Wednesday, 2019-08-14

*** rcernin has joined #openstack-cinder00:08
*** whoami-rajat has joined #openstack-cinder00:09
*** markvoelker has joined #openstack-cinder00:10
*** dviroel has quit IRC00:12
*** markvoelker has quit IRC00:16
*** KeithMnemonic has joined #openstack-cinder00:20
*** markvoelker has joined #openstack-cinder00:21
openstackgerritzhufl proposed openstack/cinder master: [Trivial]Remove never used helpers
openstackgerritzhufl proposed openstack/cinder master: [Trivial]Remove unused helpers
openstackgerritzhufl proposed openstack/cinder master: [Trivial]Remove never used helpers
*** trident has quit IRC00:38
*** dviroel has joined #openstack-cinder00:39
*** KeithMnemonic has quit IRC00:47
*** KeithMnemonic has joined #openstack-cinder00:49
*** rcernin has quit IRC00:58
*** spsurya has joined #openstack-cinder01:05
openstackgerritBrin Zhang proposed openstack/cinder master: Rollback the quota_usages table when failed to create a incremental backup
*** rcernin has joined #openstack-cinder01:13
*** baojg has joined #openstack-cinder01:17
*** markvoelker has quit IRC01:23
*** markvoelker has joined #openstack-cinder01:23
*** markvoelker has quit IRC01:28
*** Kuirong has joined #openstack-cinder01:47
*** markvoelker has joined #openstack-cinder01:56
*** markvoelker has quit IRC02:00
*** tejdeep has joined #openstack-cinder02:04
tejdeepHi everyone, rsd driver bug fix received +2 code reviews if everything looks ok can we merge this patch. This is needed by Intel RSD CI to support future patches.02:07
*** s0ru has quit IRC02:07
*** KeithMnemonic has quit IRC02:17
*** whoami-rajat has quit IRC02:28
*** boxiang has joined #openstack-cinder02:59
openstackgerritCY Chiang proposed openstack/cinder master: doc: Fix 3PAR hpe3par:convert_to_base extra spec is documented
*** strigazi has quit IRC03:09
*** strigazi has joined #openstack-cinder03:10
*** dviroel has quit IRC03:12
*** sapd1_x has joined #openstack-cinder03:15
*** whoami-rajat has joined #openstack-cinder03:20
*** psachin has joined #openstack-cinder03:33
*** gkadam has joined #openstack-cinder03:41
*** gkadam has quit IRC03:41
*** gkadam has joined #openstack-cinder03:43
*** tejdeep has quit IRC03:50
*** udesale has joined #openstack-cinder03:54
*** gkadam has quit IRC04:00
*** markvoelker has joined #openstack-cinder04:01
*** markvoelker has quit IRC04:05
*** sapd1_x has quit IRC04:08
*** markvoelker has joined #openstack-cinder04:52
*** markvoelker has quit IRC04:57
openstackgerritRajat Dhasmana proposed openstack/cinderlib master: Follow Up: Fixing NITs
openstackgerritzhufl proposed openstack/cinder master: [Trivial]Remove never used helpers
*** dpawlik has joined #openstack-cinder06:11
*** tejdeep has joined #openstack-cinder06:15
*** brinzhang_ has joined #openstack-cinder06:16
*** rcernin has quit IRC06:16
*** brinzhang_ has quit IRC06:18
*** brinzhang_ has joined #openstack-cinder06:18
*** brinzhang has quit IRC06:19
*** brinzhang_ has quit IRC06:19
*** brinzhang_ has joined #openstack-cinder06:19
*** dpawlik has quit IRC06:20
*** dpawlik has joined #openstack-cinder06:23
*** rcernin has joined #openstack-cinder06:30
*** lemko has joined #openstack-cinder06:43
openstackgerritRajat Dhasmana proposed openstack/cinderlib master: Add default volume type flow in dbms persistence plugin
*** trident has joined #openstack-cinder07:01
*** tesseract has joined #openstack-cinder07:10
*** kaisers has quit IRC07:21
*** e0ne has joined #openstack-cinder07:22
*** kaisers has joined #openstack-cinder07:23
*** tejdeep has quit IRC07:41
openstackgerritzhufl proposed openstack/cinder master: api-ref: add metadata parameter in the response of create-backup
openstackgerritZhengMa proposed openstack/cinder master: Leverage hw accelerator in image compression
*** rcernin has quit IRC07:56
*** helenafm has joined #openstack-cinder08:00
*** markvoelker has joined #openstack-cinder08:07
*** tkajinam has quit IRC08:11
*** markvoelker has quit IRC08:15
*** rcernin has joined #openstack-cinder08:23
*** rcernin has quit IRC08:29
whoami-rajatgeguileo: Hey, could you please take a look at . Thanks!08:37
*** rcernin has joined #openstack-cinder08:37
*** markvoelker has joined #openstack-cinder08:41
*** boxiang has quit IRC08:41
*** boxiang has joined #openstack-cinder08:41
*** markvoelker has quit IRC08:45
*** boxiang has quit IRC08:58
openstackgerritHelen Walsh proposed openstack/cinder master: PowerMax Docs -  Upgrade notification
*** boxiang has joined #openstack-cinder08:58
*** rcernin has quit IRC09:04
*** markvoelker has joined #openstack-cinder09:10
*** lemko has quit IRC09:13
*** markvoelker has quit IRC09:15
*** ociuhandu has joined #openstack-cinder09:49
*** Kuirong has quit IRC09:54
*** Kuirong has joined #openstack-cinder09:58
*** markvoelker has joined #openstack-cinder10:11
*** ociuhandu has quit IRC10:14
*** ociuhandu has joined #openstack-cinder10:15
*** baojg has quit IRC10:15
*** markvoelker has quit IRC10:15
*** baojg has joined #openstack-cinder10:16
*** dviroel has joined #openstack-cinder10:19
openstackgerritHelen Walsh proposed openstack/cinder master: PowerMax Docs -  Upgrade notification
openstackgerritHelen Walsh proposed openstack/cinder master: PowerMax Docs -  New features and supported software
*** brinzhang_ has quit IRC10:39
*** dpawlik has quit IRC10:41
*** deiter has joined #openstack-cinder10:48
*** eharney has quit IRC10:48
*** helenafm has quit IRC10:58
*** udesale has quit IRC11:05
*** Kuirong has quit IRC11:05
*** deiter has quit IRC11:08
*** dpawlik has joined #openstack-cinder11:08
*** udesale has joined #openstack-cinder11:12
*** ut2k3 has joined #openstack-cinder11:16
*** udesale has quit IRC11:16
ut2k3Hi guys, I have unfortunately two Volumes stock in "detaching" - Is there any chance I can force detach these from the instances? `nova volume-detach...` as well reseting once their state with `cinder reset-state...` did not help11:18
*** markvoelker has joined #openstack-cinder11:21
*** markvoelker has quit IRC11:26
*** markvoelker has joined #openstack-cinder11:31
*** carloss has joined #openstack-cinder11:34
*** markvoelker has quit IRC11:36
smcginnisut2k3: Did you try with the --attach-status argument? "cinder reset-state --attach-status detached"11:37
ut2k3Ah nice, I will try11:49
* smcginnis wonders if we should rethink our reset-state command since it always is a source of confusion11:53
*** markvoelker has joined #openstack-cinder12:02
*** ut2k3 has quit IRC12:06
*** ociuhandu has quit IRC12:14
*** rcernin has joined #openstack-cinder12:34
*** rcernin has quit IRC12:40
openstackgerritPablo Caruana proposed openstack/cinder master: Google Cloud Storage connection failure: 'bytes' object has no attribute 'encode'
*** KeithMnemonic has joined #openstack-cinder12:46
*** davidsha has joined #openstack-cinder12:47
*** mriedem has joined #openstack-cinder13:00
*** eharney has joined #openstack-cinder13:07
hemnasup peeps13:19
jungleboyjhemna:  Morning.13:19
*** deiter has joined #openstack-cinder13:20
potsHi folks, I was wondering if anyone had time to review the Seagate driver patch (renaming / enabling the old dothill driver).  I have some follow on work I'd like to do but it's awkward managing the renamed driver in the CI.  It's  Thanks!13:33
jungleboyjpots:  Thanks for the reminder.  Will take a look today.13:34
hemnapots: so this is basically a rename ?13:58
hemnait also changes the lenovo driver13:59
smcginnisThat's that whole OEM'd driver thing.14:00
potsyes, the current iteration of the patch also incorporates one another patch ( just so it can pass CI.  The rename affects the Lenovo and HPMSA drivers since they inherit the (renamed) classes.14:00
jungleboyjpots:  ++14:01
smcginnisWould be nice if there were Lenovo CI results on that.14:01
potsThere's a new CI (Seagate CI) that tests it as a standalone driver as well.14:01
smcginnisHPMSA is passing too. Guess we just need to mark the Lenovo driver as unsupported, right jungleboyj? :P14:02
potsI think you can see positive results from previous uploads (the last one is just a rebase).  Something broke in the CI yesterday that caused the retests to fail, but I've got them queued up.14:03
hemnaHPMSA might be tagged as passing, but there are no logs14:04
potsthe latest run failed in so no logs14:05
hemnawell, I can't +2 anything that doesn't have logs14:05
smcginnisThat shouldn't be reporting success.14:05
hemnaespecially for driver changes14:05
hemnathat's a failing CI14:05
jungleboyjsmcginnis:  We keep unsupporting drivers and we won't have anything left.14:06
hemnajungleboyj: well, how many unsupported do we have right now?14:06
potssure, I am re-running the CIs for -15, just wanted to show that they were passing yesterday with the same patch, so you can have some confidence they will pass again today.14:07
hemnapots: I'll look again when CI is passing14:07
jungleboyjWell, if we unsupport the ones that aren't passing py3 it would be something like 21 I think.  Would have to go count again.14:08
jungleboyjPlus 12 that just aren't running CI anymore.14:08
jungleboyjSo we have 30+ drivers that could be unsupported in Train.14:08
hemnaso 21 drivers aren't passying py3 ?14:09
hemnapassing even14:09
jungleboyjRight.  Give me a minute and I can count.14:11
*** helenafm has joined #openstack-cinder14:14
jungleboyjOk, since I sent out the e-mail we are down to 11 that are not Py3 compliant. 10 that are being marked unsupported because CI is no longer running.14:15
jungleboyjSo, 21 drivers.14:15
hemnacurrently have 6 marked as unsupported (in master)14:15
jungleboyjRight, the other 7 are the IBM drivers.  Waiting to hear back from them but I should probably just put those patchs up.14:15
hemnado you think it's worth creating a new flag that marks a driver as non py3 compliant?14:16
jungleboyjWe went round and round on that in last week's meeting.  Good discussion for next week.14:16
hemnathe driver would be unsupported still14:16
hemnabut the log message could be a bit different14:16
jungleboyjFor those drivers I was going to make it clear in the release note that that was the reason for the deprecation.14:17
hemnathat's probably a good idea, but I kinda like a different message14:17
hemnaat startup too14:17
jungleboyjhemna:  Hmmm, yeah, that might be good.  Lets chat about it next week.14:18
hemna"As of Train release Cinder is python3 only.  It was determined that this driver is not python3 compliant and is marked as unsupported."14:18
hemnasomething like that14:18
jungleboyjYeah, that makes sense.14:18
hemnawe could decide to treat those drivers slightly different than the typical unsupported flag14:18
eharneyTrain isn't python 3 only, right?14:19
hemnayah, it was just an example message14:19
*** dpawlik has quit IRC14:20
openstackgerritMerged openstack/cinder stable/queens: Prevent double-attachment race in attachment_reserve
hemnaso we basically have 21 drivers that we aren't getting support from vendors14:21
hemnaeither due to lack of fixing py3 issues or non running/working CI ?14:21
jungleboyjhemna:  Yep.14:21
*** ociuhandu has joined #openstack-cinder14:21
smcginnisWe could have a check if we are running under py3 and require that the unsupported flag be set in cinder.conf to continue to run those drivers that are not actually tested.14:23
smcginnisBut also need to warn that they will be gone in the next release if they don't get updated, because current plan is that Unnameable will be py3 only.14:23
hemnayah, that's a good idea14:23
jungleboyjUnnameable!  I love it.  That is almost as good as Unicorn.14:24
smcginnisTurning out to be a very fitting name. ;)14:24
*** ociuhandu has quit IRC14:26
jungleboyjsmcginnis:  Yeah, and now they are talking about not naming or doing other things.  All because one letter didn't line up well with the place in question.14:31
jungleboyjThis is why we can't have nice things.14:32
hemnanot naming the release?14:36
hemnawhat kind of nonsense is this14:36
jungleboyjYeah.  I like that Ed Leaf finally came in and noted that this was one very isolated case.14:37
openstackgerritJon Bernard proposed openstack/os-brick stable/stein: Check path alive before get scsi wwn
*** ociuhandu has joined #openstack-cinder14:57
*** boxiang has quit IRC14:58
*** boxiang has joined #openstack-cinder14:58
openstackgerritJon Bernard proposed openstack/os-brick stable/rocky: Check path alive before get scsi wwn
openstackgerritJon Bernard proposed openstack/os-brick stable/queens: Check path alive before get scsi wwn
*** helenafm has quit IRC15:06
*** ociuhandu has quit IRC15:11
*** ociuhandu has joined #openstack-cinder15:12
*** enriquetaso has joined #openstack-cinder15:17
*** ianychoi has joined #openstack-cinder15:20
*** dave-mccowan has joined #openstack-cinder15:21
*** senrique_ has joined #openstack-cinder15:22
*** senrique__ has joined #openstack-cinder15:22
*** enriquetaso has quit IRC15:24
pots_hemna: logs that were missing before for HPMSA driver CI of Seagate driver patch are there now--that run succeeded, it was just the copy to the log server that failed.  Still have to re-run the Lenovo CI on that test.15:25
potshemna: ^15:26
*** dave-mccowan has quit IRC15:26
*** senrique_ has quit IRC15:26
smcginnisBetter fix the issues first before rerunning the CI.15:26
openstackgerritEric Harney proposed openstack/cinder master: Tests: Combine similar encrypted volume tests w/ ddt
davidshasmcginnis: Hey, quick question, you're familar with swordfish?15:45
smcginnisdavidsha: Indeed I am.15:50
davidshasmcginnis: I was wondering would you know how swordfish would represent the NVMe/NQN endpoints? I know in RSD they have a OEM resource /Fabrics/Endpoints that holds that info, just wondering is there something else that could be used to make it closer to swordfish.15:53
*** enriquetaso has joined #openstack-cinder15:56
*** tesseract has quit IRC15:56
smcginnisdavidsha: Let me see if I can track anything down. Last month we did have a joint meeting with the NVMe working group to start to map the object models together. I think once we got past some terminology, the mapping was going better than I had thought it would. But quite possible we didn't get to this level of detail for something like this.15:56
*** senrique__ has quit IRC15:57
davidshasmcginnis: ack, thanks!15:58
smcginnisdavidsha: Not sure, but would it work with the combination of the PciID ClassCode in the Redfish endpoint:
smcginnisdavidsha: Along with the TransportProtocol from
smcginnisdavidsha: I guess I don't know enough technical details of how that's all represented to know what would really be needed.16:00
davidshasmcginnis: From what I can see from the rsd cinder driver, we need a host and target nqn with and IPv4 address and tcp port, but what you've given me is perfect! Thanks!16:02
smcginnisdavidsha: Awesome, glad I could help.16:05
*** sfernand has joined #openstack-cinder16:13
*** tejdeep has joined #openstack-cinder16:14
*** dave-mccowan has joined #openstack-cinder16:15
*** thgcorrea has joined #openstack-cinder16:15
*** throne82 has joined #openstack-cinder16:15
*** tejdeep has quit IRC16:23
*** dklyle has quit IRC16:36
*** dklyle has joined #openstack-cinder16:37
*** ociuhandu_ has joined #openstack-cinder16:41
*** fungi has quit IRC16:42
*** fungi has joined #openstack-cinder16:43
*** ociuhandu has quit IRC16:44
*** markvoelker has quit IRC16:44
*** ociuhandu_ has quit IRC16:45
*** e0ne has quit IRC16:51
*** davidsha has quit IRC16:51
*** markvoelker has joined #openstack-cinder16:51
*** tejdeep has joined #openstack-cinder16:53
*** ociuhandu has joined #openstack-cinder16:56
*** psachin has quit IRC16:57
*** senrique_ has joined #openstack-cinder16:59
*** ociuhandu has quit IRC17:00
*** enriquetaso has quit IRC17:01
potssmcginnis: your review suggestions would change the config options for the lenovo/hpmsa drivers.  Is there an accepted protocol for migrating from vendor-specific to generic options, e.g. do I need to mark the old ones deprecated and accept either one for at least one cycle?17:04
smcginnispots: Yep, exactly. There are oslo.config settings to mark a config as deprecated. Then in the code it's best to check for the old config and use that instead of the new config's default value until the deprecation period is over.17:05
smcginnisSorry, I know it's existing for the other drivers, but I don't think we want to introduce it for the seagate one.17:05
jungleboyjsmcginnis:  ++ Yeah, that is probably good to do.  Bummer that it impacts the other drivers.17:07
smcginnisWell, it helped identify that in the other drivers. Those are really good ones that we should get cleaned up.17:08
potssmcginnis: ok, sounds good--i just didn't want to exceed the scope of what should be in a single patch.17:09
smcginnisGood call, thanks for that. But since it impacts the "new" Seagate driver, I think it's in scope to avoid adding something that we want to remove.17:10
jungleboyjsmcginnis:  ++ Yeah, guess that is a win win.17:11
*** spsurya has quit IRC17:14
potssmcginnis: you called out the "<vendor-name>_backend_name" option.  this represents the name of the pool on the hardware array associated with the backend.  I think it's confusing but didn't want to make an incompatible change.  Do you think this should be changed as well, e.g. to "lenovo_storage_pool"?   It seems like there should be a generic opt17:28
potsion for this.17:28
smcginnispots: Ah, I see. Yeah, that is a little confusing. I would recommend deprecating the current backend name option since that has different meaning in Cinder as a whole and adding a new one like you suggest.17:29
*** dpawlik has joined #openstack-cinder17:43
*** e0ne has joined #openstack-cinder18:08
*** e0ne has quit IRC18:10
*** ociuhandu has joined #openstack-cinder18:11
*** eharney_ has joined #openstack-cinder18:15
*** dpawlik has quit IRC18:15
*** eharney has quit IRC18:17
*** ociuhandu has quit IRC18:34
*** sfernand has quit IRC18:53
*** thgcorrea has quit IRC18:54
*** boxiang has quit IRC19:19
*** kaisers has quit IRC19:32
*** kaisers has joined #openstack-cinder19:34
*** deiter has quit IRC19:35
openstackgerritSean McGinnis proposed openstack/cinder master: Default API v2 to disabled
*** eharney_ has quit IRC19:38
*** senrique__ has joined #openstack-cinder19:42
*** senrique_ has quit IRC19:45
*** my_nd has joined #openstack-cinder19:45
my_ndHi there, is there a way to get an estimated time for a creation of a backup using cinder-backup? I have an openstack pike environment using NFS as a cinder-backup backend19:53
smcginnismy_nd: Not really. We can't really tell ahead of time how long something like that will take since there are so many variables.19:54
my_ndthanks @smcginnis19:55
my_ndand why when creating a backup for a in-use volume...first cinder creates a temporal volume called backup--vol--xxxxx? is this temporal volume using disk space of the cinder-volume backend, rigth?19:58
smcginnismy_nd: If a volume is in-use, first a snapshot needs to be made so we can access the data to backup and it's a consistent point in time. Depending on the storage, that may consume some additional space until the backup operation completes and the snap can be removed.19:59
my_ndwhen you indicate that first a snapshot needs to be made, it means that cinder does this snapshot automatically when executing openstack volume backup create --force or I have to do it manually?20:07
smcginnismy_nd: Cinder will take care of creating and deleting it.20:08
my_ndthanks smcginnis20:10
smcginnisNo problem20:10
*** e0ne has joined #openstack-cinder20:14
*** senrique__ has quit IRC20:17
*** dpawlik has joined #openstack-cinder20:45
*** eharney_ has joined #openstack-cinder21:05
*** e0ne has quit IRC21:11
*** e0ne has joined #openstack-cinder21:13
*** e0ne has quit IRC21:13
*** markvoelker has quit IRC21:26
*** dpawlik has quit IRC21:35
*** mriedem has quit IRC21:53
*** throne82 has quit IRC22:05
*** carloss has quit IRC22:14
*** markvoelker has joined #openstack-cinder22:23
*** eharney_ has quit IRC22:30
*** markvoelker has quit IRC22:48
*** tkajinam has joined #openstack-cinder22:50
*** tyreymer has joined #openstack-cinder23:05
*** markvoelker has joined #openstack-cinder23:25
*** markvoelker has quit IRC23:30
*** tejdeep has quit IRC23:38
*** tyreymer has quit IRC23:41
*** tejdeep has joined #openstack-cinder23:55

Generated by 2.15.3 by Marius Gedminas - find it at!