opendevreview | Merged openstack/cinder master: Add missing microversion entry for 3.67 https://review.opendev.org/c/openstack/cinder/+/830669 | 00:42 |
---|---|---|
opendevreview | Merged openstack/cinder master: Fix request id mess on Cinder-API https://review.opendev.org/c/openstack/cinder/+/827860 | 01:31 |
TusharTgite | hi, core reviewers need a workflow +1 on this https://review.opendev.org/c/openstack/cinder-tempest-plugin/+/828229 | 04:41 |
opendevreview | Atsushi Kawai proposed openstack/cinder master: Hitachi: Add port scheduler https://review.opendev.org/c/openstack/cinder/+/828696 | 08:07 |
opendevreview | Mounika Sreeram proposed openstack/cinder master: [SVf] Resize of GMCV volumes in group https://review.opendev.org/c/openstack/cinder/+/830787 | 08:47 |
opendevreview | Francesco Pantano proposed openstack/devstack-plugin-ceph master: Deploy with cephadm https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/826484 | 08:52 |
opendevreview | Francesco Pantano proposed openstack/devstack-plugin-ceph master: Deploy with cephadm https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/826484 | 08:56 |
opendevreview | Francesco Pantano proposed openstack/devstack-plugin-ceph master: Deploy with cephadm https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/826484 | 09:47 |
opendevreview | Rajat Dhasmana proposed openstack/cinder master: Support volume re-image https://review.opendev.org/c/openstack/cinder/+/606346 | 10:01 |
opendevreview | Lior Friedman proposed openstack/os-brick master: Add NVMeoF Multipathing support https://review.opendev.org/c/openstack/os-brick/+/830800 | 10:19 |
opendevreview | Rajat Dhasmana proposed openstack/python-cinderclient master: Add volume reimage command https://review.opendev.org/c/openstack/python-cinderclient/+/606891 | 10:38 |
opendevreview | Rajat Dhasmana proposed openstack/python-cinderclient master: Add volume reimage command https://review.opendev.org/c/openstack/python-cinderclient/+/606891 | 10:48 |
*** dviroel_ is now known as dviroel | 11:28 | |
*** bhagyashris_ is now known as bhagyashris | 11:57 | |
opendevreview | Lior Friedman proposed openstack/os-brick master: Add NVMeoF Multipathing support https://review.opendev.org/c/openstack/os-brick/+/830800 | 12:32 |
*** carloss is now known as carloss|afk | 12:56 | |
yuval | hey | 13:02 |
yuval | ops | 13:03 |
yuval | good day | 13:03 |
TusharTgite_ | rosmaita: gentle reminder https://review.opendev.org/c/openstack/cinder/+/773985 | 13:18 |
rosmaita | TusharTgite_: tu | 13:18 |
rosmaita | i mean, ty | 13:18 |
jungleboyj | rosmaita: Good morning. Are you thinking that Cinder will be requesting Tuesday-Friday morning time at the PTG like we have done in the past? | 13:46 |
rosmaita | jungleboyj: whoami-rajat is handling arrangements ... he did say yesterday at the meeting that he proposes using our "usual" slots, and that if anyone had problems with the scheduling, they should contact him right away | 13:48 |
rosmaita | jungleboyj: https://etherpad.opendev.org/p/zed-ptg-cinder-planning | 13:48 |
jungleboyj | Ok. Trying to fill out a Doodle Poll for the TC. | 13:48 |
jungleboyj | Yeah, looks like he is looking at the usual times. | 13:49 |
rosmaita | yeah, he also said he will schedule 1300-1600 and leave 1600-1700 for slack space | 13:50 |
rosmaita | no one raised any objections yesterday, so i don't think there will be any changes | 13:50 |
*** dviroel is now known as dviroel|brb | 14:01 | |
*** carloss|afk is now known as carloss | 14:22 | |
*** dviroel|brb is now known as dviroel | 14:23 | |
opendevreview | Walt proposed openstack/cinder master: Bugfix: Account for consumed space better https://review.opendev.org/c/openstack/cinder/+/826510 | 14:27 |
hemna | can I get some eyes on that one? | 14:32 |
hemna | one of the problems I'm seeing in our deployments releated to scheduling, is that the allocated space can sometimes go negative. I believe this is due to 1) almost nobody is reporting provisioned_capacity_gb and then the volume manager manufactures allocated_capacity_gb from the volumes that are in the DB, but only accounts for volumes 'in-use' | 14:34 |
hemna | and 'available'. | 14:34 |
hemna | that skips a lot of volumes that are in other valid states | 14:35 |
TusharTgite_ | hi core reviewrs need a workflow +1 on this https://review.opendev.org/c/openstack/cinder-tempest-plugin/+/828229 | 14:37 |
hemna | one could argue that even volumes in error state could be counted as allocated space | 14:37 |
hemna | depending on how it got to that error state | 14:37 |
hemna | error creating is probably the only way it shouldn't be counted | 14:38 |
geguileo | hemna: I think I agree | 14:49 |
geguileo | hemna: we should probably just count everything that has the host field set | 14:50 |
geguileo | hemna: looking at your patch got me thinking... This is horribly broken for A/A | 14:50 |
hemna | the accounting and capacity filter is completely hosed I think and has been for ages | 14:59 |
hemna | especially after reading the definition of terms that rosmaita handed me yesterday | 14:59 |
hemna | the good thing is that I have a modified version of the https://github.com/openstack/cinder/blob/master/cinder/utils.py#L713-L747 which works for both thin and thick provisioning | 15:01 |
geguileo | hemna: I was referring specifically to how cinder volume services would report the allocated capacity in a clustered deployment, because they have different accounting on it depending on the volumes they created/deleted/extended | 15:01 |
hemna | the virtual free can tell the scheduler if there is space available correctly. it's being calculated differently in the CapacityWeigher vs CapacityFilter | 15:01 |
hemna | and I think they are both wrong | 15:02 |
hemna | geguileo ugh, so yah that's also bad | 15:02 |
hemna | I've been having to bounce cinder-volume services when I see the allocated capacity go negative | 15:02 |
hemna | and bouncing cinder leads to inconsistent volume states (attachment process) | 15:02 |
geguileo | ouch, ouch, ouch | 15:03 |
hemna | so, I think I'm going to update my other patch here: https://review.opendev.org/c/openstack/cinder/+/830456. to include my updated calculation | 15:04 |
hemna | it's a subtle change, but I think provides an accurate count as defined by the definition of terms | 15:04 |
hemna | https://gist.github.com/hemna/7300859555e18a0be4a6c8a99eb167a1 | 15:06 |
hemna | and then the filter scheduler should simply call that to see if the virtual free space - requested size > 0 | 15:07 |
hemna | as that code accounts for reserved % and over subscription ratio already baked in | 15:08 |
hemna | and then cinder will have a unified approach to calculating the space on a backend for the scheduler | 15:08 |
*** dviroel is now known as dviroel|lunvh | 16:03 | |
*** dviroel|lunvh is now known as dviroel|lunch | 16:03 | |
opendevreview | Merged openstack/cinder master: Support volume re-image https://review.opendev.org/c/openstack/cinder/+/606346 | 16:09 |
*** dviroel|lunch is now known as dviroel | 16:58 | |
opendevreview | Rajat Dhasmana proposed openstack/cinder master: Followup: Address review comments on re-image patch https://review.opendev.org/c/openstack/cinder/+/830879 | 17:32 |
whoami-rajat | rosmaita, jungleboyj followup patch for re-image ^^ | 17:34 |
opendevreview | Rajat Dhasmana proposed openstack/python-cinderclient master: Add volume reimage command https://review.opendev.org/c/openstack/python-cinderclient/+/606891 | 17:41 |
whoami-rajat | rosmaita, eharney ^^ addressed the comments | 17:42 |
opendevreview | Rajat Dhasmana proposed openstack/python-cinderclient master: Add volume reimage command https://review.opendev.org/c/openstack/python-cinderclient/+/606891 | 17:44 |
eharney | whoami-rajat: looks good, thanks | 18:54 |
whoami-rajat | thanks | 18:55 |
opendevreview | Brian Rosmaita proposed openstack/python-brick-cinderclient-ext master: Update requirements for Yoga https://review.opendev.org/c/openstack/python-brick-cinderclient-ext/+/830897 | 20:03 |
rosmaita | need reviews on this patch so we can release brick-cinderclient-ext: https://review.opendev.org/c/openstack/python-brick-cinderclient-ext/+/830897 | 20:30 |
rosmaita | e0ne eharney geguileo hemna jungleboyj smcginnis whoami-rajat enriquetaso ^^ | 20:30 |
hemna | ok | 20:30 |
rosmaita | thanks! | 20:31 |
eharney | rosmaita: done | 20:36 |
hemna | done | 20:37 |
*** dviroel is now known as dviroel|afk | 20:54 | |
opendevreview | Merged openstack/python-cinderclient master: Add volume reimage command https://review.opendev.org/c/openstack/python-cinderclient/+/606891 | 20:54 |
opendevreview | Hironori Shiina proposed openstack/cinder master: Set attach status of temporal volume during backup https://review.opendev.org/c/openstack/cinder/+/830901 | 20:55 |
opendevreview | Walt proposed openstack/cinder master: Bugfix: Account for consumed space better https://review.opendev.org/c/openstack/cinder/+/826510 | 20:56 |
opendevreview | Merged openstack/python-brick-cinderclient-ext master: Update requirements for Yoga https://review.opendev.org/c/openstack/python-brick-cinderclient-ext/+/830897 | 21:08 |
opendevreview | yuval proposed openstack/cinder master: lightos - bugfix compression stats should be True https://review.opendev.org/c/openstack/cinder/+/830538 | 21:11 |
opendevreview | Brian Rosmaita proposed openstack/python-cinderclient master: Prepare for Yoga cinderclient release https://review.opendev.org/c/openstack/python-cinderclient/+/830904 | 21:21 |
rosmaita | last patch needed before cinderclient release: https://review.opendev.org/c/openstack/python-cinderclient/+/830904 | 21:33 |
rosmaita | e0ne eharney geguileo hemna jungleboyj smcginnis whoami-rajat enriquetaso ^^ | 21:33 |
opendevreview | Merged openstack/python-cinderclient master: Prepare for Yoga cinderclient release https://review.opendev.org/c/openstack/python-cinderclient/+/830904 | 22:01 |
*** dviroel|afk is now known as dviroel | 23:37 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!