| *** xyang_ has joined #openstack-manila | 00:05 | |
| openstackgerrit | Vu Cong Tuan proposed openstack/manila master: Replace assertEqual([], items) with assertEmpty(items) https://review.openstack.org/473314 | 00:26 |
|---|---|---|
| *** hongbin has quit IRC | 00:54 | |
| *** markstur has quit IRC | 01:09 | |
| *** markstur has joined #openstack-manila | 01:10 | |
| *** ianychoi has joined #openstack-manila | 01:14 | |
| *** markstur has quit IRC | 01:14 | |
| *** markstur has joined #openstack-manila | 01:31 | |
| *** markstur has quit IRC | 01:35 | |
| openstackgerrit | zhongjun proposed openstack/manila master: Add like filter https://review.openstack.org/462468 | 01:59 |
| *** dsariel has quit IRC | 02:05 | |
| *** sapcc-bot1 has joined #openstack-manila | 02:10 | |
| *** toabctl has quit IRC | 02:11 | |
| *** toabctl has joined #openstack-manila | 02:12 | |
| *** markstur has joined #openstack-manila | 02:12 | |
| *** sapcc-bot has quit IRC | 02:14 | |
| *** markstur has quit IRC | 02:17 | |
| *** dsariel has joined #openstack-manila | 02:28 | |
| *** xyang_ has quit IRC | 02:33 | |
| *** markstur has joined #openstack-manila | 02:54 | |
| *** tuanluong has joined #openstack-manila | 02:56 | |
| *** markstur has quit IRC | 02:58 | |
| *** markstur has joined #openstack-manila | 03:35 | |
| *** markstur has quit IRC | 03:37 | |
| *** markstur has joined #openstack-manila | 03:37 | |
| *** arnewiebalck_ has joined #openstack-manila | 04:07 | |
| *** gouthamr_ has joined #openstack-manila | 04:07 | |
| *** netapp has quit IRC | 04:07 | |
| *** netapp has joined #openstack-manila | 04:08 | |
| *** gouthamr has quit IRC | 04:09 | |
| *** hoonetorg has quit IRC | 04:11 | |
| *** hoonetorg has joined #openstack-manila | 04:21 | |
| *** gcb has quit IRC | 04:24 | |
| *** gcb has joined #openstack-manila | 04:24 | |
| *** EmilienM has quit IRC | 04:30 | |
| openstackgerrit | Merged openstack/python-manilaclient master: Updated from global requirements https://review.openstack.org/473034 | 04:32 |
| *** EmilienM has joined #openstack-manila | 04:38 | |
| *** arnewiebalck_ has quit IRC | 05:13 | |
| *** rraja has joined #openstack-manila | 05:15 | |
| *** gouthamr_ has quit IRC | 05:29 | |
| *** markstur has quit IRC | 05:37 | |
| *** markstur has joined #openstack-manila | 05:58 | |
| *** markstur has quit IRC | 06:02 | |
| *** jprovazn has joined #openstack-manila | 06:22 | |
| *** rraja has quit IRC | 06:51 | |
| *** rraja has joined #openstack-manila | 06:52 | |
| openstackgerrit | zhongjun proposed openstack/manila master: Enable IPv6 in manila(network plugins and drivers) https://review.openstack.org/406776 | 07:00 |
| *** a-pugachev has joined #openstack-manila | 07:03 | |
| *** a-pugachev has quit IRC | 07:23 | |
| *** gcb has quit IRC | 07:34 | |
| *** gcb has joined #openstack-manila | 07:37 | |
| openstackgerrit | zhongjun proposed openstack/manila master: Add export-location filter in share and share instance list API https://review.openstack.org/461712 | 07:43 |
| *** ianychoi has quit IRC | 08:43 | |
| *** lpetrut has joined #openstack-manila | 08:50 | |
| *** a-pugachev has joined #openstack-manila | 09:09 | |
| *** jprovazn has quit IRC | 09:23 | |
| openstackgerrit | zhongjun proposed openstack/manila master: Enable IPv6 in manila(network plugins and drivers) https://review.openstack.org/406776 | 09:25 |
| *** markstur has joined #openstack-manila | 09:36 | |
| *** markstur has quit IRC | 09:41 | |
| *** jprovazn has joined #openstack-manila | 09:44 | |
| *** ianychoi has joined #openstack-manila | 09:49 | |
| *** markstur has joined #openstack-manila | 10:13 | |
| *** markstur has quit IRC | 10:17 | |
| openstackgerrit | Valeriy Ponomaryov proposed openstack/manila master: Add share groups quotas https://review.openstack.org/473464 | 10:19 |
| *** markstur has joined #openstack-manila | 10:50 | |
| *** ganso has joined #openstack-manila | 10:50 | |
| *** markstur has quit IRC | 10:54 | |
| *** tuanluong has quit IRC | 11:16 | |
| *** markstur has joined #openstack-manila | 11:27 | |
| *** makowals_ has joined #openstack-manila | 11:27 | |
| *** makowals has quit IRC | 11:28 | |
| *** markstur has quit IRC | 11:31 | |
| *** a-pugachev has quit IRC | 11:45 | |
| *** a-pugachev has joined #openstack-manila | 11:53 | |
| *** markstur has joined #openstack-manila | 12:04 | |
| *** markstur has quit IRC | 12:09 | |
| *** chlong has quit IRC | 12:20 | |
| *** markstur has joined #openstack-manila | 12:25 | |
| *** markstur has quit IRC | 12:29 | |
| openstackgerrit | Merged openstack/manila master: [Docs] Correct glusterfs references https://review.openstack.org/473787 | 12:32 |
| *** gouthamr has joined #openstack-manila | 12:44 | |
| *** catintheroof has joined #openstack-manila | 12:52 | |
| *** markstur has joined #openstack-manila | 13:02 | |
| *** markstur has quit IRC | 13:06 | |
| openstackgerrit | Valeriy Ponomaryov proposed openstack/python-manilaclient master: Add share group quotas https://review.openstack.org/473482 | 13:09 |
| *** gcb has quit IRC | 13:14 | |
| *** xyang1 has joined #openstack-manila | 13:14 | |
| *** dsariel has quit IRC | 13:16 | |
| *** openstackgerrit has quit IRC | 13:18 | |
| *** eharney has joined #openstack-manila | 13:24 | |
| *** sapcc-bot1 has quit IRC | 13:36 | |
| *** carthaca_ has quit IRC | 13:36 | |
| *** carthaca_ has joined #openstack-manila | 13:36 | |
| *** sapcc-bot has joined #openstack-manila | 13:36 | |
| *** markstur has joined #openstack-manila | 13:38 | |
| *** markstur has quit IRC | 13:43 | |
| *** a-pugachev has quit IRC | 13:49 | |
| *** a-pugachev has joined #openstack-manila | 14:00 | |
| *** chlong has joined #openstack-manila | 14:02 | |
| *** cknight has joined #openstack-manila | 14:02 | |
| *** chlong has quit IRC | 14:06 | |
| *** xyang_ has joined #openstack-manila | 14:08 | |
| *** dsariel has joined #openstack-manila | 14:10 | |
| *** dustins has joined #openstack-manila | 14:10 | |
| *** eharney has quit IRC | 14:13 | |
| *** markstur has joined #openstack-manila | 14:15 | |
| *** eharney has joined #openstack-manila | 14:16 | |
| *** markstur has quit IRC | 14:20 | |
| *** openstackgerrit has joined #openstack-manila | 14:28 | |
| openstackgerrit | Merged openstack/manila master: Replace assertEqual([], items) with assertEmpty(items) https://review.openstack.org/473314 | 14:28 |
| *** chlong has joined #openstack-manila | 14:34 | |
| *** chlong has quit IRC | 14:35 | |
| *** cfey has quit IRC | 14:36 | |
| *** hongbin has joined #openstack-manila | 14:37 | |
| *** chlong has joined #openstack-manila | 14:38 | |
| *** markstur has joined #openstack-manila | 14:52 | |
| *** markstur has quit IRC | 14:56 | |
| *** markstur has joined #openstack-manila | 14:57 | |
| *** markstur has quit IRC | 14:57 | |
| *** markstur has joined #openstack-manila | 14:58 | |
| *** rraja has quit IRC | 15:06 | |
| xyang1 | bswartz: ping | 15:11 |
| bswartz | pong | 15:12 |
| xyang1 | bswartz: can you please take a look of the comments on https://review.openstack.org/#/c/473864/ | 15:12 |
| bswartz | yes I saw that this morning | 15:12 |
| bswartz | we need to discuss among the various affected driver authors how we want to address the problem | 15:12 |
| bswartz | I'm uncertain which approach I prefer | 15:13 |
| xyang_ | bswartz: every driver that I see so far needs the id of ShareSnapshotInstance. It's just the way how they referenced it is different | 15:14 |
| xyang_ | bswartz: some use share_id, some use share_name, some directly use share_instance_id in ShareSnapshotInstance, but they are all looking for id of ShareInstance | 15:15 |
| openstackgerrit | Jan Provaznik proposed openstack/manila master: User Messages https://review.openstack.org/313549 | 15:15 |
| bswartz | yes that's what they should do, unless it's migration/replication code | 15:15 |
| xyang_ | bswartz: I don't know which driver really needs the id of the Share | 15:16 |
| gouthamr | even when a driver is performing migration or replication, what's the need to reference the instance id? | 15:16 |
| bswartz | some things are tied to the share ID rather than the instance | 15:16 |
| bswartz | like access rules | 15:16 |
| gouthamr | we designed the driver interfaces such that they look like different resources altogether | 15:16 |
| xyang_ | I am not that familiar with what is needed in migration and replication, I'm focusing on create_snapshot here | 15:17 |
| xyang_ | In create_share, drivers use share['id'] or share['name'] passed in which is based on the id of ShareInstance, not Share | 15:17 |
| gouthamr | yep, sorry i didn't dig into this review earlier.. -> snapshot['share_id'] was always the id of the share instance.. and that changed breaking drivers | 15:18 |
| xyang_ | In create_snapshot, they need to find the id of the ShareSnapshotInstance so they can create snapshot | 15:19 |
| xyang_ | gouthamr: yes | 15:19 |
| openstackgerrit | Valeriy Ponomaryov proposed openstack/python-manilaclient master: Add share group quotas https://review.openstack.org/473482 | 15:20 |
| xyang_ | so I don't know how this current fix in https://review.openstack.org/#/c/473864/ is helpful if no driver is really looking for the id of Share. Drivers need id of ShareInstance, not id of Share | 15:20 |
| xyang_ | we could either get rid of the share_id from ShareSnapshotInstance completely and have every driver references share_instance_id instead, or have share_id in ShareSnapshotInstance still refer to id of ShareInstance to maintain the backward compatibility that was there before | 15:22 |
| xyang_ | share_name in ShareSnapshotInstance still works as it is still based on id of ShareInstance | 15:23 |
| xyang_ | so if we decide to get rid of share_id because the name is confusing, we should get rid of share_name too | 15:24 |
| vponomaryov | xyang_: moving of "share_id" is a result of squashing of DB tables | 15:27 |
| vponomaryov | xyang_: removed table did have it | 15:27 |
| vponomaryov | xyang_: so, it is dictated by incompatibility caused by existing workaround for "nothing" | 15:27 |
| vponomaryov | xyang_: each share driver receives "instances" in its interfaces | 15:28 |
| vponomaryov | xyang_: what problem do you see in updating existing affected drivers to use correct attr? | 15:29 |
| vponomaryov | xyang_: like ZFSonLinux driver has been doing all the time | 15:29 |
| xyang_ | vponomaryov: have drivers use share_instance_id directly? | 15:29 |
| vponomaryov | xyang_: moreover, it uses both values for storing appropriate info for replication needs | 15:30 |
| vponomaryov | xyang_: yes | 15:30 |
| xyang_ | vponomaryov: that was the proposal I gave during our discussion with Ben on IRC. He rejected it | 15:30 |
| xyang_ | vponomaryov: He said driver should not have knowledge of share_instance_id | 15:31 |
| vponomaryov | xyang_: Ben can be wrong as any other person in the World | 15:31 |
| vponomaryov | xyang_: then, he is not sure anymore ) | 15:31 |
| xyang_ | vponomaryov: your current fix has not fixed the problem that the bug was opened for | 15:31 |
| vponomaryov | xyang_: and third, I want see clear arguments for not doing it in the proper way | 15:31 |
| vponomaryov | xyang_: false, bug as about "not populated", it is POPULATED now | 15:32 |
| vponomaryov | xyang_:so, the BUG is fixed, it is drivers that were dependent on bug are not fixed | 15:32 |
| vponomaryov | s/bug as/bug is/ | 15:33 |
| xyang_ | vponomaryov: populated with undesired value is not a fix | 15:33 |
| vponomaryov | xyang_: "undesired by you", "desired by me" | 15:33 |
| vponomaryov | ^_^ | 15:33 |
| vponomaryov | xyang_: I 2 arguments for it | 15:33 |
| vponomaryov | that I have mentioned several times | 15:33 |
| vponomaryov | s/I 2/I have 2/ | 15:34 |
| xyang_ | vponomaryov: undesired by drivers affected by the bug. that's not just Dell EMC drivers | 15:34 |
| vponomaryov | xyang_: again, "what problem do you see in updating existing affected drivers to use correct attr?" | 15:34 |
| xyang_ | vponomaryov: I have already answered that earlier | 15:35 |
| vponomaryov | because Ben restricted it? | 15:35 |
| vponomaryov | xyang_: in case you want to keep old "incorrect" behavior in drivers, that does not really have any value, we should revert back "DB tables squash" back and do not squash them tables | 15:37 |
| vponomaryov | s/them/these/ | 15:38 |
| vponomaryov | xyang_: if you want, I can fix EMC drivers in addition to common logic fix | 15:39 |
| vponomaryov | xyang_: and it will work just fine | 15:39 |
| xyang_ | vponomaryov: that's a small fix to use share_instance_id. we have the code, but didn't submit because I thought your fix in common code will fix the bug | 15:41 |
| xyang_ | vponomaryov: I want to see what Ben decide on this | 15:41 |
| *** dsariel has quit IRC | 15:42 | |
| vponomaryov | xyang_: what do you think personally? | 15:42 |
| bswartz | sorry guys I haven't read all of the scrollback | 15:42 |
| bswartz | I think we have deeper issues with our driver interface, and this problem is an example of why we should fix them | 15:42 |
| bswartz | the problem is that a proper cleanup of the interface might take a lot of effort | 15:43 |
| vponomaryov | bswartz: rename arguments in driver interfaces | 15:43 |
| bswartz | vponomaryov: according to gouthamr, nothing in the driver currently needs any real share ID or snapshot ID | 15:43 |
| xyang_ | vponomaryov: use share_instance_id is a straight forward fix. I don't have problem with that. but if share_id in ShareSnapshotInstance does not point to share_instance_id any more, I don't see it is even needed there | 15:43 |
| bswartz | they're all using the instance IDs -- even replication and migraiton | 15:43 |
| bswartz | if you can find examples of places in a driver somewhere that needs a real snapshot ID or share ID, that would be a help support a need to rename these fields | 15:44 |
| vponomaryov | xyang_, bswartz: share_groups logic uses this attr | 15:44 |
| gouthamr | yes, as far as i can tell.. snapshot['share_instance_id'] makes no sense | 15:44 |
| bswartz | otherwise I think it's simpler to continue to hide the real share and snapshot IDs | 15:45 |
| gouthamr | when the driver doesn't do anything with 'share_instance_id' yet | 15:45 |
| bswartz | vponomaryov: that's new code though -- we could change it to match the current style | 15:45 |
| bswartz | I would strongly prefer to have a driver interface that never exposed share IDs and snapshot IDs to drivers | 15:45 |
| vponomaryov | gouthamr: not "snapshot['share_instance_id']", it is "snapshot_instance['share_instance_id']" | 15:45 |
| gouthamr | ouch.. why should teh drivers worry about snapshot instances now? | 15:46 |
| bswartz | the rationale for hiding the real IDs from the drivers back when we did that work (in liberty?) was that we wanted to reserve the right to change them without affecting driver operation | 15:46 |
| vponomaryov | sure, we can also remove all references to "share_id" from share groups logic and return back "proxy" attr | 15:47 |
| gouthamr | vponomaryov: snapshot_replicas were used to hide that fact.. | 15:47 |
| gouthamr | a snapshot_replica is a snapshot_instance.. | 15:47 |
| gouthamr | it's easier to consume when you just assume that they are all the same as snapshots | 15:47 |
| bswartz | I'm in favor of continuing to use the proxy attr in the short term | 15:47 |
| bswartz | longer term we should not be passing sqlalchemy objects down to the drivers at all | 15:48 |
| vponomaryov | bswartz: but it will not block availability of share and snapshot IDs to driver | 15:48 |
| bswartz | we should be passing down some kind of synthetic objects -- at one time we had planned to use OVO -- but I would be okay with something else like simple dictionaries | 15:48 |
| bswartz | vponomaryov: if we wanted to block the availability of those fields to the drivers, how would you do it? | 15:49 |
| vponomaryov | bswartz: remove it from objects before providing it to drivers | 15:49 |
| gouthamr | the rationale behind passing primitive dicts in the share manager was something else too... sqlalchemy lazy loading | 15:50 |
| vponomaryov | bswartz: each share instance already has "share_id" | 15:50 |
| bswartz | wouldn't converting the sqlalchemy model objects to plain dictionaries achieve that? | 15:50 |
| vponomaryov | bswartz: BUT | 15:50 |
| vponomaryov | bswartz: it is really useful for replication needs, for example, in ZFSonLinux driver | 15:50 |
| gouthamr | vponomaryov: and other drivers too | 15:50 |
| vponomaryov | bswartz: share facade data and share instance data are different data and both existing. So, share ID and snapshot ID should be available for drivers | 15:51 |
| bswartz | I would hope that we could implement replication as a list of replicas with no knowledge of the actual Share ID those replicas belong to | 15:51 |
| vponomaryov | sorry, got to go | 15:51 |
| bswartz | are there places in the ZFS driver today that depend on share ID? | 15:51 |
| bswartz | I have to go to | 15:51 |
| vponomaryov | bswartz: yes | 15:51 |
| gouthamr | we can table this to tomorrow's meeting | 15:52 |
| bswartz | okay I'd like to find all of those places | 15:52 |
| bswartz | because those are at the core of this issue | 15:52 |
| bswartz | bbiab | 15:52 |
| bswartz | whew, made it back before cinder meeting | 15:59 |
| *** gouthamr_ has joined #openstack-manila | 16:00 | |
| *** gouthamr2 has joined #openstack-manila | 16:04 | |
| *** gouthamr_ has quit IRC | 16:08 | |
| *** a-pugachev has quit IRC | 16:11 | |
| *** cfey_ has joined #openstack-manila | 16:18 | |
| *** dustins has quit IRC | 16:19 | |
| *** dustins_ has joined #openstack-manila | 16:19 | |
| *** jmlowe has quit IRC | 16:23 | |
| *** arnewiebalck_ has joined #openstack-manila | 16:24 | |
| *** makowals has joined #openstack-manila | 16:35 | |
| *** makowals_ has quit IRC | 16:36 | |
| *** makowals_ has joined #openstack-manila | 16:41 | |
| *** makowals has quit IRC | 16:43 | |
| *** jmlowe has joined #openstack-manila | 16:46 | |
| *** makowals_ has quit IRC | 16:53 | |
| *** gouthamr_ has joined #openstack-manila | 16:55 | |
| *** makowals has joined #openstack-manila | 16:55 | |
| *** gouthamr_mbl has joined #openstack-manila | 16:58 | |
| *** gouthamr2 has quit IRC | 16:58 | |
| *** makowals has quit IRC | 17:00 | |
| *** gouthamr_ has quit IRC | 17:01 | |
| *** catintheroof has quit IRC | 17:02 | |
| *** catintheroof has joined #openstack-manila | 17:03 | |
| *** lpetrut has quit IRC | 17:10 | |
| *** pcaruana has quit IRC | 17:29 | |
| *** gouthamr_mbl has quit IRC | 17:30 | |
| *** gouthamr_ has joined #openstack-manila | 17:30 | |
| *** gouthamr_ has quit IRC | 17:34 | |
| *** makowals has joined #openstack-manila | 17:34 | |
| *** makowals has quit IRC | 17:40 | |
| *** makowals has joined #openstack-manila | 17:41 | |
| *** makowals has quit IRC | 17:41 | |
| *** makowals has joined #openstack-manila | 17:42 | |
| *** arnewiebalck_ has quit IRC | 17:46 | |
| *** makowals has quit IRC | 17:59 | |
| *** arnewiebalck_ has joined #openstack-manila | 18:30 | |
| *** arnewiebalck_ has quit IRC | 18:58 | |
| *** dsariel has joined #openstack-manila | 19:10 | |
| *** a-pugachev has joined #openstack-manila | 19:12 | |
| *** a-pugachev has quit IRC | 19:35 | |
| *** dustins_ is now known as dustins | 19:52 | |
| *** jprovazn has quit IRC | 19:58 | |
| *** xyang_ has quit IRC | 20:03 | |
| *** xyang_ has joined #openstack-manila | 20:04 | |
| *** xyang_ has quit IRC | 20:06 | |
| *** xyang1 has quit IRC | 20:18 | |
| *** gouthamr has quit IRC | 20:45 | |
| *** cknight has quit IRC | 20:48 | |
| *** jmlowe has quit IRC | 21:05 | |
| *** gouthamr has joined #openstack-manila | 21:11 | |
| *** eharney has quit IRC | 21:16 | |
| *** a-pugachev has joined #openstack-manila | 21:24 | |
| *** dustins has quit IRC | 21:31 | |
| *** hoonetorg has quit IRC | 21:34 | |
| *** jmlowe has joined #openstack-manila | 21:40 | |
| *** ganso has quit IRC | 22:15 | |
| *** chlong has quit IRC | 22:37 | |
| *** a-pugachev has quit IRC | 23:03 | |
| *** catintheroof has quit IRC | 23:09 | |
| *** dsariel has quit IRC | 23:12 | |
| *** ghs39ghsggf has joined #openstack-manila | 23:20 | |
| *** ghs39ghsggf has left #openstack-manila | 23:20 | |
| *** hongbin has quit IRC | 23:46 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!