Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2025-02-25
17:04:54 bauzas we're a cloud API, libvirt is an hypervisor
17:05:05 bauzas diffreent usecases
17:05:22 sean-k-mooney bauzas: today it gets updated on live migration...
17:05:41 sean-k-mooney so you update it in cinder and the new atachment used for live migation will get the new qos...
17:06:14 sean-k-mooney nova does not offically supprot that
17:06:26 sean-k-mooney i personally consier that a but but its apprely what happens today
17:06:27 bauzas we had events for cross-project communication, right?
17:06:36 sean-k-mooney we do it does nto use that
17:06:45 sean-k-mooney its because we only use those event on the volume
17:06:53 sean-k-mooney where as they are chging the qos on teh voluem type
17:06:58 bauzas anyway, looks we definitely need to design it at the PTG
17:07:09 sean-k-mooney basiclly they are updating there version fo a flavor and applying it to all instnaces
17:07:16 bauzas and I need to leave now, python meeting downtown
17:07:29 sean-k-mooney o/
17:09:11 MengyangZhang[m] thanks for all the discussion. sounds like we will continue this in PTG, let me know how to register it coz I think it already passed the registration deadline of Feb 21, no?
17:12:18 MengyangZhang[m] Oh i think i managed to register on openinfra site
17:16:07 opendevreview Stefan Hoffmann proposed openstack/os-vif master: [ovs] Add method to get external_ids from Interface https://review.opendev.org/c/openstack/os-vif/+/942715
17:35:32 vsaienko hello nova team, please add to your review queue. Fixes non working ironic serial console https://review.opendev.org/c/openstack/nova/+/942575
17:35:44 vsaienko already has +1 from Ironic folks
17:39:16 JayF and I think we should be able to backport it given the osdk fix is backporting, too
19:07:09 opendevreview Douglas Viroel proposed openstack/nova master: Add support for showing scheduler_hints in server details https://review.opendev.org/c/openstack/nova/+/938604
19:09:32 mikal Morning
19:16:54 opendevreview Merged openstack/nova master: Per-Property ImageMetaPropsWeigher https://review.opendev.org/c/openstack/nova/+/941601
20:07:24 opendevreview Douglas Viroel proposed openstack/nova master: DNM - Testing microversion 2.100 changes https://review.opendev.org/c/openstack/nova/+/942728
21:25:03 opendevreview Alan Bishop proposed openstack/nova master: Show swap_progress in attachment details https://review.opendev.org/c/openstack/nova/+/941476
#openstack-nova - 2025-02-26
00:16:07 sean-k-mooney[m] bauzas: can you review this in your morning and the repoducer patch above it https://review.opendev.org/c/openstack/nova/+/932931/2 i would like to make sure thats included in epoxy so we can start backporting it to stable ideally beforee bobcat goes eol.
00:17:30 opendevreview melanie witt proposed openstack/nova master: nova-manage: Add --max-rows option to 'db purge' command https://review.opendev.org/c/openstack/nova/+/927339
00:19:57 sean-k-mooney[m] so i think that is related to a donwstream bug report right ^ should the user really have to pass —max rows or should nova just do it in bataches automatically
00:21:41 melwitt sean-k-mooney: it is related to a downstream bug report but the bug reporter noted that what we need is the same ability we have for archive_deleted_rows (which allows someone to control what size they want the batches to be). --max-rows defaults to 1000
00:24:51 sean-k-mooney[m] i commented on the review as well but ya i noticed that defaulting when i looked more closely
00:25:30 sean-k-mooney[m] i guess that kind of indirectly does what i asked for so im ok with that i think
00:26:16 sean-k-mooney[m] basiclly it would be good for nova to hide this impelation detail form the user. its unlikely that archiveing 1000 instance would exceeed the transaction limit size
00:26:27 sean-k-mooney[m] posible but unlikely
00:26:40 sean-k-mooney[m] * purgeing not archiveing
00:27:25 sean-k-mooney[m] to me the fact a transaction size limit is a internal detail that a user of nova manage should ideallly not need to care about
00:27:26 melwitt fwiw the database transaction limit and packet size etc is all configurable on the database side so it's not something you can assume
00:27:46 sean-k-mooney[m] right but it is something we can look up
00:27:54 melwitt in my experience, operators do like to tune this to how they want, so I haven't thought about hiding bc of that
00:28:40 sean-k-mooney[m] im still of the opion that nova should not have shadow tables and archiving should be done externally to noav using other tooling :)
00:28:49 melwitt not really because we can't go looking into their innodb config or whatever
00:29:11 sean-k-mooney[m] you can make an sql query to look it up i touhgt
00:29:22 sean-k-mooney[m] the same way you can do that to get the connection limit
00:29:26 melwitt oh. I guess you could do that
00:30:59 sean-k-mooney[m] https://dev.mysql.com/doc/refman/8.4/en/group-replication-system-variables.html#sysvar_group_replication_transaction_size_limit
00:31:19 sean-k-mooney[m] melwitt that could be a future thing
00:31:28 melwitt I feel like that would be a separate thing where if we did it we would change archive to work the same way, at the same time. I think they should be consistent.
00:31:31 sean-k-mooney[m] the fact you providing a default kind of paper over the problem
00:32:06 sean-k-mooney[m] ya im not going to ask you to put it in the current patch fro that reason
00:32:14 melwitt and this patch is addressing the fact that today, if your shadow tables are too large, you are just screwed and you can't purge them with nova tooling and you have to truncate them manually
00:32:19 sean-k-mooney[m] i just wanted to raise that design point
00:32:36 melwitt fair enough :)
00:33:11 sean-k-mooney[m] yep so personally i think we should be deprecating the nova tooling and saying you should use some fo the external toolign for this
00:33:24 sean-k-mooney[m] but i know that wont happen any time soon
00:34:03 melwitt I think people might find that harsh. what kind of external tooling are you thinking of?
00:36:51 sean-k-mooney[m] there are two diffent set of tooling
00:37:22 sean-k-mooney[m] there is a tool that manages archiveing and purging records for all service in a common way inculdiign for serice that dont have shadow tables
00:38:04 sean-k-mooney[m] ill see if i can find the link but it handels nova neuron ectar all the same way with the ablity to archive them to a diffent database service instance or locally to files
00:38:13 melwitt oh, right, that does sound familiar
00:38:21 sean-k-mooney[m] there si allso osops tools which has a nova specifc srcipt https://opendev.org/openstack/osops/src/branch/master/tools/generic/nova/openstack_db_archive.sh
00:39:26 melwitt yeah, I see what you mean now. I agree I think that would be ideal for there to be a unified thing that will work for all of the projects
00:39:59 melwitt when you said "external" I thought you meant external to openstack and was confused
00:41:22 sean-k-mooney[m] https://github.com/ovh/osarchiver
00:41:34 sean-k-mooney[m] well i was thinking of ^
00:41:56 sean-k-mooney[m] thats the thing that basicaly does this for all proejcts in a consistent way
00:42:08 melwitt oh, interesting
00:42:59 sean-k-mooney[m] it basically does not use shadowtable at all
00:43:20 sean-k-mooney[m] it can arcive without deletign or it can archive and delete the rows in teh service dbs
00:43:27 sean-k-mooney[m] to get a similar effect
00:44:12 sean-k-mooney[m] given you can restore thing from the shadown tabels to the main tables i would actully say its more usefual because it mvoes the data to an entirly differnt db in teh same format
00:44:29 sean-k-mooney[m] which is you ever wanted to restore those rows back for any reason would be simpler
00:45:11 sean-k-mooney[m] its like doing incremental backups of only the rows marked as deleted
00:45:40 sean-k-mooney[m] the reason osarchiver is not hosted in opendev is because we didnt agree which repo to put it in
00:45:45 sean-k-mooney[m] which is a shame
00:45:50 melwitt yeah, I was just thinking that if people kept incremental backups there would be no use for shadow tables
00:46:14 melwitt hah, nice
00:46:53 sean-k-mooney[m] there was a maining list tread about it like 4 years ago i belvie and it just stalled out
00:47:29 melwitt reposhed
00:47:59 sean-k-mooney[m] melwitt: anyway i think that a better technial solution i just dont think we will ever get to that at least not unless something changes
00:48:29 sean-k-mooney[m] i wanted to use that in our new installer but…. it was too much effort to advocate for that and get by in
00:49:23 melwitt yeah. especially difficult if it's not in opendev
00:50:02 sean-k-mooney[m] it was more that pm didnt want use to focus on day two operations like automating backups of db in the operators
00:51:04 sean-k-mooney[m] so providing incremental backups and “archiving” of this in a centralised way was not a priority
00:51:16 sean-k-mooney[m] so we just use nova manage in a k8s cron job instead
00:51:39 sean-k-mooney[m] i.e. we keep is simpel using the upstream tooling
00:52:15 sean-k-mooney[m] im not saying it could never be in scope but it was not in the mvp list for v1
00:52:32 melwitt yeah
00:54:00 sean-k-mooney[m] ill try and add that to my review list for tomorrow
00:54:29 sean-k-mooney[m] its not nessisally impacted by ff but while i have context i might as well
00:55:26 melwitt haha thanks
00:55:48 sean-k-mooney[m] if you feel like reviewing something small this and the repoducer under neath is waiting for a second +2 https://review.opendev.org/c/openstack/nova/+/932931/2 although i pinged sylain to hopefully take a look above
00:57:01 sean-k-mooney[m] i think its time i went to sleep chat to you tomorrow o/
00:57:07 melwitt sylvain would be ideal for that one but if he doesn't get to it I can do it
00:57:13 melwitt seeya tomorrow o/
01:24:46 opendevreview Merged openstack/nova master: Fix device type when booting from ISO image https://review.opendev.org/c/openstack/nova/+/909611
01:39:28 opendevreview melanie witt proposed openstack/nova master: DNM test headers https://review.opendev.org/c/openstack/nova/+/942752
06:38:33 opendevreview Amit Uniyal proposed openstack/nova master: WIP: remove swap.disk from disk.info https://review.opendev.org/c/openstack/nova/+/942759
06:45:22 opendevreview melanie witt proposed openstack/nova master: DNM test headers https://review.opendev.org/c/openstack/nova/+/942752
08:50:33 opendevreview Merged openstack/nova master: Add repoducer test for bug 2074219 https://review.opendev.org/c/openstack/nova/+/932930
09:11:53 bauzas gibi: sean-k-mooney[m] : I'd appreciate if you could get a swing for https://review.opendev.org/c/openstack/nova/+/845757/
09:12:08 bauzas I'll also try to look at the functests failing

Earlier   Later