Wednesday, 2021-04-14

« Tuesday, 2021-04-13 Index Thursday, 2021-04-15 »
*** mlavalle has quit IRC00:02
*** martinkennelly has quit IRC00:13
*** gyee has quit IRC00:38
*** LinPeiWen has joined #openstack-nova00:49
melwittguilhermesp: you can just do the openstack server reboot --hard to regenerate the guest xml00:55
*** tkajinam has quit IRC00:55
*** tkajinam has joined #openstack-nova00:56
guilhermespindeed! thanks melwitt00:58
melwittdon't think you needed the undefine but I've been wrong before00:59
*** zzzeek has quit IRC01:07
*** zzzeek has joined #openstack-nova01:08
*** zzzeek has quit IRC01:33
*** swp20 has quit IRC01:33
*** zzzeek has joined #openstack-nova01:35
*** k_mouza has joined #openstack-nova01:50
*** k_mouza has quit IRC01:54
*** zzzeek has quit IRC01:54
*** artom has quit IRC01:56
*** zzzeek has joined #openstack-nova01:57
guilhermespmelwitt: i'm my case ( an openstack-helm deployment )  my nova compute pod wasnt starting for some reason after a power shutdown. All guests were present on the node but in shutoff state and refusing to start, causing the compute node to crash loop and never be UP02:26
guilhermespso using cli wasnt an option... after undefine all guests that the pod was able to start the compute and things got back to normal :) ( guests were recreated ofc )02:27
*** hamalq has quit IRC02:29
*** zzzeek has quit IRC02:55
*** zzzeek has joined #openstack-nova02:57
*** zzzeek has quit IRC03:30
*** wangzhh has joined #openstack-nova03:31
*** zzzeek has joined #openstack-nova03:31
*** rcernin has quit IRC03:54
*** rcernin has joined #openstack-nova04:00
*** terdei has quit IRC04:27
*** terdei has joined #openstack-nova04:30
*** ratailor has joined #openstack-nova04:39
*** vishalmanchanda has joined #openstack-nova04:45
*** whoami-rajat_ has joined #openstack-nova05:20
*** ralonsoh has joined #openstack-nova05:21
*** mdbooth has quit IRC05:22
*** mdbooth has joined #openstack-nova05:22
*** mkrai has joined #openstack-nova05:25
*** wangzhh has quit IRC05:41
*** tinwood has quit IRC06:22
*** hemanth_n has joined #openstack-nova06:24
*** waleedm_ has joined #openstack-nova06:24
*** tinwood has joined #openstack-nova06:25
*** csatari has quit IRC06:34
*** macz_ has joined #openstack-nova06:34
*** PrinzElvis has quit IRC06:34
*** knikolla has quit IRC06:34
*** knikolla has joined #openstack-nova06:36
*** csatari has joined #openstack-nova06:36
*** PrinzElvis has joined #openstack-nova06:37
*** macz_ has quit IRC06:38
openstackgerritzitoun proposed openstack/nova master: Explaining when a base image is considered unused to avoid confusion why some images aren't deleted.  https://review.opendev.org/c/openstack/nova/+/78605906:47
*** bbowen has quit IRC06:56
*** lpetrut has joined #openstack-nova06:57
*** mkrai has quit IRC07:03
*** dklyle has quit IRC07:10
*** luksky has joined #openstack-nova07:11
*** andrewbonney has joined #openstack-nova07:14
*** xinranwang has quit IRC07:20
*** ociuhandu has joined #openstack-nova07:27
*** k_mouza has joined #openstack-nova07:30
*** ociuhandu has quit IRC07:31
*** ociuhandu has joined #openstack-nova07:31
*** k_mouza has quit IRC07:34
*** ociuhandu has quit IRC07:41
*** ociuhandu has joined #openstack-nova07:42
*** ociuhandu has quit IRC07:44
*** ociuhandu has joined #openstack-nova07:44
*** waleedm_ has quit IRC07:48
*** waleedm_ has joined #openstack-nova07:48
*** tosky has joined #openstack-nova07:50
*** bbowen has joined #openstack-nova07:51
*** derekh has joined #openstack-nova07:54
*** rpittau|afk is now known as rpittau07:55
*** bbowen_ has joined #openstack-nova07:56
*** bbowen has quit IRC07:58
*** lucasagomes has joined #openstack-nova08:01
*** kashyap has joined #openstack-nova08:03
*** martinkennelly has joined #openstack-nova08:05
*** slaweq_ has joined #openstack-nova08:19
*** slaweq has quit IRC08:19
*** rcernin has quit IRC08:24
*** macz_ has joined #openstack-nova08:35
*** mkrai has joined #openstack-nova08:38
*** ricolin has quit IRC08:39
*** macz_ has quit IRC08:39
*** slaweq_ is now known as slaweq08:51
hemanth_nsean-k-mooney: hi, this is pending final review for a long time https://review.opendev.org/c/openstack/nova/+/761824 (rocky backport), this is not urgent so please review only when you get some time.. thanks09:08
*** rcernin has joined #openstack-nova09:12
*** hoonetorg has joined #openstack-nova09:13
*** dtantsur|afk is now known as dtantsur09:45
openstackgerritStephen Finucane proposed openstack/nova master: docs: Add man pages for 'nova-policy'  https://review.opendev.org/c/openstack/nova/+/72991809:51
*** swp20 has joined #openstack-nova09:57
*** mkrai has quit IRC10:01
*** whoami-rajat_ is now known as whoami-rajat10:17
openstackgerritBalazs Gibizer proposed openstack/nova-specs master: Change error code when vDPA is unsupported  https://review.opendev.org/c/openstack/nova-specs/+/78125910:30
*** ociuhandu has quit IRC10:38
*** mkrai has joined #openstack-nova10:45
*** Luzi has joined #openstack-nova10:46
openstackgerritMerged openstack/nova-specs master: Change error code when vDPA is unsupported  https://review.opendev.org/c/openstack/nova-specs/+/78125910:54
*** ociuhandu has joined #openstack-nova10:59
*** ociuhandu has quit IRC10:59
*** ociuhandu has joined #openstack-nova11:00
*** rcernin has quit IRC11:04
*** ociuhandu_ has joined #openstack-nova11:04
openstackgerritStephen Finucane proposed openstack/nova master: docs: Document location of nova.conf files  https://review.opendev.org/c/openstack/nova/+/78620211:04
*** ociuhandu has quit IRC11:08
*** tkajinam has quit IRC11:14
*** ratailor has quit IRC11:17
*** swp20 has quit IRC11:18
*** ociuhandu_ has quit IRC11:20
*** ociuhandu has joined #openstack-nova11:20
*** waleedm__ has joined #openstack-nova11:27
*** waleedm_ has quit IRC11:30
*** hemna has quit IRC11:39
*** hemna has joined #openstack-nova11:39
*** zoharm has joined #openstack-nova11:45
*** rcernin has joined #openstack-nova11:45
*** ociuhandu has quit IRC11:52
*** ociuhandu has joined #openstack-nova11:53
zigoHi. Could we get reviews on that trivial doc (config file help) patch? https://review.opendev.org/c/openstack/nova/+/78605911:55
*** rcernin has quit IRC11:59
*** ociuhandu has quit IRC12:02
gibizigo: done12:02
zigoThanks.12:02
*** ociuhandu has joined #openstack-nova12:03
*** ociuhandu has quit IRC12:08
*** waleedm__ has quit IRC12:10
*** ociuhandu has joined #openstack-nova12:14
*** ociuhandu has quit IRC12:17
*** ociuhandu has joined #openstack-nova12:18
stephenfinooh, we have our 23.0.0 tag \o/12:18
stephenfingibi: Does this mean we can start merging larger patches again now?12:18
gibistephenfin: yes it is12:18
sean-k-mooneystephenfin: the release was yesterday so yes12:18
* stephenfin has an alembic migration series _almost_ completed locally12:18
stephenfinhurrah12:18
gibisounds cool12:18
* stephenfin will try wrap this up and spend tomorrow on reviews so12:19
stephenfinsean-k-mooney: I fixed https://review.opendev.org/c/openstack/nova/+/729918/ and added a follow-up explaining where nova.conf comes from12:19
*** luksky has quit IRC12:20
stephenfinjust fyi12:20
*** luksky has joined #openstack-nova12:20
*** luksky has quit IRC12:21
*** luksky has joined #openstack-nova12:21
*** hemanth_n has quit IRC12:24
*** macz_ has joined #openstack-nova12:36
*** macz_ has quit IRC12:41
*** Luzi has quit IRC12:44
gibicore: there are two patches needed on master before we get deep into xena:13:01
gibihttps://review.opendev.org/c/openstack/nova/+/78217113:01
gibihttps://review.opendev.org/c/openstack/nova/+/77892313:01
gibiboth needs a second core13:01
openstackgerritJohannes Kulik proposed openstack/nova master: vmware: OptionValue.value cannot be Boolean  https://review.opendev.org/c/openstack/nova/+/78624613:12
*** tkajinam has joined #openstack-nova13:14
bauzasgibi: ack lookinf13:18
bauzaslol13:18
* bauzas just clicked and saw the changes :p13:18
gibi:D13:19
sean-k-mooneywe likely should look at doing the libvirt min microverion bumps early too13:19
sean-k-mooneybefore m1 in any case13:19
sean-k-mooneynot as urgent though13:20
sean-k-mooneyfor https://review.opendev.org/c/openstack/nova/+/782171/2/nova/objects/service.py#20513:20
sean-k-mooneyshould we really be removing the victoria alias13:20
sean-k-mooneyi would have just appended wallaby13:21
sean-k-mooneyotherwise what is the point of having a dict13:21
*** k_mouza has joined #openstack-nova13:24
gibisean-k-mooney: I can remove the dict. I thought originally that we need a list but the code only uses that latest one so the dict is not needed13:29
*** mdbooth has quit IRC13:29
gibibut we both need a service number and the name13:30
*** mdbooth has joined #openstack-nova13:31
sean-k-mooneyyep well we could leave it as is too it just feels a little clunky currently13:31
sean-k-mooneynot that i really want to open this can of worms but our pms keep brinign up the idea of having nova support more then n-1 computes13:32
sean-k-mooneyim partly wondering if we want this to be confiugrable at some point13:33
sean-k-mooneynot that i actully want to really support n-2 or n-313:33
sean-k-mooneyim happy we now codify the n-1 support13:33
*** ricolin has joined #openstack-nova13:34
gibisean-k-mooney: we cannot simply make the n-2 support configurable, as n-2 support means keeping the compatibilty code and service version checks longer in the code, so it needs in change of the behavior of the developers13:34
gibibut anyhow if we want to support a bigger version gap later then it make sense to keep the Victoria version mapping in the dict13:36
*** artom has joined #openstack-nova13:36
openstackgerritBalazs Gibizer proposed openstack/nova master: Update min supported service version for Xena.  https://review.opendev.org/c/openstack/nova/+/78217113:36
gibisean-k-mooney: ^^13:37
sean-k-mooneyya it is more work then that13:38
sean-k-mooneygibi: we always have a tension downstreeam13:38
sean-k-mooneyfrom osp 16 on we move to a FFU only model13:39
sean-k-mooneyso the only way to go from train is to FFU directly to wallaby when 17 comes out13:39
sean-k-mooneythat is an issue for a mulit sight multi cell deployment as all nodes need to be upgrade at once13:39
gibiI see13:40
sean-k-mooneyso the tension is between custoemr saying we want to do upgrades on a slower cadance and the fact that we can only support n-1 upstream13:40
sean-k-mooneywe have other customer that want new feature faster too so not all are the same obviously13:41
*** k_mouza has quit IRC13:47
*** hemanth_n has joined #openstack-nova13:47
*** hemanth_n has quit IRC13:52
*** k_mouza has joined #openstack-nova13:54
*** amodi has joined #openstack-nova14:06
noonedeadpunkHi there! Have a question regarding nova-conductor and rabbit... We had a deployment on V with enabled SSL for rabbitmq (but without trusted cert or rootca). And while upgrade to W, we disabled ssl for rabbitmq. However, conductor still tries to reach rabbit via ssl and fails with CERTIFICATE_VERIFY_FAILED (whic was the reason why ssl was disabled at the first place)14:15
noonedeadpunkAm I missing so nova-manage command to make conductor migrate and re-read config?14:16
noonedeadpunk*some14:16
*** slaweq has quit IRC14:16
noonedeadpunkas with same config code that's deployed just on W works nicely14:17
gibinoonedeadpunk: look at the cell_mappings table in your nova api database. I guess it still contains rabbit urls with ssl14:17
noonedeadpunkgibi: yep, that's absolutely true14:18
gibinoonedeadpunk: you can use nova-manage cell_v2 update_cell to make the adjustments14:19
*** slaweq has joined #openstack-nova14:20
*** k_mouza has quit IRC14:20
sean-k-mooneyunless your useing templated transport urls ya yuou likely need to update them14:21
noonedeadpunkok, thanks, that worked! Now I need to review wtf we do in https://opendev.org/openstack/openstack-ansible-os_nova/src/branch/master/tasks/nova_db_setup.yml and https://opendev.org/openstack/openstack-ansible-os_nova/src/branch/master/tasks/nova_db_post_setup.yml14:23
sean-k-mooneynoonedeadpunk: i can never recall if templated urls is a nova thing or ooo. i think its a nova feature14:24
* noonedeadpunk no idea what templated urls are14:26
noonedeadpunkhttps://docs.openstack.org/nova/latest/user/cells.html#template-urls-in-cell-mappings ?14:26
sean-k-mooneynoonedeadpunk: yes14:26
sean-k-mooneyit allows you to not sore passward and logins info in the db for rabbit14:27
sean-k-mooneynova baisically render teh full url from the template when it tires to connect to rabbit14:27
sean-k-mooneyfilling in the placeholder from it nova.conf14:27
sean-k-mooneythat way if you cahnge your rabbit password for example you dont need to update the db14:28
sean-k-mooneyjust the config files on the nodes14:28
noonedeadpunkeventually I think that's what I'm looking for:)14:28
sean-k-mooneyi think the tls options can be set that way too14:29
noonedeadpunkthat's kind of `fragment` I guess.14:30
sean-k-mooneyi think its in query14:30
sean-k-mooneybut maybe fragment14:31
sean-k-mooneyim trying to find an example transport url with tls enabeld to take a look14:31
noonedeadpunkThe only thing I didn't get now - how to define this variable. I mean where? In [DEFAULT] or [database]?14:31
noonedeadpunkyeah, you're right, it's query14:32
sean-k-mooneywell by default you are using https://docs.openstack.org/nova/latest/configuration/config.html#DEFAULT.transport_url14:32
sean-k-mooneyand just hardcoding eveything which is fine14:33
sean-k-mooneyi assume the ssl would come from https://docs.openstack.org/nova/latest/configuration/config.html#oslo_messaging_rabbit.ssl14:33
noonedeadpunkthat's exactly what I have atm14:33
noonedeadpunkwell, it also comes from query indeed, ie `transport_url = rabbit://nova:pasword@172.29.236.100:5671//nova?ssl=1`14:35
*** tesseract has joined #openstack-nova14:37
*** zoharm has quit IRC14:39
sean-k-mooneyi think you would set somehtin like this "{scheme}://{username}:{password}@{hostname}/{path}?{query}"14:41
sean-k-mooneyin the cell mapping14:41
noonedeadpunkBut um. I mean that in nova.conf I should still have regular transport_url ?14:42
noonedeadpunkSo it's transport_url got splited and auto substituted in case cell mapping contains variables??14:43
sean-k-mooneythat im not sure im trying to figure out14:43
sean-k-mooneyunfortuetly our downstream sos report replace it with transport_url=********* since it could have password so i dont have an example with tls enable to look at14:43
noonedeadpunkas eventually http://paste.openstack.org/show/804476/ is not working in nova.conf... (test env, so I'm not bothered about password there :p)14:44
sean-k-mooneythe unit test seam to be seetting the values in the nova.conf in the tansport url14:46
sean-k-mooneyhttps://github.com/openstack/nova/blame/50fdbc752a9ca9c31488140ef2997ed59d861a41/nova/tests/unit/objects/test_cell_mapping.py#L216-L23114:46
noonedeadpunkyep, looks like this way indeed14:49
sean-k-mooneynoonedeadpunk: i think ya you shoudl be still setting actula datat in the nova.conf14:50
*** macz_ has joined #openstack-nova14:50
*** k_mouza has joined #openstack-nova14:50
sean-k-mooneyand then we split the url and teake the info in the db and templeate it14:50
noonedeadpunkSo I see literally no point in not using templates under any conditions...14:50
*** hoonetorg has quit IRC14:50
*** haleyb has quit IRC14:50
noonedeadpunksean-k-mooney: thanks a lot!14:50
sean-k-mooneynoonedeadpunk: i think we always use the in ooo/OSP since it was added in rocky14:51
noonedeadpunkI think for us to migrate to it, we need to replace https://opendev.org/openstack/openstack-ansible-os_nova/src/branch/master/tasks/nova_db_setup.yml#L62 and add task to nova-manage cell_v2 update_cell during upgrades as far as I got the process?14:52
*** lpetrut has quit IRC15:00
melwittgibi: heya, in case you wanted to look at PS2 on the reject force up if evacuations not complete https://review.opendev.org/c/openstack/nova/+/784130 to check before approving15:01
gibimelwitt: thanks for the heads up I will look now15:01
gibimelwitt, lyarwood: I'm +A ^^15:06
melwittsweet thanks gibi15:06
lyarwoodack thanks :)15:07
noonedeadpunksean-k-mooney: well, today I learned, and pushed https://review.opendev.org/c/openstack/openstack-ansible-os_nova/+/786258 to use templates. Better late, then never...15:08
gibilyarwood: btw, I went back to the libvirt detach patch series. I hope I can push some update today15:08
*** dklyle has joined #openstack-nova15:09
lyarwoodgibi: oh cool, I've had my head downstream all day with internal escalations15:09
*** ociuhandu has quit IRC15:09
gibilyarwood: sounds bad15:10
*** ociuhandu has joined #openstack-nova15:12
*** mlavalle has joined #openstack-nova15:17
openstackgerritJohannes Kulik proposed openstack/nova master: vmware: Handle empty list attributes on vSphere objects  https://review.opendev.org/c/openstack/nova/+/78626215:18
*** artom has quit IRC15:21
*** k_mouza has quit IRC15:21
*** k_mouza_ has joined #openstack-nova15:21
*** tesseract has quit IRC15:28
gibiwoot http://lists.openstack.org/pipermail/openstack-announce/2021-April/002047.html15:29
lyarwood\o/15:30
*** ociuhandu_ has joined #openstack-nova15:36
*** rcernin has joined #openstack-nova15:39
*** ociuhandu has quit IRC15:40
*** ociuhandu_ has quit IRC15:40
*** rcernin has quit IRC15:43
*** ircuser-1 has joined #openstack-nova15:44
*** hoonetorg has joined #openstack-nova16:03
*** lucasagomes has quit IRC16:04
*** zoharm has joined #openstack-nova16:16
*** gyee has joined #openstack-nova16:17
*** zoharm has quit IRC16:19
*** rpittau is now known as rpittau|afk16:21
*** hamalq has joined #openstack-nova16:28
*** k_mouza_ has quit IRC16:32
openstackgerritBalazs Gibizer proposed openstack/nova master: Replace blind retry with libvirt event waiting in detach  https://review.opendev.org/c/openstack/nova/+/77024616:32
openstackgerritBalazs Gibizer proposed openstack/nova master: Move instance power state check to _detach_with_retry  https://review.opendev.org/c/openstack/nova/+/77891816:32
openstackgerritBalazs Gibizer proposed openstack/nova master: Consolidate device detach error handling  https://review.opendev.org/c/openstack/nova/+/77897816:33
openstackgerritBalazs Gibizer proposed openstack/nova master: Move the guest.get_disk test to test_guest  https://review.opendev.org/c/openstack/nova/+/77715116:35
openstackgerritBalazs Gibizer proposed openstack/nova master: Enable mypy on libvirt/guest.py  https://review.opendev.org/c/openstack/nova/+/77715516:35
openstackgerritBalazs Gibizer proposed openstack/nova master: Follow up type hints for a634103  https://review.opendev.org/c/openstack/nova/+/77715916:36
openstackgerritBalazs Gibizer proposed openstack/nova master: libvirt: Remove dead error handling code  https://review.opendev.org/c/openstack/nova/+/77970416:36
*** ralonsoh has quit IRC16:45
*** k_mouza has joined #openstack-nova16:46
openstackgerritMerged openstack/nova master: docs: Document options for remaining nova-* commands  https://review.opendev.org/c/openstack/nova/+/72619516:54
openstackgerritMerged openstack/nova master: docs: Populate "See Also" section with meaningful links  https://review.opendev.org/c/openstack/nova/+/72991716:55
sean-k-mooneynoonedeadpunk: cool16:56
sean-k-mooneynoonedeadpunk: i think ooo was previously the only tool to make use of that but its a nice feature to have16:56
*** derekh has quit IRC17:05
*** k_mouza has quit IRC17:06
*** k_mouza has joined #openstack-nova17:10
*** k_mouza has quit IRC17:25
*** mkrai has quit IRC17:29
*** artom has joined #openstack-nova17:34
*** rcernin has joined #openstack-nova17:39
openstackgerritStephen Finucane proposed openstack/nova master: db: Remove dead code  https://review.opendev.org/c/openstack/nova/+/78629117:43
openstackgerritStephen Finucane proposed openstack/nova master: db: Remove 'nova.db.sqlalchemy.utils'  https://review.opendev.org/c/openstack/nova/+/78629217:44
openstackgerritStephen Finucane proposed openstack/nova master: db: Remove unused DB methods  https://review.opendev.org/c/openstack/nova/+/78629317:44
openstackgerritStephen Finucane proposed openstack/nova master: test: Remove checks for SQLite < 3.7  https://review.opendev.org/c/openstack/nova/+/78629417:44
openstackgerritStephen Finucane proposed openstack/nova master: db: Use module-level imports for sqlalchemy  https://review.opendev.org/c/openstack/nova/+/78629517:44
openstackgerritStephen Finucane proposed openstack/nova master: db: Fold in indexes  https://review.opendev.org/c/openstack/nova/+/78629617:44
openstackgerritStephen Finucane proposed openstack/nova master: db: Fold in ForeignKey constraints  https://review.opendev.org/c/openstack/nova/+/78629717:44
openstackgerritStephen Finucane proposed openstack/nova master: db: Remove 'nova.db.base' module  https://review.opendev.org/c/openstack/nova/+/78629817:44
openstackgerritStephen Finucane proposed openstack/nova master: db: Copy docs from 'nova.db.*' to 'nova.db.sqlalchemy.*'  https://review.opendev.org/c/openstack/nova/+/78629917:44
openstackgerritStephen Finucane proposed openstack/nova master: db: Synchronize function signatures  https://review.opendev.org/c/openstack/nova/+/78630017:44
openstackgerritStephen Finucane proposed openstack/nova master: db: Clean up migration code  https://review.opendev.org/c/openstack/nova/+/78630117:44
*** dtantsur is now known as dtantsur|afk17:44
*** rcernin has quit IRC17:44
*** andrewbonney has quit IRC17:59
*** k_mouza has joined #openstack-nova18:24
*** k_mouza has quit IRC18:27
*** whoami-rajat has quit IRC18:38
*** vishalmanchanda has quit IRC18:44
*** jmlowe has quit IRC19:18
*** k_mouza has joined #openstack-nova19:28
*** k_mouza has quit IRC19:32
noonedeadpunksean-k-mooney: we really want to follow best practises but not always have time for reading all renos for all releases :( So eventually missing cool things :(19:36
noonedeadpunkoh, wait. Isn't cell name not unique now?19:42
noonedeadpunkhttp://paste.openstack.org/show/804488/ <- Is it expected?19:43
noonedeadpunkthis happens only with the template by the way19:44
noonedeadpunkthis condition feels wrong for the template https://opendev.org/openstack/nova/src/branch/master/nova/cmd/manage.py#L623-L62419:47
noonedeadpunkbecause `cell.database_connection` and `cell.transport_url` are always resolved while checked, and `database_connection` with `transport_url` are shown as template...19:50
noonedeadpunkwrote down https://bugs.launchpad.net/nova/+bug/1923899 - will try to check what can be done there tomorrow, unless there already be a suggestions19:56
openstackLaunchpad bug 1923899 in OpenStack Compute (nova) "Nova create duplicate cells when templates are used" [Undecided,New]19:56
*** ociuhandu has joined #openstack-nova20:00
*** ociuhandu has quit IRC20:04
openstackgerritMerged openstack/nova master: Explaining when a base image is considered unused to avoid confusion why some images aren't deleted.  https://review.opendev.org/c/openstack/nova/+/78605920:05
noonedeadpunkIt feels like template conversion is processed somewhere in models, as eventually here it's already converted from template to url https://opendev.org/openstack/nova/src/branch/master/nova/objects/cell_mapping.py#L25820:13
noonedeadpunkunless it's done somewhere in oslo....20:13
noonedeadpunkas kind of see nothing in models... https://opendev.org/openstack/nova/src/branch/master/nova/db/sqlalchemy/api_models.py#L11120:14
noonedeadpunkok, that's where conversion is done https://opendev.org/openstack/nova/src/branch/master/nova/objects/cell_mapping.py#L159-L16920:18
sean-k-mooneynoonedeadpunk: the cell names are uniqe yes20:19
sean-k-mooneybut i dont think we enforce that with a db constraint20:20
sean-k-mooneyim not sure what will happen with http://paste.openstack.org/show/804488/20:20
noonedeadpunkyeah, there's totally a bug in tempaltes usage20:20
sean-k-mooneyother then it will proably break20:20
sean-k-mooneynoonedeadpunk: it look like maybe you don thave support for multiple cells?20:22
sean-k-mooneylooking at https://review.opendev.org/c/openstack/openstack-ansible-os_nova/+/786258/1/tasks/nova_db_setup.yml20:22
noonedeadpunknot sure just what's the best way to pass to avoid formatting here https://opendev.org/openstack/nova/src/branch/master/nova/objects/cell_mapping.py#L159-L169 considering it's going through oslo.versionedobjects20:22
sean-k-mooneyit was previously hardcoded as   --name {{ nova_cell1_name }}20:22
noonedeadpunkwell, my point is more about that if I run excatly same create_cell command twice - I get exact same 2 cells with different uuids20:23
sean-k-mooneyya which is porably a nova bug20:23
noonedeadpunkbecause verification if that cell already exists is just broken for tempalate user scenario - as it compares what is passed as an argument with already resolved template20:23
sean-k-mooneyi cant think of a  valid reason to allow that20:23
sean-k-mooneyright we should have a unique constraint on the cell name20:24
noonedeadpunkand you can't do that anyway when you use just old good fixed stuff20:24
noonedeadpunkor avoid resolution of the template when we checking if same cell already exist20:25
sean-k-mooneyhttps://opendev.org/openstack/nova/src/branch/master/nova/db/sqlalchemy/api_models.py#L111-L12920:26
noonedeadpunkSo I'd say when we call here https://opendev.org/openstack/nova/src/commit/9e5feb051c15166988e87b10132927135bd56884/nova/cmd/manage.py#L618 we should pass smth like resolve=false20:26
sean-k-mooneyso the unique constatin is only the uuid currently20:26
sean-k-mooneywe shoudl also have one  on the name20:26
noonedeadpunkyeah, I saw that...20:26
sean-k-mooneydansmith: do you know why that is the case today ^20:26
sean-k-mooneydansmith: i.e. why we allow to cell records with the same name?20:27
sean-k-mooneynoonedeadpunk: as far as i know we only use the id in other tables20:27
sean-k-mooneynoonedeadpunk: so other then cell0 which is speciall we dont really use the cell name20:27
sean-k-mooneynoonedeadpunk: its just for operators really20:28
dansmithyeah, name is just for the humans20:28
noonedeadpunkwell, I'm probably even ok that cell name is not unique. I just see no reason in adding cells with same name, transport_url and database_connection20:28
dansmithbut no I dunno any reason to or not to be unique on name for any reason20:28
dansmithnoonedeadpunk: well, we had initially talked about sharding on the db but not the mq, or vice versa,20:29
dansmithso I think that's why we didn't make those unique20:29
noonedeadpunkAnd there's an attempt to avoid that with condition and raising `The specified transport_url and/or database_connection combination already exists for another cell with uuid ef89aebc-142d-4755-a377-6ed23b6c1ed4.`20:29
sean-k-mooneydansmith: i was debating if this was for some kind of loadbalancing20:29
dansmithhowever, I think we'd have work to do before that would actually work20:29
dansmithat least testing20:29
sean-k-mooneykind of like the way you can list multple glance? rabbnit? instances in the config instead of using haproxy20:30
dansmithno, that was never the reasoning for lack of uniqueness that I recall,20:31
dansmithmore the opposite,20:31
dansmithwhere you wanted to be able to have two DBs and not two MQs or the like20:31
sean-k-mooneyok20:31
sean-k-mooneynoonedeadpunk: in any case i guess you need to figure out why you are creatin gmultiple cells20:31
sean-k-mooneynoonedeadpunk: i guess its not checking if it already exits?20:32
noonedeadpunkum... any time I run exact same comand - new cell is created. I just want to use same cell, and get exit code 2, when cell with provided transport_url and database_connection already exist20:32
noonedeadpunkand when I get exit code 2, I want to update cell if necessary20:33
noonedeadpunksean-k-mooney: it's checking, but check is broken20:33
noonedeadpunkI'm kind of trying to explain that for the last several minutes :)20:33
sean-k-mooneyso nova manage bug?20:33
noonedeadpunkhttps://opendev.org/openstack/nova/src/commit/9e5feb051c15166988e87b10132927135bd56884/nova/cmd/manage.py#L623-L62420:33
sean-k-mooney hum ok20:34
sean-k-mooneythat should retrun true an print the message20:35
sean-k-mooneydid you get that message printed20:35
noonedeadpunkcell.database_connection here is resolved and `mysql+pymysql://nova:534d29a090559b5a6345122e999a8ef29f67@172.29.236.101/nova?charset=utf8` but database_connection is `{scheme}://{username}:{password}@{hostname}/{path}?{query}`20:35
sean-k-mooneylooking at http://paste.openstack.org/show/804488/ no20:35
sean-k-mooneynoonedeadpunk: ah ok20:35
noonedeadpunkas when we run objects.CellMappingList.get_all(ctxt) we end up in https://opendev.org/openstack/nova/src/branch/master/nova/objects/cell_mapping.py#L159-L16920:36
sean-k-mooneythat porbaly happing when we load the object form the db20:36
sean-k-mooneyyep20:36
noonedeadpunkyep, and we trigger it here https://opendev.org/openstack/nova/src/branch/master/nova/objects/cell_mapping.py#L25820:36
sean-k-mooneyso we need to do this https://opendev.org/openstack/nova/src/branch/master/nova/objects/cell_mapping.py#L162-L16520:37
sean-k-mooneyin the nova manage command20:37
noonedeadpunkSo I was thinking to add an extra argumet or adjust context, so that it contained like `template_resolve=false`20:37
sean-k-mooneynoonedeadpunk: i think it would be simpler just to reolve the unresolved one20:37
sean-k-mooneyby calling cell_mapping.format_db_url20:38
sean-k-mooneyin this if https://opendev.org/openstack/nova/src/commit/9e5feb051c15166988e87b10132927135bd56884/nova/cmd/manage.py#L623-L62420:38
noonedeadpunkwell, that would have side effect, that in case config on the node where we run nova-manage differs, new cell will be created anyway20:38
noonedeadpunkso I think we should compare what we have in database and what we're about to add there20:39
sean-k-mooneywell the  cell.database_connection value is being formated with your local nova.conf too20:40
sean-k-mooneynoonedeadpunk: so for your actul usecase you could list the cells and check if the cell is present with that name already20:41
noonedeadpunkyes, but I may depend from node to node?20:41
sean-k-mooneyand only create it if it does not exits20:42
noonedeadpunk*it may depend20:42
sean-k-mooneywell nova-manage is only run on the controler nodes right and typically only one of the contoler per cell20:43
sean-k-mooneyyou do not normally run it on compute nodes for example20:43
noonedeadpunkI mean shouldn't be there a valid protection, that exact same cells are not present in DB with different UUIDs?20:44
sean-k-mooneynoonedeadpunk: checking the db content wont work i this case actully20:44
sean-k-mooneyi think the cell name is really the unique thing20:45
sean-k-mooneyingoring the sharding usecase20:45
sean-k-mooneythat is not currently implemetned20:45
sean-k-mooneydansmith: actully i dont think we can support the sharding usecase with the current db schema without denormalising the database20:47
noonedeadpunkbut at least combination of name+transport_url+database_connection should be unique IMO. And considering templates, I see the only way to do that is to check here https://opendev.org/openstack/nova/src/commit/9e5feb051c15166988e87b10132927135bd56884/nova/cmd/manage.py#L623-L624 that cell.database_connection is still in tempate format, but not transformed20:47
sean-k-mooneydansmith: we would realy need to have a second db table that mapped db urls to the cell id or similar20:47
noonedeadpunkI kind of submitted bug regarding this https://bugs.launchpad.net/nova/+bug/1923899 as well20:48
openstackLaunchpad bug 1923899 in OpenStack Compute (nova) "Nova create duplicate cells when templates are used" [Undecided,New]20:48
dansmithsean-k-mooney: yeah, above I said that was the original intent to be able to do that, but that we'd still have work to do to make it work20:48
noonedeadpunksorry for late discussion, already midnight here, so need to head out :(20:49
sean-k-mooneyno worries20:49
sean-k-mooneyi think if we jsut put a unique constarint on name that would be enough. if we want to supprot sharding we would need to have a second table with something like (id, cell_id, db20:50
noonedeadpunkeventually what I intended to do is to rely on cell_create command return code, and if it's 2 - update_cell, if 0 - proceed with setup. And for me it's now always 0 when using templates :(20:50
sean-k-mooney(id, cell_id, db_url, db_table)20:51
sean-k-mooneynoonedeadpunk: ya for now i guess you have to check and not use the return code20:51
sean-k-mooneybut i think the bug is more or less valid and we should be returning 220:51
noonedeadpunkyeah:(20:52
sean-k-mooneydansmith: im not sure how ye were thinking about shardin gby the way so you might not need to do it by table so maybe the other table would just be (id, cell_id, db_url) not (id, cell_id, db_url, db_table)20:53
sean-k-mooneybut the impotant thing is we would need to model the many to one relation ship while maintianing only on uuid per cell20:55
*** rcernin has joined #openstack-nova21:02
mnaseri've been noticing an issue where in a scenario that libvirt loses config, when using vgpu, and doing a hard reboot generates xml that _does not_ contain the vgpu21:06
mnaserto reproduce: openstack server stop <instance>; virsh undefine <instance>; openstack server start <instance>; virsh dumpxml <instance> ... no mdev21:06
mnaserhttps://github.com/openstack/nova/blob/master/nova/virt/libvirt/driver.py#L3453-L345621:12
mnaserthis is starting to sound like mdev state is actually conserved in libvirt domains?21:13
*** Hazelesque has quit IRC21:19
*** rcernin has quit IRC21:45
*** rcernin has joined #openstack-nova21:46
*** jkulik has quit IRC21:52
*** jkulik has joined #openstack-nova21:54
melwittmnaser: this seems kinda related https://bugs.launchpad.net/nova/+bug/190080021:57
openstackLaunchpad bug 1900800 in OpenStack Compute (nova) "VGPUs is not recreated on host reboot" [Low,Confirmed] - Assigned to Sylvain Bauza (sylvain-bauza)21:57
mnasermelwitt: that’s exactly it22:01
*** rcernin has quit IRC22:20
*** luksky has quit IRC22:34
*** luksky has joined #openstack-nova22:34
*** luksky has quit IRC22:40
*** tkajinam has quit IRC22:43
*** tkajinam has joined #openstack-nova22:52
*** luksky has joined #openstack-nova22:53
*** hamalq has quit IRC22:57
*** hamalq has joined #openstack-nova22:57
*** macz_ has quit IRC23:05
*** rcernin has joined #openstack-nova23:06
*** luksky has quit IRC23:18
*** tosky has quit IRC23:29
*** irclogbot_1 has quit IRC23:50
*** irclogbot_1 has joined #openstack-nova23:56
« Tuesday, 2021-04-13 Index Thursday, 2021-04-15 »

Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!