Wednesday, 2021-05-05

*** dave-mccowan has quit IRC00:08
*** dave-mccowan has joined #openstack-cinder00:10
*** sapd1_x has quit IRC00:28
*** martinkennelly has quit IRC01:05
*** tkajinam has joined #openstack-cinder01:13
openstackgerritMerged openstack/devstack-plugin-ceph master: Remove the stable branch jobs from master gate
openstackgerritMerged openstack/cinder stable/wallaby: Open local image files with "rb" mode
openstackgerritMerged openstack/cinder master: Remove unused _get_non_shared_target_hosts from cmd/manage
openstackgerritYuehuiLei proposed openstack/os-brick master: setup.cfg: Replace dashes with underscores
openstackgerritYuehuiLei proposed openstack/python-brick-cinderclient-ext master: setup.cfg: Replace dashes with underscores
*** carloss has quit IRC02:30
*** jv has quit IRC02:41
*** jv has joined #openstack-cinder03:00
*** rcernin has quit IRC03:05
*** rcernin has joined #openstack-cinder03:16
*** enriquetaso has quit IRC03:26
*** psachin has joined #openstack-cinder03:37
*** ociuhandu has joined #openstack-cinder03:49
*** ociuhandu has quit IRC03:54
*** sapd1_x has joined #openstack-cinder04:15
*** jv has quit IRC04:24
*** amar7ibm has quit IRC04:25
*** GirishChilukuri has joined #openstack-cinder04:43
*** m75abrams has joined #openstack-cinder04:45
*** vishalmanchanda has joined #openstack-cinder04:51
*** jv has joined #openstack-cinder04:51
openstackgerritMerged openstack/os-brick stable/train: multipath/iscsi: remove devices from multipath monitoring
*** jv has quit IRC05:23
*** sapd1_x has quit IRC05:30
*** GirishChilukuri has quit IRC06:03
*** GirishChilukuri has joined #openstack-cinder06:12
*** jv has joined #openstack-cinder06:13
*** char_ has quit IRC06:16
*** GirishChilukuri has quit IRC06:16
*** hack-char has joined #openstack-cinder06:17
*** whoami-rajat has joined #openstack-cinder06:18
*** zenkuro has quit IRC06:18
*** GirishChilukuri has joined #openstack-cinder06:21
*** GirishChilukuri has quit IRC06:24
*** GirishChilukuri has joined #openstack-cinder06:26
*** pcheli has joined #openstack-cinder06:41
*** pcheli has quit IRC06:50
*** e0ne has joined #openstack-cinder07:07
*** gokhani has joined #openstack-cinder07:08
*** ociuhandu has joined #openstack-cinder07:10
*** rpittau|afk is now known as rpittau07:12
*** GirishChilukuri has quit IRC07:14
*** GirishChilukuri has joined #openstack-cinder07:15
*** sapd1_x has joined #openstack-cinder07:18
*** rcernin has quit IRC07:21
*** rcernin has joined #openstack-cinder07:25
*** GirishChilukuri has quit IRC07:25
openstackgerritLuigi Toscano proposed openstack/cinder master: zuul: fixes for the A/A job (nodeset, variables)
*** GirishChilukuri has joined #openstack-cinder07:26
*** tosky has joined #openstack-cinder07:30
*** sapd1_x has quit IRC07:42
*** sapd1_x has joined #openstack-cinder07:49
*** ociuhandu has quit IRC07:54
*** ociuhandu has joined #openstack-cinder07:56
*** rcernin has quit IRC07:59
*** ociuhandu has quit IRC08:00
*** jv has quit IRC08:03
*** ociuhandu has joined #openstack-cinder08:12
*** ianychoi_ has quit IRC08:13
gokhaniHi folks, In my victoria environment, I am using cinder with  Netapp Ontap backend. When I try to attach volume to instance, it is very slow and it takes about 5 minutes. When I examine logs, I see lots of "Searching for a device in session 29 and hctl ['17', '*', '*', 1] yield: None device_name_by_hctl" debug logs. then get "LUN 1 on iSCSI portal08:20
gokhani10.8.48.13:3260 not found on sysfs after logging in" message. I realized that it tries all iscsi sessions. But in my environment I have 2 backends netapp_sas and netapp_ssd. I have 2 sessions per backend. when I attach volumes, it is weird and it tries all iscsi sessions both netapp_sas and netapp_ssd backend sessions. Normally it needs to try only08:20
gokhaniits backend session which is created. do I miss something to do?  logs are in ı need your help.08:20
*** rcernin has joined #openstack-cinder08:23
*** pcheli has joined #openstack-cinder08:26
*** rcernin has quit IRC08:29
*** rcernin has joined #openstack-cinder08:38
*** rcernin has quit IRC08:42
*** rcernin has joined #openstack-cinder08:46
*** alexe9191 has joined #openstack-cinder08:49
*** alexe919182 has joined #openstack-cinder08:50
*** alexe9191 has quit IRC08:53
*** alexe919182 is now known as alexe919108:53
*** martinkennelly has joined #openstack-cinder09:01
*** pcheli has quit IRC09:10
*** pcheli has joined #openstack-cinder09:10
*** pcheli has quit IRC09:16
*** rcernin has quit IRC09:30
*** pcheli has joined #openstack-cinder09:35
*** udesale has joined #openstack-cinder09:45
*** GirishChilukuri has quit IRC09:45
*** pcheli has quit IRC09:46
*** GirishChilukuri has joined #openstack-cinder09:48
*** rcernin has joined #openstack-cinder09:54
*** GirishChilukuri has quit IRC10:01
*** rcernin has quit IRC10:05
*** GirishChilukuri has joined #openstack-cinder10:11
*** pcheli has joined #openstack-cinder10:13
*** GirishChilukuri has quit IRC10:13
*** GirishChilukuri has joined #openstack-cinder10:18
*** pcheli has quit IRC10:23
*** tosky has quit IRC10:38
*** tosky has joined #openstack-cinder10:38
*** carloss has joined #openstack-cinder10:41
*** e0ne has quit IRC10:42
*** tosky_ has joined #openstack-cinder10:45
*** tosky has quit IRC10:47
*** tosky_ is now known as tosky10:49
*** GirishChilukuri_ has joined #openstack-cinder10:51
*** rcernin has joined #openstack-cinder10:52
*** GirishChilukuri has quit IRC10:55
*** rcernin has quit IRC10:58
openstackgerritMerged openstack/cinder stable/victoria: Fix PowerStore iSCSI targets filtering
*** rcernin has joined #openstack-cinder11:02
*** GirishChilukuri_ has quit IRC11:02
*** ociuhandu has quit IRC11:06
*** rcernin has quit IRC11:06
*** sapd1_x has quit IRC11:18
*** ociuhandu has joined #openstack-cinder11:22
*** dsariel has joined #openstack-cinder11:24
*** rcernin has joined #openstack-cinder11:24
*** ociuhandu has quit IRC11:27
*** rcernin has quit IRC11:28
*** ociuhandu has joined #openstack-cinder11:38
*** gokhani has quit IRC11:42
*** ociuhandu has quit IRC11:42
*** rcernin has joined #openstack-cinder11:45
*** dsariel has quit IRC11:53
*** ociuhandu has joined #openstack-cinder11:54
*** rcernin has quit IRC11:55
*** gokhani has joined #openstack-cinder11:56
*** ociuhandu has quit IRC11:59
*** Luzi has joined #openstack-cinder12:00
*** dviroel is now known as dviroel|nho12:01
*** jv has joined #openstack-cinder12:02
*** zenkuro has joined #openstack-cinder12:05
openstackgerritAmar proposed openstack/cinder master: [SVF]: Fix extend issue for a clone of rep-volume
*** ociuhandu has joined #openstack-cinder12:11
*** rcernin has joined #openstack-cinder12:18
*** ociuhandu has quit IRC12:18
*** amar7ibm has joined #openstack-cinder12:19
*** eharney has joined #openstack-cinder12:22
*** rcernin has quit IRC12:23
*** ociuhandu has joined #openstack-cinder12:33
*** eharney has quit IRC12:37
*** ociuhandu has quit IRC12:44
*** eharney has joined #openstack-cinder12:52
*** enriquetaso has joined #openstack-cinder12:52
openstackgerritAmar proposed openstack/cinder master: [SVF]: Fix extend issue for a clone of rep-volume
*** alexe9191 has quit IRC13:12
*** ociuhandu has joined #openstack-cinder13:12
*** dsariel has joined #openstack-cinder13:17
openstackgerritMerged openstack/cinder-specs master: setup.cfg: Replace dashes with underscores
*** ociuhandu has quit IRC13:27
*** ociuhandu has joined #openstack-cinder13:51
rosmaitaCourtesy reminder: Cinder meeting in #openstack-meeting-alt at 1400 UTC13:57
rosmaitajungleboyj rosmaita smcginnis tosky whoami-rajat m5z e0ne geguileo eharney walshh_ jbernard lseki sfernand rajinir enriquetaso hemna ^^13:57
whoami-rajatthanks rosmaita13:58
*** ociuhandu has quit IRC13:58
*** Luzi has quit IRC13:59
*** ociuhandu has joined #openstack-cinder14:00
*** e0ne has joined #openstack-cinder14:01
*** m75abrams has quit IRC14:13
openstackgerritRajat Dhasmana proposed openstack/cinder stable/train: [SVF]:Reduce slowness by caching pool information
openstackgerritEric Harney proposed openstack/cinder stable/train: Backup manager: Synchronously call remove_export
openstackgerritEric Harney proposed openstack/cinder stable/train: API validation: Add cinder_host type to support ipv6 in manage
whoami-rajatrosmaita: so the tests are for required features only?15:01
enriquetaso#startmeeting cinder_bs15:01
openstackMeeting started Wed May  5 15:01:59 2021 UTC and is due to finish in 60 minutes.  The chair is enriquetaso. Information about MeetBot at
openstackUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.15:02
*** openstack changes topic to " (Meeting topic: cinder_bs)"15:02
openstackThe meeting name has been set to 'cinder_bs'15:02
enriquetasoI don't have much for today's meeting. Cinder has eleven more reported bugs from 2021-04-21 to 2021-05-05.15:02
enriquetasoYou can check them in the mailing list. Topic: '[cinder] Bug deputy report for week of 2021-05-05 - OpenStack-dev mailing list'15:02
enriquetaso#topic bug_1 'Cinder backup python3 encoding issue'15:02
*** openstack changes topic to "bug_1 'Cinder backup python3 encoding issue' (Meeting topic: cinder_bs)"15:02
openstackLaunchpad bug 1925809 in Cinder "cinder backup python3 encoding issue" [Medium,New]15:02
enriquetasoSummary: Looks like there may be a couple of encoding python3 issues in the backup ceph driver.15:02
enriquetasoThis can be fixed by encoding the zero stream like:15:03
*** amar7ibm has quit IRC15:03
enriquetaso                zeroes = '\0' * self.chunk_size15:03
enriquetaso                zeroes = zeroes.encode('utf-8')15:03
enriquetasothough I'm not 100% sure if that does anything untoward or properly zeroes the remaining space.  What do you think about this fix?15:03
*** amar7ibm has joined #openstack-cinder15:03
eharneyif it's a bytes vs str confusion bug we should fix it15:03
eharney(and then land type annotations that would prevent that...)15:04
enriquetasoyes, mypy is important15:05
enriquetaso#action (enriquetaso) proposed a fix for that15:05
rosmaitai thought \0 was the null character?15:05
enriquetasoi think so15:05
rosmaitathat should be the same in ascii and utf-8, i think15:06
eharneyit failed with a TypeError indicating it wants bytes rather str regardless of if they're the same15:06
eharneyrather than*15:06
rosmaitacurrent error is 2021-04-23 09:43:02.803 20 ERROR cinder.backup.manager TypeError: a bytes-like object is required, not 'str'15:06
eharneyi think that may come from some of the C ceph lib bindings which check such things now15:07
eharneybut it's not too obvious from the backtrace15:07
enriquetasoneed to debug a bit more then15:08
toskydo we know which version of ceph *and* libraries?15:08
eharneyi don't think it matters too much, it probably would have been better to pass bytes on the older versions too15:08
enriquetasocinder (16.2.1) (though I believe it will affect all versions including master)15:08
eharneyone of the bugs i fixed for ceph pacific support was a similar thing where we used to pass str and it no longer allowed it15:09
enriquetasoOK, i'll search your patch to see how did you fix it15:10
enriquetasoany other suggestions? moving on..15:11
enriquetaso#topic Open Discussion15:11
*** openstack changes topic to "Open Discussion (Meeting topic: cinder_bs)"15:11
rosmaitaonly thing i'm wondering is wehther we should do the conversion first15:11
rosmaitawhat i mean is, we create a chunk_size str and then convert to bytes15:12
enriquetasorosmaita, good question, let me check the pacific patch quick15:12
rosmaitai wonder whether we should create a chunk_size byte array to begin with?15:12
eharneyfor "zeroes"?  yes probably15:13
eharneywe don't need to make a str and then encode it..15:13
rosmaitaalso looks like we'll need to fix this line:
eharneyi'm not sure what you want to get from 773694.  i was just noting that current ceph libraries enforce type checks about bytes vs. str where they didn't previously15:15
rosmaitawell, it does remind me that b'\o15:16
rosmaitai mean15:16
rosmaitawill do what we want without encoding anything15:16
eharneyiirc you can just do bytearray(chunk_size) to get what is needed15:17
rosmaitaeven better15:18
eharneybecause it inits to 0s15:18
enriquetasocool, thanks15:19
rosmaitai like that way better15:19
rosmaitamuch clearer about what we're trying to accomplish there15:19
enriquetasoguess we don't have any other topic for today15:20
eharneyi'll mention one bug15:20
eharney is a good one that i just patched up15:21
openstackLaunchpad bug 1926630 in Cinder "Creating an encrypted volume type without a cipher accepted but will fail to create an encrypted volume" [Medium,In progress] - Assigned to Eric Harney (eharney)15:21
eharney(mostly asking for reviews i guess)15:21
eharneyi added validation around the issues directly pointed to by the bug, but i wouldn't be surprised if there are one or two more corner cases to find in this API or related ones15:21
openstackLaunchpad bug 1926630 in Cinder "Creating an encrypted volume type without a cipher accepted but will fail to create an encrypted volume" [Medium,In progress] - Assigned to Eric Harney (eharney)15:22
eharneybut basically an admin is currently allowed to create encryption types that don't work, which seems to not be dangerous, but is definitely annoying since the CLI clients don't really make you do the right thing15:22
rosmaitai was looking at that yesterday15:22
eharneymy first patch hardens this up for existing types, and the second prevents admins from making bad types, basically15:23
rosmaitais there a canonical list of acceptable ciphers in x-y-z format?15:23
eharneyi think not, but tbh we've never advised that anyone use anything other than aes-xts-plain6415:23
rosmaitaok, definitely good to fix this15:24
eharneya few others are definitely possible and should work, but... probably not best practice15:24
enriquetasoso far, all the bugs related to encryption uses --cipher aes-xts-plain64 . There's a bug using cryptsetup --batch-mode luksFormat --key-file=- --cipher aes-xts-plain64 --key-size 256 /dev/dm-1915:26
eharneybtw, i tried to use our validation schema code to check some of this, and wrote my own code instead when the schema required fields specs didn't seem to work like i'd expect them to15:26
eharneyenriquetaso: that is the only configuration we've ever documented using15:26
eharneyexcept for when we used to document 512 bit keys i guess, but those don't even work in barbican15:27
enriquetasoso, the current validation schema code isn't working properly ?15:28
eharneythe current validation schema for encrypted type creation doesn't require one or two fields that probably should be required15:28
enriquetasoOK, thanks for working on this eharney15:29
rosmaitaeharney: and it does require an optional field!15:30
eharneydoes it?15:30
*** udesale has quit IRC15:30
rosmaitaline 3515:31
rosmaitabet it was a typo, they meant provider and cipher?15:31
eharneywell, requiring control_location is reasonable, but tbh the whole control_location thing probably needs a revisit anyway15:31
rosmaitawe're showing control_loc as optional in the api-ref:
eharneyi'm really not sure what the current state is of control_location=back-end15:32
eharneyrosmaita: humm15:33
eharneyit's been hardwired into my brain for so many years to pass it from the CLI that i probably haven't really thought about it..15:33
eharneyshould probably go investigate control_location=back-end and make sure it doesn't hold nasty surprises for an admin that doesn't really know how this works15:34
enriquetasoi should fill a bug for the control_location optional vs required15:35
rosmaitaenriquetaso: ++15:35
eharneyrosmaita: i guess it would make sense to make cipher optional and default to the one that everyone wants, maybe worth digging into15:35
rosmaitawe need to figure this out15:35
eharneyi'm a little unclear on how much should be validated via API schema vs. in the code etc...15:36
enriquetasoi have the same concern15:36
eharneya general sense i had while working on this is that if it's in the code, it's at least in front of you in the right place to understand it more explicitly, and you can throw more specific errors15:36
eharneybecause for something like encryption details, even if the API schema validates it, we probably need to validate it everywhere else anyway15:37
rosmaitawell, we can at least get the required/optional elements and their general datatypes into the schema correctly15:37
eharneyi think we need to decide which ones are really required/optional to do that15:37
eharneyalso our api-ref says key_size is usually 256 but shows 128 in the example   *facepalm*15:37
rosmaitawhat happens if you use --key_size None ?15:38
rosmaitai mean, the default isn't None, wouldn't it be an actual integer?15:39
eharneyi think i tried that yesterday and found that it is validated15:39
eharneybecause i was poking around trying to find concerns similar to what the bug reported15:39
rosmaitakey_size = {'type': ['string', 'integer', 'null'],15:40
rosmaitaso yeah15:40
eharneybut it has minimum: 0, so presumably it can't actually be null?  i don't know..15:41
eharneyusing MAX_INT for the maximum is pretty funny though15:41
rosmaitathat would be quite a key15:41
enriquetaso#action (enriquetaso) fix the key_size in the example
rosmaitaok, so it looks like we need some doc + schema + api-ref cleanup around encryption15:42
eharneyi think this still leaves my current patches ok to go but i'll look over them again15:43
enriquetasothanks eharney15:43
enriquetasoOK, so control_location is not optinal?15:44
rosmaitayes, i agree that your patches are good for now, and then we need to follow up15:44
rosmaitaenriquetaso: it looks like the schema-validation will require it15:44
eharneyenriquetaso: it's probably reasonable for it to be optional, i'm not sure if it currently is or not15:44
rosmaitabut that doesn't mean that it's really required15:44
rosmaitai mean in the "big picture" sense of required15:45
enriquetasooh, ok15:45
enriquetasoany other topic/bug to discuss ?15:46
rosmaitanothing from me15:46
enriquetasoOK, i'll end up the meeting then15:48
enriquetasoSee you next week15:48
*** openstack changes topic to "The Block Storage Project | |"15:48
openstackMeeting ended Wed May  5 15:48:38 2021 UTC.  Information about MeetBot at . (v 0.1.4)15:48
openstackMinutes (text):
rosmaitathanks sofi!15:52
*** rpittau is now known as rpittau|afk15:52
*** pcheli has joined #openstack-cinder16:08
*** ociuhandu has quit IRC16:10
*** pcheli has quit IRC16:19
*** amar7ibm has quit IRC16:32
*** ociuhandu has joined #openstack-cinder16:33
*** ociuhandu has quit IRC16:33
*** ociuhandu has joined #openstack-cinder16:34
*** psachin has quit IRC16:36
*** dviroel|nho is now known as dviroel16:38
*** ociuhandu has quit IRC16:40
*** ociuhandu has joined #openstack-cinder16:43
*** zzzeek has joined #openstack-cinder16:45
*** ociuhandu has quit IRC16:48
*** zzzeek has quit IRC16:50
*** zzzeek has joined #openstack-cinder16:52
*** enriquetaso has quit IRC17:02
*** enriquetaso has joined #openstack-cinder17:02
*** zzzeek has quit IRC17:04
openstackgerritMerged openstack/cinder stable/ussuri: API validation: Add cinder_host type to support ipv6 in manage
*** zzzeek has joined #openstack-cinder17:08
*** pcheli has joined #openstack-cinder17:09
*** pcheli has quit IRC17:54
*** dsariel has quit IRC18:01
*** knikolla has joined #openstack-cinder18:10
*** gokhani has quit IRC18:13
*** MrClayPole_ has joined #openstack-cinder18:19
*** e0ne has quit IRC18:20
*** zenkuro has quit IRC18:22
*** MrClayPole has quit IRC18:23
*** zenkuro has joined #openstack-cinder18:24
*** ociuhandu has joined #openstack-cinder18:36
*** ociuhandu has quit IRC18:41
*** vishalmanchanda has quit IRC18:46
*** e0ne has joined #openstack-cinder19:01
*** e0ne has quit IRC19:22
*** e0ne has joined #openstack-cinder19:38
*** martinkennelly has quit IRC20:07
*** jmlowe has quit IRC20:08
*** jmlowe has joined #openstack-cinder20:09
*** martinkennelly has joined #openstack-cinder20:14
*** e0ne has quit IRC20:26
*** e0ne has joined #openstack-cinder20:38
openstackgerritEric Harney proposed openstack/cinder-tempest-plugin master: WIP: Test data on encrypted volume clone operations
openstackgerritMerged openstack/cinder stable/train: [Pure] Fix failing consistency group tempest tests
*** e0ne has quit IRC21:31
*** tosky has quit IRC23:03
*** rcernin has joined #openstack-cinder23:04
openstackgerritSofia Enriquez proposed openstack/cinder master: [rbd] Fix create encrypted volume from snapshot

Generated by 2.17.2 by Marius Gedminas - find it at!