Thursday, 2020-04-02

*** enriquetaso has joined #openstack-cinder00:25
*** ociuhandu has joined #openstack-cinder00:34
*** ociuhandu has quit IRC00:38
*** Xuchu has joined #openstack-cinder00:58
*** Liang__ has joined #openstack-cinder01:08
*** Xuchu has quit IRC01:39
*** zhanglong has joined #openstack-cinder01:40
*** ruffian_sheep has joined #openstack-cinder01:40
*** Xuchu has joined #openstack-cinder01:43
*** ruffian_sheep has quit IRC01:48
*** raunak12 has quit IRC02:49
*** xuanyd has joined #openstack-cinder03:03
xuanydwhoami-rajat: hello, the path has update, please review again, thanks! https://review.opendev.org/#/c/705876/03:05
*** sapd1_x has joined #openstack-cinder03:10
*** Liang__ is now known as LiangFang03:22
*** enriquetaso has quit IRC03:24
*** ykarel|away is now known as ykarel04:33
*** rishabhhpe has joined #openstack-cinder04:42
*** udesale has joined #openstack-cinder04:56
*** sapd1_x has quit IRC05:02
*** yaawang has quit IRC05:05
*** yaawang has joined #openstack-cinder05:05
*** threestrands has joined #openstack-cinder05:06
*** fnordahl has joined #openstack-cinder05:10
*** sapd1_x has joined #openstack-cinder05:14
*** ociuhandu has joined #openstack-cinder05:58
*** ociuhandu has quit IRC06:02
*** spatel has joined #openstack-cinder06:20
openstackgerritXuchu Jiang proposed openstack/cinder master: Delete volume with additional removehostmappings parameter  https://review.opendev.org/71576606:21
*** mnasiadka has quit IRC06:21
*** mnasiadka has joined #openstack-cinder06:22
*** ociuhandu has joined #openstack-cinder06:23
*** spatel has quit IRC06:25
*** sapd1_x has quit IRC06:29
*** vishalmanchanda has joined #openstack-cinder06:30
*** jv has quit IRC06:39
*** zhanglong has quit IRC06:39
*** zhanglong has joined #openstack-cinder06:40
*** sapd1_x has joined #openstack-cinder06:41
*** gary_perkins has quit IRC06:55
*** gary_perkins has joined #openstack-cinder06:58
*** ociuhandu has quit IRC07:03
*** Xuchu has quit IRC07:03
*** ociuhandu has joined #openstack-cinder07:04
*** Xuchu has joined #openstack-cinder07:05
*** ociuhandu has quit IRC07:09
*** yaawang has quit IRC07:34
*** yaawang has joined #openstack-cinder07:34
*** ociuhandu has joined #openstack-cinder07:37
*** sapd1_x has quit IRC07:39
*** threestrands has quit IRC08:01
*** tosky has joined #openstack-cinder08:02
*** yaawang has quit IRC08:03
*** yaawang has joined #openstack-cinder08:04
*** ykarel is now known as ykarel|lunch08:08
*** e0ne has joined #openstack-cinder08:15
*** yaawang has quit IRC08:32
*** yaawang has joined #openstack-cinder08:34
*** ociuhandu has quit IRC08:36
*** ociuhandu has joined #openstack-cinder08:38
*** martinkennelly has joined #openstack-cinder08:39
*** ociuhandu has quit IRC08:43
*** ociuhandu has joined #openstack-cinder09:10
*** LiangFang has quit IRC09:22
*** xuanyd has quit IRC09:30
*** Xuchu has quit IRC09:30
*** Xuchu has joined #openstack-cinder09:35
*** ykarel|lunch is now known as ykarel09:42
*** tkajinam has quit IRC09:47
openstackgerritXuchu Jiang proposed openstack/cinder master: Delete volume with additional removehostmappings parameter  https://review.opendev.org/71576609:55
*** zhanglong has quit IRC10:08
*** spatel has joined #openstack-cinder10:09
*** spatel has quit IRC10:14
*** Xuchu has quit IRC10:21
openstackgerritLuigi Toscano proposed openstack/cinder master: DNM Check that revert tests fail on RBD  https://review.opendev.org/71691910:26
openstackgerritSeema S B proposed openstack/cinder master: Add different RBAC rule for different HTTP REST API  https://review.opendev.org/67879910:33
*** n-saito has quit IRC10:35
*** Xuchu has joined #openstack-cinder10:49
openstackgerritOpenStack Release Bot proposed openstack/os-brick stable/ussuri: Update .gitreview for stable/ussuri  https://review.opendev.org/71694010:54
openstackgerritOpenStack Release Bot proposed openstack/os-brick stable/ussuri: Update TOX/UPPER_CONSTRAINTS_FILE for stable/ussuri  https://review.opendev.org/71694110:54
openstackgerritOpenStack Release Bot proposed openstack/os-brick master: Update master for stable/ussuri  https://review.opendev.org/71694210:54
openstackgerritMerged openstack/devstack-plugin-open-cas master: Add tempest tests to zuul  https://review.opendev.org/71377211:04
*** tkajinam has joined #openstack-cinder11:12
hemnaiœ6pö333fhjfd11:27
hemnamy cat says good morning11:27
rosmaitagood morning11:27
*** vishakha has quit IRC11:36
*** udesale_ has joined #openstack-cinder11:40
*** udesale has quit IRC11:43
*** rcernin has quit IRC11:54
*** ykarel is now known as ykarel|afk12:03
walshh_is anyone available to look at https://review.opendev.org/#/c/711217/ It has +2 and clean CI.  Thank you12:05
*** jv__ has joined #openstack-cinder12:13
*** ykarel|afk is now known as ykarel12:13
*** jv__ has quit IRC12:20
hemnawalshh_ added my feedback12:21
hemnaI think we have many drivers that don't use driver.do_setup() and driver.check_for_setup_errors() like this, but we should encourage driver maintainers to use those instead of doing validation checks at driver.__init__ time.12:22
hemnaFWIW,  this is where the driver is created in the volume manager: https://github.com/openstack/cinder/blob/master/cinder/volume/manager.py#L260-L26712:27
hemnathe volume manager calls driver.do_setup() in it's init_host() call, which gives the volume manager a chance to create profiling and output the driver name and version info to the logs.12:28
hemnait's probably a little late at this stage to encourage this, but drivers really should use do_setup() and check_for_setup_errors()12:29
walshh_thanks hemna12:31
hemnaI suppose you can do a follow up to refactor that a bit if you want.12:31
*** ykarel is now known as ykarel|afk12:33
walshh_We will look into it, we have 4 other methods in the init that are candidate to the moved to do_setup so perhaps we could do it in a separate review12:33
walshh_candidates*12:34
hemnayah, I figured it would take a bit of work to refactor that12:34
hemnaif you file a bug on it, then I'll +A this one12:34
walshh_great feedback, thank you12:34
*** jv__ has joined #openstack-cinder12:34
*** enriquetaso has joined #openstack-cinder12:35
hemnasmcginnis ping12:42
hemnawhat's the best way to skip certain tempest tests?  like if I have a policy that will always cause a test to fail12:44
hemnaI have lots of those12:44
*** ajitha has joined #openstack-cinder12:45
rosmaitawhoami-rajat: around? question about https://review.opendev.org/#/c/705876/12:52
rosmaitayour comment on PS 2312:52
whoami-rajatrosmaita, yep12:53
*** jv__ has quit IRC12:53
hemnawalshh_ have that bug url for me?12:54
*** Xuchu has quit IRC12:54
smcginnishemna: I believe you just pass a regex to match (or in your case not match) a pattern.12:54
hemnaand if I have like....20+ of them?12:55
rosmaitawhoami-rajat: sorry, am trying to formulate my question12:55
*** ykarel|afk is now known as ykarel12:55
whoami-rajatrosmaita, np, i'm here12:55
rosmaitawhoami-rajat: so what you were able to make happen is that you can create a volume with the incorrect src_backup_id12:56
rosmaitaand that seems to be because the src_backup_id is only written to new volumes, not to restores?12:57
whoami-rajatrosmaita, yes and my concern was it's a static field once set doesn't change again12:57
smcginnishemna: There might be some tempest.conf settings you can use, but I'm not finding anything.12:58
smcginnishemna: Looking around in https://docs.openstack.org/tempest/latest/run.html12:58
hemnahrmm ok12:58
hemnawe have a policy.json that restricts a lot of stuff, which will cause tempest.api.volume test failures12:58
whoami-rajatrosmaita, yes12:58
smcginnisMaybe easier using the --blacklist-file or --whitelist-file?12:58
rosmaitawhoami-rajat: it's static in the sense that it doesn't get modified by current workflow, but there's no reason why it could not be modified, is that right?12:58
hemnatrying to find the 'right' way to just skip those with a blacklist or something12:58
hemnaok I suppose I could create a blacklist file and add each of the desired tests to be skipped there12:59
hemnathanks, I'll try that12:59
rosmaitawhoami-rajat: because it's just volume metadata, so could be changed/removed by volume owner/admin13:00
smcginnishemna: I have noticed some CI's passing a pretty long regex string, so it's certainly possible to do it all in one command line. But of course, complicated regex == hard to read. ;)13:01
whoami-rajatrosmaita, from a code perspective, i think a volume should have the backup id of the latest backup it's restored from13:01
whoami-rajatrosmaita, i can't say regarding the owner/admin changing it manually :/13:01
smcginnishemna: Just as an example, search for "tox -eall" in here: http://s3.eu-geo.objectstorage.softlayer.net/ibm-storage-ci-log/storwize-tempest-dsvm-full-iscsi/751/console.txt13:01
*** ociuhandu has quit IRC13:02
hemnayah that's a little nasty13:02
rosmaitawhoami-rajat: right, i wasn't clear in my question ... i just meant that there's no reason why the code could not handle the case of changing a src_backup_id13:03
*** jv__ has joined #openstack-cinder13:04
whoami-rajatrosmaita, yep, we can handle the case in the code13:04
whoami-rajatrosmaita, from the recent change, i see they've considered the src_backup_id during the update but i think it still lacks some verification as to if the src_backup_id is a valid UUID and a valid backup13:06
rosmaitaok, i will update that review, thanks13:06
whoami-rajatrosmaita, np, but i'm still unsure what the query was :P13:09
*** laurent\ has quit IRC13:16
*** ianychoi_ has joined #openstack-cinder13:17
openstackgerritAndreas Jaeger proposed openstack/cinderlib master: Update hacking for Python3  https://review.opendev.org/71699713:19
openstackgerritAndreas Jaeger proposed openstack/cinderlib master: Update hacking for Python3  https://review.opendev.org/71699713:19
*** ianychoi has quit IRC13:20
*** ianychoi_ is now known as ianychoi13:23
*** rishabhhpe has quit IRC13:25
*** ykarel is now known as ykarel|afk13:29
rosmaitawhoami-rajat: when you have  a few minutes, see if my comment on PS31 of https://review.opendev.org/#/c/705876/ makes sense13:32
*** xakaitetoia has joined #openstack-cinder13:34
*** Xuchu has joined #openstack-cinder13:45
*** xuanyd has joined #openstack-cinder13:53
*** sfernand has joined #openstack-cinder13:55
whoami-rajatrosmaita, that conditional statement was very confusing, the only case they are avoiding is when a volume is restored several times with the same backup, they won't update the db and performance wise i think that makes sense13:57
*** Xuchu has quit IRC13:57
*** vishalmanchanda has quit IRC13:58
xuanydrosmatia: hello, thanks for review and thanks for suggestions. i think that, the src_backup_id is mark the volume's init data, So we just need to ensure that src_backup_id is up to date14:03
openstackgerritAndreas Jaeger proposed openstack/cinderlib master: Update hacking for Python3  https://review.opendev.org/71699714:05
xuanydrosmaita:  hello, thanks for review and thanks for suggestions. i think that, the src_backup_id is mark the volume's init data, So we just need to ensure that src_backup_id is up to date14:11
sfernandHi everyone14:15
sfernandcould anyone take a look at this patch? https://review.opendev.org/#/c/712799/114:15
sfernandThis adds Active/Active replication to the SolidFire driver, also enables the SUPPORTS_ACTIVE_ACTIVE flag14:15
*** ociuhandu has joined #openstack-cinder14:17
xuanydrosmaita: about the last comment, When recovering to the original volume, if src_backup_id is empty, we do not write src_backup_id14:18
rosmaitaxuanyd: right, but if src_backup_id is not empty, you do14:21
rosmaitaso you get a situation were sometimes you do, sometimes you don't14:21
*** Xuchu has joined #openstack-cinder14:23
*** N3l1x has joined #openstack-cinder14:25
*** ykarel|afk is now known as ykarel14:28
*** ociuhandu has quit IRC14:34
xuanydrosmatia: got it, I will split the conditions to make it clearer, thanks14:34
rosmaitaxuanyd: ok, i'll take a look after you update14:35
xuanydrosmatia: how about change to :if backup.volume_id == volume.id and old_src_backup_id != backup.id:14:38
xuanydrosmaita: Can you tell me which scenarios will go wrong about path 3114:46
rosmaitaxuanyd: i am in a meeting now, will look in 15 min or so14:47
openstackgerritMerged openstack/os-brick master: Update master for stable/ussuri  https://review.opendev.org/71694214:49
openstackgerritMerged openstack/os-brick stable/ussuri: Update .gitreview for stable/ussuri  https://review.opendev.org/71694014:50
openstackgerritMerged openstack/os-brick stable/ussuri: Update TOX/UPPER_CONSTRAINTS_FILE for stable/ussuri  https://review.opendev.org/71694114:50
*** throne82 has joined #openstack-cinder14:59
throne82Hi! I have an bug - https://bugs.launchpad.net/cinder/+bug/187036715:01
openstackLaunchpad bug 1870367 in Cinder "Fail to extend attached volume using generic NFS driver" [Undecided,New]15:01
throne82How can we proceed with this ?15:01
xuanydrosmaita: Hello, are you free now15:08
rosmaitaxuanyd: in 5 minutes15:08
rosmaitaxuanyd: free now15:15
*** ociuhandu has joined #openstack-cinder15:18
xuanydrosmaita: Can you tell me which scenarios will go wrong about path 3115:19
rosmaitalooking15:20
hemnamep15:21
xuanydrosmaita: I think the current conditions can be met to restore the new volume and restore the original volume with src_backup_id15:22
*** ociuhandu has quit IRC15:23
rosmaitaok, i will try to talk through this and we can see if it is a problem or not15:25
xuanydthanks15:28
rosmaitasorry for the delay ... i am confusing myself!15:29
xuanydNever mind, there are more conditions15:32
*** ociuhandu has joined #openstack-cinder15:34
rosmaitaxuanyd: so the good thing is that lines 616-622 make sure that if a volume has src_backup_id metadata, then it is correct15:38
rosmaitaso that is goo15:38
rosmaita*good15:38
rosmaitaso actually i think your code is fine15:39
xuanydYes, thanks for review, i will rename the src_backup_id to old_src_backup_id in line 61415:41
rosmaitaok, cool15:41
rosmaitaxuanyd: thanks for your patience!15:42
xuanydGlad to discuss the problem with you too.15:44
*** ykarel is now known as ykarel|away15:47
openstackgerritXuan Yandong proposed openstack/cinder master: Add features for add backup id to volume metadata  https://review.opendev.org/70587615:48
xuanydrosmaita: I submitted the update, please review it again latter, thanks very much15:51
rosmaitaxuanyd: i will keep an eye on it15:51
*** xuanyd has quit IRC15:55
openstackgerritMerged openstack/cinder master: PowerMax Driver - Replication array serial check  https://review.opendev.org/71121716:19
*** KeithMnemonic has joined #openstack-cinder16:25
*** udesale_ has quit IRC16:26
*** ajitha has quit IRC16:43
*** ociuhandu has quit IRC16:52
openstackgerritHelen Walsh proposed openstack/cinder master: PowerMax Driver - Refreshing stale metadata  https://review.opendev.org/71297717:02
*** martinkennelly has quit IRC17:08
*** martinkennelly has joined #openstack-cinder17:09
*** KeithMnemonic has quit IRC17:12
*** martinkennelly has quit IRC17:24
*** martinkennelly has joined #openstack-cinder17:25
*** martinkennelly has quit IRC17:27
*** KeithMnemonic has joined #openstack-cinder17:28
*** martinkennelly has joined #openstack-cinder17:28
*** ociuhandu has joined #openstack-cinder17:32
*** ociuhandu has quit IRC17:36
*** rishabhhpe has joined #openstack-cinder17:37
*** xakaitetoia has quit IRC17:38
*** rishabhhpe has quit IRC17:46
*** rishabhhpe has joined #openstack-cinder17:47
*** xakaitetoia has joined #openstack-cinder17:48
*** xakaitetoia has quit IRC17:56
*** andrebeltrami has quit IRC18:06
*** rishabhhpe has quit IRC18:08
*** xakaitetoia has joined #openstack-cinder18:24
*** xakaitetoia has quit IRC18:29
*** xakaitetoia has joined #openstack-cinder19:01
*** xakaitetoia has quit IRC19:07
*** xakaitetoia has joined #openstack-cinder19:34
*** xakaitetoia has quit IRC19:39
*** ociuhandu has joined #openstack-cinder19:52
*** ociuhandu has quit IRC20:02
*** ociuhandu has joined #openstack-cinder20:02
*** ociuhandu has quit IRC20:07
*** xakaitetoia has joined #openstack-cinder20:09
*** xakaitetoia has quit IRC20:17
enriquetasoHello ganso o/ : I'm not able to reproduce step 3 in https://bugs.launchpad.net/cinder/+bug/186974620:18
openstackLaunchpad bug 1869746 in Cinder "Cinder throws error creating incremental backup from parent in another project" [Undecided,New]20:18
enriquetasoI'm able to see volumes by other tenants from the admin user. The documentation says: "--all-tenants [<all_tenants>] Shows details for all tenants. Admin only."20:18
enriquetasoBecause of that, I can't see the backup created by the admin with the demo user, I guess it's not possible unless the demo user has admin role. Does the demo has admin roles?20:18
gansoenriquetaso: no it doesn't have admin roles. I am confused, are you sure it is step 3? Don't you mean step 2?20:19
enriquetasoganso, i'm able to create the full backup as user admin :) but I can't see the backup with the demo user, because of that I can't create an incremental from it, but I can create a full backup instead.20:21
gansoenriquetaso: in step 1, you create the volume as the user: "openstack volume create --size 1 v1". In step 2, you change your credentials to admin, and run "openstack volume backup create <v1_id>". Step 3 you go back to demo credentials and run "openstack volume backup create <v1_id or just v1> --incremental".20:21
gansoenriquetaso: you don't need to see the backup to issue another incremental one20:21
enriquetasooh true, confirmed then :P20:24
enriquetasothanks ganso, i'll try to propose a fix this week20:26
gansoenriquetaso: if you are the user you don't need to know about other backups that have been taken by the admin. Those steps are the simplest to reproduce the issue. On a more realistic scenario, the user would have a full backup, then the admin has another one, and then the user, oblivious to the fact the admin has taken another one, tries an incremental and gets the error. Regardless of that, the docs state that even if the user has no backups and20:26
gansospecifies --incremental, it will ignore the parameter and create a full one automatically, which creates the use case of the user always using --incremental with no full backups and hitting the error as well20:26
gansoenriquetaso: I created this etherpad for us to discuss https://etherpad.openstack.org/p/cinder-backup-bug20:26
gansoenriquetaso: we need to choose the proper solution. I like (c) best, but it has a few concerns, and I am not sure we will be able to backport it :(20:30
enriquetasoThanks for creating the etherpad ganso! I'll check the solutions an write my opinion then20:39
gansoenriquetaso: great! thanks! let's work together on this and discuss the solution with the team if there are any other concerns we missed20:40
enriquetasosounds good to me20:41
*** Xuchu has quit IRC20:52
*** ociuhandu has joined #openstack-cinder21:02
*** e0ne has quit IRC21:06
*** ociuhandu has quit IRC21:06
*** dpawlik has quit IRC21:12
*** Xuchu has joined #openstack-cinder21:29
*** Lucas_Gray has joined #openstack-cinder21:34
*** Lucas_Gray has quit IRC21:40
openstackgerritHelen Walsh proposed openstack/cinder master: PowerMax Driver - Version comparison correction  https://review.opendev.org/71332921:52
*** rcernin has joined #openstack-cinder22:22
*** openstackstatus has quit IRC22:56
*** openstack has joined #openstack-cinder23:01
*** ChanServ sets mode: +o openstack23:01
*** martinkennelly has quit IRC23:21
*** Xuchu has quit IRC23:24
*** Xuchu has joined #openstack-cinder23:27
*** enriquetaso has quit IRC23:34
*** rosmaita has quit IRC23:46
*** Xuchu has quit IRC23:56
*** Xuchu has joined #openstack-cinder23:58

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!