| *** chuanm7 is now known as chuanm | 00:12 | |
| *** chuanm7 is now known as chuanm | 01:01 | |
| *** chuanm0 is now known as chuanm | 01:17 | |
| *** chuanm6 is now known as chuanm | 11:04 | |
| *** chuanm9 is now known as chuanm | 12:55 | |
| jbernard | #startmeeting cinder | 14:02 |
|---|---|---|
| opendevmeet | Meeting started Wed Mar 8 14:02:42 2023 UTC and is due to finish in 60 minutes. The chair is jbernard. Information about MeetBot at http://wiki.debian.org/MeetBot. | 14:02 |
| opendevmeet | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 14:02 |
| opendevmeet | The meeting name has been set to 'cinder' | 14:02 |
| happystacker | hello | 14:02 |
| eharney | hi | 14:02 |
| jbernard | #roll call | 14:02 |
| jbernard | #topic roll call | 14:03 |
| Mounika | Hi | 14:03 |
| mubeen | Hi | 14:03 |
| jungleboyj | o/ | 14:03 |
| Sathya | hi | 14:03 |
| zaitcev | o/ | 14:03 |
| jbernard | hello everyone o/ , rajat is out at this time so I'm going to try to work the meetbot for today's meeting | 14:03 |
| felipe_rodrigues | o/ | 14:04 |
| HelenaDantas[m] | o/ | 14:04 |
| happystacker | o/ | 14:04 |
| thiagoalvoravel | o/ | 14:04 |
| MatheusAndrade[m] | o/ | 14:04 |
| caiquemello[m] | o/ | 14:04 |
| tobias-urdin | o/ | 14:05 |
| lucasmoliveira059 | \o | 14:05 |
| jbernard | ok, let's get started | 14:06 |
| jbernard | #topic annoucements | 14:06 |
| jbernard | Cinder RC1 has been released! | 14:06 |
| jbernard | the stable/2023.1 branch has been cut and master is now the 2023.2 (Bobcat) development branch | 14:06 |
| jbernard | There will be an RC-2, so continue to prioritize reviews on this etherpad: | 14:07 |
| jbernard | #link https://etherpad.opendev.org/p/cinder-antelope-fixes-rc | 14:07 |
| jbernard | patch authors: make sure the notes about your patch on ^^ are up to date on an ongoing basis, it affects whether someone reviews or waits for an update | 14:07 |
| jungleboyj | \o/ | 14:08 |
| jbernard | all patches must merge to master FIRST and then be BACKPORTED to stable/2023.1 | 14:08 |
| jbernard | patch authors: ^^ pay attention an propose the backport as soon as your patch has merged (do NOT propose earlier) | 14:08 |
| jbernard | for 2023.2 (Bobcat) release planning, | 14:09 |
| jbernard | We will conduct Cinder virtual PTG from 28th March (Tuesday) to 31st March (Friday), 2023 | 14:09 |
| jbernard | Timings will be 1300-1700 UTC -- This time slot has worked for us in the past several PTGs so keeping it same | 14:09 |
| jbernard | Please add your topics to the planning etherpad | 14:09 |
| jbernard | #link https://etherpad.opendev.org/p/bobcat-ptg-cinder-planning | 14:10 |
| rosmaita | o/ | 14:10 |
| jbernard | #topic Cinder retype for migration , passing the new volume type id to the drivers | 14:10 |
| jbernard | #link https://bugs.launchpad.net/cinder/+bug/2001619 | 14:11 |
| jbernard | ^ links to the previously raised bug | 14:11 |
| Sathya | Hi | 14:11 |
| Sathya | Eralier we had this discussion with geguileo, on cinder retype for migration, where volume type with additional extra specs | 14:11 |
| Sathya | should have a chance to enter driver specific code | 14:11 |
| Sathya | Now i have observed that the new volume type is not being sent to the driver thorugh | 14:12 |
| Sathya | driver.migrate_volume(ctxt,volume,host) | 14:12 |
| Sathya | where as new volume type id value is being passed to cinder generic migration | 14:12 |
| Sathya | self._migrate_volume_generic(ctxt, volume, host, new_type_id) | 14:12 |
| Sathya | I believe this is a important parameter to be passed, but since it's migrate_volume function defnition | 14:12 |
| Sathya | inclusion of this parameter might pose some issues to other drivers | 14:12 |
| geguileo | Sathya: I don't remember what I said in the conversation, but I vaguely remember drivers not needing that... | 14:13 |
| Sathya | calling to check driver whether it could handle extra specs | 14:14 |
| Sathya | https://review.opendev.org/c/openstack/cinder/+/869999 | 14:15 |
| Sathya | this is the gerrit link for our initial discussion | 14:16 |
| jbernard | Sathya: I think we can discuss this further in #openstack-cinder after the meeting | 14:20 |
| Sathya | sure | 14:20 |
| jbernard | do you need anything in addition to further design discussion, like a request for review, what would help most? | 14:20 |
| geguileo | just one last question | 14:21 |
| geguileo | Sathya: Why wasn't the migration implemented on the "retype" method of your driver? | 14:21 |
| Sathya | our backend has now come up with this new non disruptive migration | 14:22 |
| geguileo | why can't you do the migration on the retype? | 14:23 |
| geguileo | you driver is called on a retype when just the backend has changed | 14:23 |
| geguileo | Sathya: maybe the call to know if the driver can do the retype needs to be in the manager's retype instead of the migrate | 14:25 |
| Sathya | we had those additional extra specs, which was not entering driver code, that's why bug was raised | 14:25 |
| geguileo | Sathya: but you were focusing on the migrate part of the code | 14:26 |
| geguileo | and apparently I didn't think of this as a whole | 14:26 |
| geguileo | because it makes more sense that it's the "retype" method in the manager the one that checks with the driver if it can do the retype | 14:26 |
| Sathya | yes it's regarding cinder retype for migration | 14:27 |
| geguileo | so I think the fix should be in the manager's retype | 14:27 |
| Sathya | yes | 14:28 |
| geguileo | please have a look at that code and how this same idea of having the driver tell whether it can do it efficiently can be applied there | 14:28 |
| Sathya | i tried several approach , it's not suiting , can we have this discussion separately | 14:30 |
| geguileo | Sathya: I don't have time these days, so it's going to be hard for me to assist you on this one | 14:31 |
| Sathya | any further discussions is it possible in video call meetings | 14:33 |
| rosmaita | Sathya: put it on the agenda for the virtual PTG | 14:34 |
| Sathya | sure | 14:34 |
| jbernard | we do have our PTG meeting in video from March 28 to March 31, so that might be an option | 14:34 |
| Sathya | ok | 14:35 |
| rosmaita | #link https://etherpad.opendev.org/p/bobcat-ptg-cinder-planning | 14:35 |
| jbernard | thanks Sathya | 14:36 |
| jbernard | #topic review requests | 14:36 |
| jbernard | there are 2 for Dell PowerMax by happystacker | 14:36 |
| jbernard | #link https://review.opendev.org/c/openstack/cinder/+/797970 | 14:37 |
| jbernard | and | 14:37 |
| jbernard | #link https://review.opendev.org/c/openstack/cinder/+/858370 | 14:37 |
| jbernard | Fujitsu Eternus by inori | 14:38 |
| jbernard | #link https://review.opendev.org/c/openstack/cinder/+/847730 (feature add) | 14:38 |
| jbernard | ^ this one has a -1 from whoami-rajat to wait until stable/2023.1 is cut, but that has just happened | 14:38 |
| jbernard | and | 14:39 |
| jbernard | Increase size of volume image metadata values (drencrom) | 14:39 |
| jbernard | #link https://review.opendev.org/c/openstack/cinder/+/868485 (need one more +2) | 14:39 |
| jbernard | #topic open discussion | 14:39 |
| zaitcev | I'm also back with https://review.opendev.org/c/openstack/cinder/+/852654, with which you helped once, and whoami-rajat had to circle back because I'm dumb and missed one of his important comments. I want to hear two things: 1. is there a better name than volume_is_fresh and 2. is the new test okay. | 14:41 |
| zaitcev | Rajat pointed out that "fresh" is essentially synonymous with "new", and both means "freshly created". But I want to convey something else: that the volume is safe for a sparse restore, because it has no old data. | 14:43 |
| zaitcev | I'm almost desperate enough to name it "volume_sparse_ok". | 14:44 |
| rosmaita | how about volume_is_safe_for_sparse_restore ? | 14:44 |
| happystacker | lol | 14:44 |
| rosmaita | i'm serious, either that or a comment in the rpc/api.py restore() that explains what "fresh" means for that parameter | 14:44 |
| zaitcev | I like the meaning of it but it's inconveniently long. | 14:44 |
| rosmaita | well, that's what the \ is for! | 14:45 |
| rosmaita | either that or "volume_has_data" as a bool (though that would flip the logic) | 14:46 |
| zaitcev | Allright. | 14:47 |
| zaitcev | As for the test, I wanted to copy what volume service does with ddt, which looks elegant, but only works because the service has its own _get_cctxt. | 14:48 |
| jbernard | ok, this is specifically for test_rpcapi.py, i will need to look at this more closely as im not immediately sure | 14:52 |
| jbernard | #link https://review.opendev.org/c/openstack/cinder/+/852654 | 14:53 |
| jbernard | ^ this is an additional review request in case anyone is looking for more | 14:53 |
| jbernard | ok, unless there is antyhing else I think we can wrap up | 14:55 |
| jbernard | thanks everyone! | 14:57 |
| jbernard | #endmeeting | 14:57 |
| opendevmeet | Meeting ended Wed Mar 8 14:57:18 2023 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 14:57 |
| opendevmeet | Minutes: https://meetings.opendev.org/meetings/cinder/2023/cinder.2023-03-08-14.02.html | 14:57 |
| opendevmeet | Minutes (text): https://meetings.opendev.org/meetings/cinder/2023/cinder.2023-03-08-14.02.txt | 14:57 |
| opendevmeet | Log: https://meetings.opendev.org/meetings/cinder/2023/cinder.2023-03-08-14.02.log.html | 14:57 |
| *** chuanm2 is now known as chuanm | 15:42 | |
| *** chuanm0 is now known as chuanm | 16:35 | |
| *** chuanm3 is now known as chuanm | 17:38 | |
| *** chuanm5 is now known as chuanm | 19:05 | |
| *** chuanm6 is now known as chuanm | 20:23 | |
| *** chuanm6 is now known as chuanm | 23:04 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!