| *** pcaruana has quit IRC | 11:54 | |
| *** pcaruana has joined #openstack-cinder | 11:55 | |
| *** rosmaita has left #openstack-cinder | 12:07 | |
| *** rcernin has joined #openstack-cinder | 12:08 | |
| *** rcernin has quit IRC | 12:16 | |
| stephenfin | geguileo++ thanks | 12:18 |
| *** rosmaita has joined #openstack-cinder | 12:21 | |
| rosmaita | geguileo: thanks for answering stephenfin | 12:22 |
| *** zhanglong has joined #openstack-cinder | 12:24 | |
| *** hoonetorg has joined #openstack-cinder | 12:25 | |
| rosmaita | stephenfin: i'm inclined to agree with lyarwood ... your commit message says "Rather than require udev be added to the container or ceph's udev rules be placed on the host, simply do the work of the udev rule and create the symlink manually", but this is turning out to be not that simple (and i say that having +1'd PS 1). I'm beginning to think the correct fix is to require that appropriate dependencies be present. | 12:28 |
| *** raghavendrat has joined #openstack-cinder | 12:34 | |
| raghavendrat | hi reviewers, | 12:48 |
| raghavendrat | this is regarding https://review.opendev.org/#/c/739033/ and https://review.opendev.org/#/c/739416/ | 12:48 |
| raghavendrat | for both, zuul and CI have passed. | 12:48 |
| raghavendrat | It would be great if someone can have a look. Thanks. | 12:48 |
| *** yaawang has quit IRC | 12:54 | |
| *** pcaruana has quit IRC | 12:56 | |
| *** yaawang has joined #openstack-cinder | 12:58 | |
| geguileo | rosmaita: Cinderlib is currently creating and deleting the symlinks for RBD | 13:04 |
| geguileo | rosmaita: so I'm in favor of adding it to os-brick | 13:04 |
| geguileo | rosmaita: that way we can move cinderlib to privsep and remove the nos_brick.py file from cinderlib that fixes all the missing stuff in the RBD os-brick connector | 13:05 |
| rosmaita | well, we will have to beef up the testing somehow, because the current unit test did not catch any of the problems you brought up on the review | 13:06 |
| rosmaita | i mean, we are lucky that you already figured this out in cinderlib and knew what might fail | 13:07 |
| rosmaita | so we do have that | 13:07 |
| geguileo | rosmaita: I currently have other 3 os-brick patches fixing the RBD connector for local attachments | 13:07 |
| geguileo | https://review.opendev.org/737311 | 13:08 |
| geguileo | https://review.opendev.org/737312 | 13:08 |
| geguileo | https://review.opendev.org/737310 | 13:08 |
| *** pcaruana has joined #openstack-cinder | 13:09 | |
| tosky | geguileo: add it to os-brick, or add os-brick as dependency? | 13:11 |
| *** rosmaita has quit IRC | 13:11 | |
| geguileo | tosky: adding the link creation/deletion to os-brick | 13:12 |
| geguileo | tosky: basically I'm in favor of that patch | 13:12 |
| tosky | and remove some custom code from cinderlib afterwards? | 13:12 |
| *** jungleboyj has joined #openstack-cinder | 13:17 | |
| geguileo | tosky: yup, the patch for cinderlib is ready https://review.opendev.org/#/c/735935/ | 13:20 |
| geguileo | tosky: you can see that removes 700 LOC ;-) | 13:20 |
| geguileo | tosky: but it requires not only my 3 patches, but also the one from stephenfin | 13:20 |
| *** lpetrut_ has joined #openstack-cinder | 13:21 | |
| tosky | geguileo: and that's after branching cinderlib for ussuri, or before that? :) | 13:21 |
| geguileo | tosky: after, that's why the patch has -W | 13:22 |
| *** raghavendrat has quit IRC | 13:28 | |
| openstackgerrit | Gorka Eguileor proposed openstack/cinder master: Add cinder-manage command to remove file locks https://review.opendev.org/734144 | 13:37 |
| geguileo | whoami-rajat: ^ updated with your requested changes | 13:39 |
| *** rosmaita has joined #openstack-cinder | 13:43 | |
| rosmaita | geguileo: sorry to drop from our discussion, had a power failure & had to relocate to a coffee shop | 13:44 |
| whoami-rajat | geguileo, thanks! | 13:48 |
| *** geguileo has quit IRC | 13:52 | |
| *** Liang__ has joined #openstack-cinder | 13:54 | |
| *** geguileo has joined #openstack-cinder | 13:55 | |
| *** knomura has joined #openstack-cinder | 13:55 | |
| rosmaita | Courtesy reminder: Cinder meeting in #openstack-meeting-alt at 1400 UTC | 13:58 |
| rosmaita | jungleboyj rosmaita smcginnis tosky whoami-rajat m5z e0ne geguileo eharney walshh_ jbernard lseki sfernand rajinir enriquetaso ^^ | 13:58 |
| *** Liang__ has quit IRC | 13:59 | |
| geguileo | rosmaita: thanks! | 13:59 |
| jungleboyj | rosmaita, Thanks. | 13:59 |
| *** dviroel has joined #openstack-cinder | 13:59 | |
| *** Liang__ has joined #openstack-cinder | 14:05 | |
| *** enriquetaso has joined #openstack-cinder | 14:06 | |
| *** bhagyashris is now known as bhagyashris|afk | 14:07 | |
| *** Liang__ is now known as LiangFang | 14:08 | |
| *** mabrams has quit IRC | 14:11 | |
| *** bhagyashris|afk is now known as bhagyashris | 14:17 | |
| *** LiangFang has quit IRC | 14:20 | |
| *** Liang__ has joined #openstack-cinder | 14:21 | |
| *** dave-mccowan has joined #openstack-cinder | 14:23 | |
| hemna | geguileo so that cinder-manage command to delete unused locks is going to delete volumes?! | 14:27 |
| geguileo | noooooooooooooooooooooooooooooooooooooooooooo | 14:27 |
| geguileo | did I write that??? | 14:27 |
| hemna | um it looks like it | 14:27 |
| * geguileo needs to learn how to communicate | 14:27 | |
| hemna | https://review.opendev.org/#/c/734144/5/cinder/cmd/manage.py | 14:27 |
| geguileo | it just delete the file locks | 14:27 |
| *** lpetrut_ has quit IRC | 14:28 | |
| hemna | that's fetching volumes from the db and then calling delete | 14:28 |
| geguileo | hemna: I see, that's wrong :-( | 14:28 |
| geguileo | the comment is wrong | 14:28 |
| geguileo | but the code doesn't delete the volumes | 14:28 |
| hemna | ok I'm quite confused | 14:28 |
| hemna | I guess the name of the function led me to believe that it was | 14:29 |
| geguileo | ooooooh, I see | 14:29 |
| geguileo | no, it's remove existing snapshots/volumes from the dictionary | 14:29 |
| *** mlycka has joined #openstack-cinder | 14:29 | |
| geguileo | and now I see how my naming was terrible | 14:29 |
| geguileo | I'll add a comment now for those methods | 14:29 |
| mlycka | Hi, hello? | 14:30 |
| hemna | can we rename the functions? | 14:30 |
| hemna | and possibly add some comments in them :) | 14:31 |
| *** dave-mccowan has quit IRC | 14:31 | |
| *** abdysn has quit IRC | 14:31 | |
| openstackgerrit | Gorka Eguileor proposed openstack/cinder master: Add cinder-manage command to remove file locks https://review.opendev.org/734144 | 14:33 |
| geguileo | hemna: ^ done | 14:33 |
| geguileo | you were right, the naming was really confusing | 14:34 |
| *** lseki has joined #openstack-cinder | 14:36 | |
| *** mlycka has quit IRC | 14:46 | |
| *** priteau has joined #openstack-cinder | 14:48 | |
| *** Liang__ has quit IRC | 14:49 | |
| hemna | can you adjust the comments in the functions? They still say "remove volumes that still exist...." | 14:51 |
| *** zhanglong has quit IRC | 14:51 | |
| hemna | del volumes[volume_id] looks like deleting a volume.... | 14:52 |
| hemna | "remove volumes from the list that still exist in the DB." | 14:53 |
| *** zhanglong has joined #openstack-cinder | 14:53 | |
| hemna | "so we can know which locks we can then delete from disk" | 14:53 |
| hemna | or something like that | 14:54 |
| rosmaita | a downside to ttx's consolidation of the meeting channels is that we can't go overtime | 15:00 |
| *** sfernand has joined #openstack-cinder | 15:00 | |
| *** zhanglong has quit IRC | 15:00 | |
| rosmaita | geguileo: i think i am coming around to your thought about the config option for the types-list call | 15:01 |
| geguileo | hemna: but it says removing them from the dict | 15:01 |
| hemna | ok sorry I missed that | 15:02 |
| geguileo | rosmaita: yeah, I was just pulling your leg about kicking us out ;-) | 15:02 |
| hemna | lgtm then | 15:02 |
| geguileo | hemna: I will move "from the dict" to earlier in the sentence | 15:02 |
| rosmaita | i kind of liked the original names, really made me look carefully at the code! :) | 15:02 |
| geguileo | rosmaita: I don't love the idea of the config option either, but I think the alternatives are going to introduce other bugs | 15:03 |
| whoami-rajat | geguileo, rosmaita I also liked the idea with a new config option, that will be used to hide/show the __DEFAULT__ type. seems the safest | 15:03 |
| rosmaita | geguileo: yeah, we default it to the way we think the behavior should be, but an operator can change it | 15:03 |
| *** zhanglong has joined #openstack-cinder | 15:03 | |
| geguileo | whoami-rajat: not only to hide/show but also to allow users to use it, right? | 15:04 |
| hemna | what's the issue with the types-list ? | 15:04 |
| whoami-rajat | geguileo, yep. right | 15:04 |
| rosmaita | hemna: read through https://bugs.launchpad.net/cinder/+bug/1886632 | 15:04 |
| openstack | Launchpad bug 1886632 in Cinder "Cannot delete __DEFAULT__ volume type" [Undecided,New] - Assigned to Rajat Dhasmana (whoami-rajat) | 15:04 |
| *** ykarel is now known as ykarel|away | 15:04 | |
| *** priteau has quit IRC | 15:04 | |
| hemna | fwiw, this fix is still dangling . https://review.opendev.org/#/c/736404/ | 15:04 |
| rosmaita | geguileo: whoami-rajat: one option for both functionalities? | 15:05 |
| *** Xuchu has quit IRC | 15:05 | |
| hemna | why would you want to delete the __DEFAULT__ volume type? it was created on purpose to be the fallback and is always supposed to be there | 15:05 |
| whoami-rajat | geguileo, also using the CONF.default_volume_type might be an issue because if some volumes use the __DEFAULT__ type and we've set the config then what exactly should we do | 15:05 |
| rosmaita | hemna: you need to read through the bug to see what the guy is really asking for | 15:06 |
| hemna | ok /me reads | 15:06 |
| rosmaita | :) | 15:06 |
| geguileo | hemna: the last comment is enough to understand the situation | 15:06 |
| whoami-rajat | rosmaita, i think so, should we allow using it if we don't allow listing it? | 15:06 |
| geguileo | rosmaita: yeah, just 1 config option to disable the __DEFAULT__ altogether | 15:07 |
| geguileo | rosmaita: that way we can check if we can do it on start | 15:07 |
| geguileo | as in check that there are no volumes using it | 15:07 |
| geguileo | and that it has a default value defined in cinder.conf | 15:07 |
| rosmaita | but here is the problem | 15:07 |
| hemna | so the problem is for deployments that have defined usable volume types, they don't want to see the __DEFAULT__ | 15:07 |
| rosmaita | suppose it is in use, but the operator no longer wants people to use it | 15:07 |
| geguileo | hemna: yup | 15:08 |
| geguileo | rosmaita: then they just need to retype them to the equivalent | 15:08 |
| geguileo | rosmaita: if the driver is properly implemented it will just be a DB change | 15:08 |
| rosmaita | "if the driver is properly implemente" | 15:08 |
| geguileo | so if __DEFAULT__ is equivalent to standard_iops, then they retype it to that | 15:09 |
| whoami-rajat | rosmaita, yeah, that was my concern as well, if volumes are using the __DEFAULT__ type and admin disables the config option, what should we do in this case? geguileo | 15:09 |
| geguileo | rosmaita: and if it's broken they go and edit the DB after they open a launchpad bug on the driver | 15:09 |
| hemna | so there is no way to hide it, and still use it | 15:09 |
| geguileo | whoami-rajat: we fail to start | 15:09 |
| smcginnis | Its use was for cases where someone doesn't provide a volume type to use. So hiding it makes the most sense to me. We don't want it to be something someone selects. It's an internal mechanism to because we can't get our code straight. | 15:09 |
| hemna | but also we migrated volumes that didn't have a type | 15:10 |
| hemna | well 'db migrated' | 15:10 |
| geguileo | hemna: ouch, true | 15:10 |
| smcginnis | It's a type that is no type. | 15:10 |
| hemna | well it's supposed to be an empty type | 15:10 |
| hemna | so can we just not show it in the api for get types ? | 15:11 |
| geguileo | smcginnis: unless the admin has set keys in its extra specs | 15:11 |
| hemna | but still show it where it's used? | 15:11 |
| rosmaita | except, that operators have the ability to update it | 15:11 |
| whoami-rajat | hemna, or we can update it as well to make it the actual default | 15:11 |
| hemna | yuhk | 15:11 |
| rosmaita | yeah | 15:11 |
| smcginnis | geguileo: We shouldn't allow that. It should be hidden. If they want a type with extra specs, they should define one. | 15:11 |
| hemna | I don't think operators should be able to add extra specs to it | 15:12 |
| geguileo | afaik we allow it | 15:12 |
| smcginnis | hemna: ++ | 15:12 |
| hemna | the intention was for it to be a placeholder re: empty | 15:12 |
| rosmaita | it seemed like such a simple change ... | 15:12 |
| smcginnis | I think that we allowed folks to modify it is a bug. | 15:12 |
| hemna | yah it should be locked | 15:12 |
| whoami-rajat | The original idea was not to allow updating it but there was a review comment stating we should allow it | 15:12 |
| hemna | and only visible to admins in cinder type-list | 15:13 |
| smcginnis | Bad call. :) | 15:13 |
| openstackgerrit | Gorka Eguileor proposed openstack/cinder master: Add cinder-manage command to remove file locks https://review.opendev.org/734144 | 15:13 |
| rosmaita | ok, i guess we all need to take an action item to go back and review the original spec to refresh ourselves on what we were trying to accomplish | 15:13 |
| rosmaita | and then read through today[s meeting logs | 15:13 |
| rosmaita | and figure out a good proposal for this | 15:13 |
| hemna | geguileo thanks man, much better :) | 15:13 |
| hemna | rosmaita, so do you disagree with smcginnis and myself on this? | 15:14 |
| hemna | make it immutable and hidden in cinder type-list | 15:15 |
| rosmaita | well, my memory is that i didn't want this thing to be modifiable, but there must have been a good reason for it | 15:15 |
| rosmaita | or my memory is just shot | 15:15 |
| hemna | anyone have the review handy? | 15:15 |
| whoami-rajat | https://review.opendev.org/#/q/topic:bp/untyped-volumes-default-volume-type+(status:open+OR+status:merged) | 15:16 |
| smcginnis | It's whole point was to be a safety because we thought it was too much work to go through our code and fix anywhere that we made the assumption that there would always be a volume type. So __DEFAULT__ should just be an empty type to make sure we don't blow up on one of those assumptions. | 15:16 |
| hemna | smcginnis++ | 15:16 |
| rosmaita | some dope approved the spec with this comment: "I think we're all in favor of doing this, so let's move it along and figure out whether __DEFAULT__ volume type should be mutable during the implementation phase. (See my question on PS9, Line 45)" | 15:17 |
| rosmaita | that was on https://review.opendev.org/#/c/651480/ | 15:18 |
| hemna | hrmm I don't see any discussion in the patch review for mutability https://review.opendev.org/#/c/639180/ | 15:19 |
| *** zhanglong has quit IRC | 15:19 | |
| geguileo | hemna: thank you for pointing out the confusing text | 15:20 |
| hemna | there was a comment about Eric not agreeing with immutability | 15:20 |
| hemna | looking for his comment about it | 15:20 |
| smcginnis | https://review.opendev.org/#/c/651480/5/specs/train/untyped-volumes-to-default-volume-type.rst@45 | 15:20 |
| rosmaita | my battery is at 3%, i need to sign off | 15:20 |
| *** rosmaita has left #openstack-cinder | 15:20 | |
| smcginnis | Saved by the battery. :) | 15:21 |
| hemna | https://review.opendev.org/#/c/651480/5/specs/train/untyped-volumes-to-default-volume-type.rst | 15:22 |
| hemna | I'm not sure eric was arguing about __DEFAULT__ volume type or default volume type | 15:23 |
| hemna | which can be 2 different things | 15:23 |
| hemna | that wasn't clear to me in his statement | 15:24 |
| hemna | regardless, I think __DEFAULT__ volume type should be immutable | 15:24 |
| whoami-rajat | https://review.opendev.org/#/c/651480/4/specs/train/untyped-volumes-to-default-volume-type.rst@44 | 15:24 |
| whoami-rajat | the original idea was not to allow updating it | 15:24 |
| hemna | if an operator has created another default volume type, that should be mutable. | 15:24 |
| *** enriquetaso has quit IRC | 15:24 | |
| hemna | so I think it's a bug that __DEFAULT__ is mutable then | 15:25 |
| whoami-rajat | i think eharney had some concerns making it immutable https://review.opendev.org/#/c/651480/5/specs/train/untyped-volumes-to-default-volume-type.rst@45 | 15:26 |
| smcginnis | I agree with hemna. It should be immutable. If they need a default with specific settings, then that is what setting a specific default volume type in cinder.conf is for. | 15:28 |
| hemna | whoami-rajat that's what I just mentioned just a second ago. I think eharney was talking about the default volume type, not __DEFAULT__ | 15:28 |
| hemna | 2 different things | 15:29 |
| smcginnis | ++ | 15:29 |
| hemna | default volume type being what the operator has created and set in cinder.conf | 15:29 |
| whoami-rajat | yeah but previously we didn't decide the name __DEFAULT__ so default volume type referred to __DEFAULT__ | 15:29 |
| openstackgerrit | Merged openstack/cinder stable/stein: 3PAR - Fix renaming volume after migration https://review.opendev.org/739416 | 15:30 |
| whoami-rajat | it also states "if Cinder creates this type, make it an empty type, therefore retypes are safe" vs. "it must be an empty type". | 15:30 |
| whoami-rajat | the idea was to have an empty type i.e. not being able to update it and i think Eric's concern was on the __DEFAULT__ type | 15:31 |
| hemna | I don't understand the concern with making __DEFAULT__ immutable | 15:33 |
| hemna | makes no sense to me | 15:33 |
| whoami-rajat | i don't have any concerns but there might be a reason to not do it in first place that I'm not able to recall | 15:34 |
| whoami-rajat | but again, i don't have any issue with doing it | 15:34 |
| smcginnis | It's a bug that we allow it to be modified. It's a bug that we show it to users through the API. | 15:35 |
| hemna | smcginnis++ | 15:37 |
| *** hamalq has joined #openstack-cinder | 15:44 | |
| *** hamalq_ has joined #openstack-cinder | 15:45 | |
| *** hamalq has quit IRC | 15:49 | |
| *** enriquetaso has joined #openstack-cinder | 15:51 | |
| *** ociuhandu_ has joined #openstack-cinder | 15:55 | |
| *** ociuhandu has quit IRC | 15:58 | |
| *** ociuhandu_ has quit IRC | 15:59 | |
| *** dviroel has quit IRC | 16:07 | |
| *** ociuhandu has joined #openstack-cinder | 16:12 | |
| *** knomura has quit IRC | 16:14 | |
| *** udesale_ has quit IRC | 16:16 | |
| *** ociuhandu has quit IRC | 16:18 | |
| *** knomura has joined #openstack-cinder | 16:21 | |
| openstackgerrit | Alex Deiter proposed openstack/cinder master: Creating snapshot on NFS backend fails https://review.opendev.org/739277 | 16:25 |
| *** gouthamr has quit IRC | 16:26 | |
| *** knomura has quit IRC | 16:26 | |
| *** ykarel|away has quit IRC | 16:46 | |
| *** whoami-rajat has quit IRC | 16:46 | |
| *** ociuhandu has joined #openstack-cinder | 16:47 | |
| *** gouthamr has joined #openstack-cinder | 16:52 | |
| *** markvoelker has joined #openstack-cinder | 16:53 | |
| *** ociuhandu has quit IRC | 16:53 | |
| *** markvoelker has quit IRC | 16:53 | |
| *** Lucas_Gray has quit IRC | 16:57 | |
| *** rosmaita has joined #openstack-cinder | 17:09 | |
| jungleboyj | smcginnis, eharney Why do I feel like we have been through this before? https://review.opendev.org/#/c/702810/8/specs/victoria/approved/on-demand-vm-quiescence.rst | 17:14 |
| eharney | jungleboyj: because it's a big feature gap that we should care about :) | 17:15 |
| jungleboyj | I thought we had talked about this before, but now that I think about it, it was for resizing volumes, not Quiesing | 17:16 |
| *** ociuhandu has joined #openstack-cinder | 17:16 | |
| *** ociuhandu has quit IRC | 17:23 | |
| smcginnis | That's a nova spec, but it might require some work on the Cinder end. | 18:00 |
| smcginnis | I know at least some storage that needs to be prepared ahead of the quiesce to flush intenal cache and things like that so the actual snap can complete in time before the host needs to unqueiese. | 18:00 |
| *** e0ne has quit IRC | 18:03 | |
| jungleboyj | Yeah, I know it was a Nova spec, but it goes along with something that has been requested on the Cinder side for quite some time as well. | 18:18 |
| jungleboyj | The Nova piece would need to happen first. | 18:18 |
| *** ociuhandu has joined #openstack-cinder | 18:35 | |
| *** ociuhandu has quit IRC | 18:41 | |
| openstackgerrit | Merged openstack/cinder master: Move make_initiator_target_all2all_map out of vutils https://review.opendev.org/729296 | 18:48 |
| openstackgerrit | Merged openstack/cinder master: tests: remove self.injected https://review.opendev.org/729343 | 18:48 |
| *** gmann_ has joined #openstack-cinder | 19:17 | |
| *** gmann_ is now known as gmann | 19:18 | |
| *** geguileo has quit IRC | 19:21 | |
| *** masayukig has joined #openstack-cinder | 19:29 | |
| *** arne_wiebalck_ has joined #openstack-cinder | 19:30 | |
| *** donnyd has joined #openstack-cinder | 19:31 | |
| *** mnaser has joined #openstack-cinder | 19:32 | |
| *** bcm has joined #openstack-cinder | 19:33 | |
| *** TheJulia has joined #openstack-cinder | 19:35 | |
| *** jungleboyj_ has joined #openstack-cinder | 19:35 | |
| *** kklimonda has joined #openstack-cinder | 19:36 | |
| *** arne_wiebalck_ is now known as arne_wiebalck | 19:41 | |
| *** _pewp_ has quit IRC | 19:51 | |
| *** hemna has quit IRC | 19:52 | |
| *** _pewp_ has joined #openstack-cinder | 19:52 | |
| *** vishalmanchanda has joined #openstack-cinder | 19:53 | |
| *** dviroel_ has joined #openstack-cinder | 19:58 | |
| *** hemna has joined #openstack-cinder | 19:59 | |
| *** mnasiadka has joined #openstack-cinder | 20:04 | |
| *** knikolla has joined #openstack-cinder | 20:05 | |
| *** carloss has joined #openstack-cinder | 20:06 | |
| *** xyang has joined #openstack-cinder | 20:09 | |
| *** johnsom_ has joined #openstack-cinder | 20:10 | |
| *** ajitha has joined #openstack-cinder | 20:12 | |
| *** lseki has quit IRC | 20:13 | |
| *** coreycb has joined #openstack-cinder | 20:14 | |
| *** walshh_ has joined #openstack-cinder | 20:14 | |
| *** simondodsley has joined #openstack-cinder | 20:14 | |
| *** lseki has joined #openstack-cinder | 20:15 | |
| *** patrickeast has joined #openstack-cinder | 20:17 | |
| *** vkmc_ has joined #openstack-cinder | 20:17 | |
| *** sri_ has joined #openstack-cinder | 20:17 | |
| *** NobodyCam has joined #openstack-cinder | 20:36 | |
| *** Lucas_Gray has joined #openstack-cinder | 20:52 | |
| *** mnaser is now known as mnaser|ic | 20:52 | |
| *** ajitha has quit IRC | 20:55 | |
| *** mnaser|ic has quit IRC | 21:08 | |
| *** mnaser|ic has joined #openstack-cinder | 21:08 | |
| *** mnaser|ic has quit IRC | 21:08 | |
| *** mnaser|ic has joined #openstack-cinder | 21:08 | |
| *** mnaser|ic is now known as vexxhost | 21:08 | |
| *** gouthamr has quit IRC | 21:09 | |
| *** gouthamr has joined #openstack-cinder | 21:11 | |
| *** sfernand has quit IRC | 21:12 | |
| *** vexxhost is now known as mnaser | 21:14 | |
| *** mnaser is now known as mnaser|ic | 21:14 | |
| *** gouthamr_ has joined #openstack-cinder | 21:15 | |
| *** johnsom_ is now known as johnsom | 21:19 | |
| *** johnsom has joined #openstack-cinder | 21:19 | |
| *** vishalmanchanda has quit IRC | 22:03 | |
| *** tosky has quit IRC | 22:30 | |
| openstackgerrit | Ajitha Robert proposed openstack/cinder master: Add multiattach in Nimble driver https://review.opendev.org/740125 | 22:55 |
| *** tkajinam has joined #openstack-cinder | 22:58 | |
| *** rcernin has joined #openstack-cinder | 23:01 | |
| *** hamalq_ has quit IRC | 23:09 | |
| *** hamalq has joined #openstack-cinder | 23:10 | |
| openstackgerrit | Merged openstack/cinder master: Add non-voting code coverage job https://review.opendev.org/738687 | 23:38 |
| *** mnaser|ic has quit IRC | 23:38 | |
| *** mnaser|ic has joined #openstack-cinder | 23:38 | |
| *** mnaser|ic has quit IRC | 23:38 | |
| *** mnaser|ic has joined #openstack-cinder | 23:38 | |
| *** mnaser|ic is now known as mnaser | 23:38 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!