Monday, 2019-08-05

« Sunday, 2019-08-04 Index Tuesday, 2019-08-06 »
*** markvoelker has joined #openstack-cinder00:02
*** markvoelker has quit IRC00:06
*** brinzhang has joined #openstack-cinder00:31
*** lixiaoy1 has joined #openstack-cinder00:37
*** enriquetaso has joined #openstack-cinder00:44
*** markvoelker has joined #openstack-cinder00:45
*** markvoelker has quit IRC01:09
*** enriquetaso has quit IRC01:23
*** lixiaoy1 has quit IRC01:51
*** lixiaoy1 has joined #openstack-cinder01:52
*** markvoelker has joined #openstack-cinder02:01
*** markvoelker has quit IRC02:06
*** n-saito has joined #openstack-cinder02:10
*** enriquetaso has joined #openstack-cinder02:21
*** markvoelker has joined #openstack-cinder02:32
*** whoami-rajat has joined #openstack-cinder02:38
*** markvoelker has quit IRC02:42
*** enriquetaso has quit IRC03:07
*** markvoelker has joined #openstack-cinder03:13
*** markvoelker has quit IRC03:17
*** psachin has joined #openstack-cinder03:29
*** yaawang has quit IRC03:42
*** yaawang has joined #openstack-cinder03:43
*** udesale has joined #openstack-cinder04:07
*** dpawlik has joined #openstack-cinder04:15
*** dpawlik has quit IRC04:20
*** markvoelker has joined #openstack-cinder04:28
*** markvoelker has quit IRC04:33
*** Luzi has joined #openstack-cinder05:06
*** yaawang has quit IRC05:08
*** yaawang has joined #openstack-cinder05:10
*** vishalmanchanda has joined #openstack-cinder05:13
*** boxiang has quit IRC05:19
*** boxiang has joined #openstack-cinder05:20
*** yaawang has quit IRC05:29
*** yaawang has joined #openstack-cinder05:30
*** yaawang has quit IRC05:34
*** yaawang has joined #openstack-cinder05:35
*** bhagyashris has joined #openstack-cinder05:38
*** ruffian_sheep has joined #openstack-cinder05:43
*** dpawlik has joined #openstack-cinder05:43
ruffian_sheepwhoami-rajat :Hi05:43
whoami-rajatruffian_sheep: hey05:44
ruffian_sheepwhoami-rajat :Did the problematic document that was said last week been solved?05:44
whoami-rajatruffian_sheep: oh, I had some work at evening so couldn't discuss with the team05:45
whoami-rajatruffian_sheep: I will discuss today and let you know05:45
ruffian_sheepwhoami-rajat :Ok, look forward to your message.05:45
*** boxiang has quit IRC05:56
*** boxiang has joined #openstack-cinder05:57
*** boxiang has quit IRC05:58
*** sahid has joined #openstack-cinder05:59
openstackgerritSandStone Storage CI proposed openstack/cinder-specs master: add sandstone volume driver bp.  https://review.opendev.org/67453306:13
openstackgerritChris M proposed openstack/cinder master: Additional multiattach fixes for dothill driver  https://review.opendev.org/67453406:15
*** yaawang has quit IRC06:15
*** yaawang has joined #openstack-cinder06:15
*** dpawlik has quit IRC06:28
*** obre has joined #openstack-cinder06:29
*** yaawang has quit IRC06:32
*** yaawang has joined #openstack-cinder06:33
*** yaawang has quit IRC06:40
*** yaawang has joined #openstack-cinder06:41
*** ruffian_sheep has quit IRC06:43
*** dpawlik has joined #openstack-cinder06:44
*** lixiaoy11 has joined #openstack-cinder07:04
*** lixiaoy1 has quit IRC07:04
*** rcernin has quit IRC07:04
*** pcaruana has joined #openstack-cinder07:08
*** markvoelker has joined #openstack-cinder07:17
*** tesseract has joined #openstack-cinder07:17
*** tosky has joined #openstack-cinder07:19
*** gkadam has joined #openstack-cinder07:26
*** Kuirong has joined #openstack-cinder07:41
openstackgerritSandStone Storage CI proposed openstack/cinder-specs master: elimiate synax error.  https://review.opendev.org/67454507:48
*** markvoelker has quit IRC07:51
*** Kuirong has quit IRC07:51
openstackgerritSandStone Storage CI proposed openstack/cinder-specs master: elimitnate trailing whitespace.  https://review.opendev.org/67454708:02
*** tkajinam has quit IRC08:11
*** tesseract-RH has joined #openstack-cinder08:22
*** tesseract has quit IRC08:22
*** tesseract-RH has quit IRC08:24
*** tesseract has joined #openstack-cinder08:24
*** e0ne has joined #openstack-cinder08:48
*** lixiaoy11 has quit IRC08:54
*** markvoelker has joined #openstack-cinder08:56
*** markvoelker has quit IRC09:30
*** lemko has joined #openstack-cinder09:41
*** Luzi has quit IRC09:43
*** dasp has quit IRC09:49
*** dasp has joined #openstack-cinder09:49
*** spsurya has joined #openstack-cinder09:50
*** Luzi has joined #openstack-cinder09:56
gmannjungleboyj: and other cores, waiting for second review on this - https://review.opendev.org/#/c/671549/09:56
*** mvkr has joined #openstack-cinder10:18
*** ociuhandu has joined #openstack-cinder10:20
*** bhagyashris has quit IRC10:25
openstackgerritAbhishek Kekane proposed openstack/cinder master: Support multiple stores of Glance  https://review.opendev.org/66167610:34
*** markvoelker has joined #openstack-cinder10:38
*** baojg has quit IRC10:41
*** markvoelker has quit IRC10:43
*** baojg has joined #openstack-cinder10:46
*** baojg has quit IRC10:50
*** baojg has joined #openstack-cinder10:53
*** rosmaita has joined #openstack-cinder11:01
*** gkadam has quit IRC11:32
*** gkadam has joined #openstack-cinder11:49
*** ociuhandu has quit IRC11:49
*** ociuhandu has joined #openstack-cinder11:50
*** markvoelker has joined #openstack-cinder11:55
*** dviroel has joined #openstack-cinder12:00
*** markvoelker has quit IRC12:02
*** markvoelker has joined #openstack-cinder12:02
*** udesale has quit IRC12:06
*** udesale has joined #openstack-cinder12:07
*** markvoelker has quit IRC12:09
*** markvoelker has joined #openstack-cinder12:11
*** ociuhandu has quit IRC12:12
openstackgerritNaoki Saito proposed openstack/cinder master: NEC Driver: Storage assist retype and a bugfix  https://review.opendev.org/67458612:19
*** ociuhandu has joined #openstack-cinder12:20
*** jroll has quit IRC12:39
*** jroll has joined #openstack-cinder12:39
*** yaawang has quit IRC12:43
*** yaawang has joined #openstack-cinder12:45
*** lixiaoy1 has joined #openstack-cinder12:51
*** lixiaoy1 has quit IRC13:11
*** yaawang has quit IRC13:13
*** yaawang has joined #openstack-cinder13:13
*** mriedem has joined #openstack-cinder13:14
*** enriquetaso has joined #openstack-cinder13:15
*** cannibalcow has joined #openstack-cinder13:18
*** jrubenst has joined #openstack-cinder13:21
*** cannibalcow has quit IRC13:24
*** jsquare has quit IRC13:28
*** jsquare has joined #openstack-cinder13:28
*** ociuhandu has quit IRC13:32
openstackgerritrenminmin proposed openstack/cinder master: Correct new_type type in VolumeManager retype method  https://review.opendev.org/65798513:41
*** Luzi has quit IRC13:46
*** pcaruana has quit IRC13:47
*** pcaruana has joined #openstack-cinder14:00
*** ociuhandu has joined #openstack-cinder14:19
*** eharney has joined #openstack-cinder14:20
*** _erlon_ has joined #openstack-cinder14:23
*** mriedem has quit IRC14:29
*** mriedem has joined #openstack-cinder14:29
*** boxiang has joined #openstack-cinder14:35
*** boxiang has quit IRC14:36
*** boxiang has joined #openstack-cinder14:40
geguileosmcginnis: jungleboyj hemna_ https://review.opendev.org/#/c/67346714:42
geguileocould you have a look?14:42
jungleboyjgeguileo:  Sure.  Looking.14:44
geguileojungleboyj: thanks!14:44
jungleboyjNP.14:44
jungleboyjAfter reading the info on the 'U' naming ... Unicorn anyone?  Though I know that isn't the right answer.14:48
jungleboyjThough Shanghai and the Letter U lining up seems to be bad timing.14:48
whoami-rajatgeguileo: Hey, as you asked to load the default type id in the dbms plugin class constructor, i'm kind of stuck there, i can't use cinder db methods to fetch the type and i don't think passing the volume_type_id in the persistence_config everytime we load a plugin is the right way. Could you please elaborate on it :(14:49
geguileowhoami-rajat: I apologize, I'm a little bit lost on the conversation14:50
geguileowhoami-rajat: what's the patch?14:50
whoami-rajatgeguileo: i'm yet to put up the patch in cinderlib,14:52
whoami-rajata snippet of our conversation,14:52
whoami-rajatwhoami-rajatgeguileo: do we need to create a volume type for every volume created with no type. can't we reuse the __DEFAULT__ type already present in db?11:2314:52
whoami-rajatgeguileowhoami-rajat: you can11:2314:52
whoami-rajatgeguileowhoami-rajat: then you have to load its ID in https://opendev.org/openstack/cinderlib/src/branch/master/cinderlib/persistence/dbms.py#L4814:52
geguileowhoami-rajat: where are you creating the default volume type in Cinder?14:53
geguileowhoami-rajat: in the migration, right?14:53
whoami-rajatgeguileo: yes14:53
geguileowhoami-rajat: and you create it with a specific ID or a specific name?14:54
whoami-rajatgeguileo: specific name14:54
geguileoOK14:54
geguileothen I don't think you need to accept any new parameter in the __init__14:55
geguileoon init you search for the default volume type by name and get its ID14:56
geguileostoring it in the class14:56
geguileoand then when you are creating a volume without a volume type you use that ID instead14:56
geguileoand when you delete you don't delete the volume type if it's the default14:56
geguileothat would ensure that all volumes have a type, and you don't delete it when deleting the volume14:57
geguileosince in cinderlib each volume with qos or extraspecs have its own volume type14:57
whoami-rajatgeguileo: oh i did the same, but faced error while running tests http://paste.openstack.org/show/755527/15:00
geguileowhoami-rajat: that looks like the migration didn't happen...15:02
geguileobecause there's no volume_types table15:02
geguileowhoami-rajat: can I have a look at the code?15:03
whoami-rajatgeguileo: but the table exist15:04
whoami-rajathttp://paste.openstack.org/show/755528/15:04
whoami-rajatgeguileo: hmm, i reversed the code to find the issue and currently facing this while loading the type in __init__, single line causing error15:05
geguileowhoami-rajat: OOOOOOhhhh15:06
geguileowhoami-rajat: I think I know what you did15:06
geguileowhoami-rajat: you tried to load the type before running the migrations15:06
geguileolol15:06
geguileowhoami-rajat: you need to load the type after this line https://opendev.org/openstack/cinderlib/src/branch/master/cinderlib/persistence/dbms.py#L7415:07
whoami-rajatgeguileo: damn, i really need to read all the code15:07
geguileolol15:07
whoami-rajatgeguileo: thanks a lot for the quick solution :)15:08
geguileowhoami-rajat: no problem15:08
openstackgerritBhaa Shakur proposed openstack/cinder master: Zadara VPSA: Move to API access key authentication  https://review.opendev.org/67071515:12
openstackgerritMerged openstack/cinder master: Replace "integrated-gate-py3" template with new "integrated-gate-storage"  https://review.opendev.org/67154915:31
*** ociuhandu has quit IRC15:40
*** ociuhandu has joined #openstack-cinder15:40
*** ociuhandu has quit IRC15:41
*** ociuhandu has joined #openstack-cinder15:42
*** ociuhandu has quit IRC15:43
*** ociuhandu has joined #openstack-cinder15:43
*** ociuhandu has quit IRC15:43
*** enriquetaso has quit IRC15:43
*** ociuhandu has joined #openstack-cinder15:45
*** ociuhandu has quit IRC15:45
*** ociuhandu has joined #openstack-cinder15:45
openstackgerritJay Bryant proposed openstack/cinder master: Mark Storpool Driver Unsupported  https://review.opendev.org/67464216:04
*** tesseract has quit IRC16:04
*** ociuhandu_ has joined #openstack-cinder16:04
*** ociuhandu has quit IRC16:07
*** lpetrut has joined #openstack-cinder16:09
*** henriqueof has quit IRC16:12
*** e0ne has quit IRC16:19
geguileosmcginnis: jungleboyj thanks for the reviews!! :-)16:23
geguileojungleboyj: you asked me if I was going to Shanghai... turns out I am  XD16:24
geguileojungleboyj: are people going the whole week or just for the PTG?16:24
jungleboyjgeguileo: Awesome!  Glad you can be there!16:24
jungleboyjI will be there for the whole week.16:24
potsgeguileo: hi, can you take a peek at https://bugs.launchpad.net/cinder/+bug/1838820 ?16:29
openstackLaunchpad bug 1838820 in os-brick "Bad args to iscsiadm for iSCSI multipath target" [Undecided,New]16:29
openstackgerritMerged openstack/cinderlib master: [DOC]:Correct variable name kv  https://review.opendev.org/67326216:30
*** ociuhandu_ has quit IRC16:33
*** ociuhandu has joined #openstack-cinder16:33
potsgeguileo: someone said this was up your alley although it looks like the bug is from a ways back.16:34
*** enriquetaso has joined #openstack-cinder16:34
openstackgerritJay Bryant proposed openstack/cinder master: Mark Synology Driver Unsupported  https://review.opendev.org/67465016:37
geguileopots: looking16:45
geguileojungleboyj: OK, I think I'll be as well16:45
*** ociuhandu has quit IRC16:45
jungleboyjgeguileo:  Good.  That will help make meetings there more productive.16:45
geguileojungleboyj: or at least less boring  ;-)16:46
jungleboyj:-)16:46
openstackgerritChris M proposed openstack/cinder master: Additional multiattach fixes for dothill driver  https://review.opendev.org/67453416:47
geguileopots: do you have the previous log entries to see what the connect volume call received as input?16:48
*** markvoelker has quit IRC16:49
jungleboyjgeguileo:  Sounds good.  :-)16:49
potsi don't think so, sorry.16:49
*** ociuhandu has joined #openstack-cinder16:50
potsi thought this might be simple based on what the _iscsiadm_update16:50
potsgeguileo: i thought this might be simple based on what the _iscsiadm_update() function accepts16:51
geguileopots: it looks more like it's related to the values being returned by _get_node_startup_values16:53
geguileopots: and it could be related to a specific driver16:53
geguileopots: that is returning a tuple or something in the target_iqn16:55
geguileopots: is this happening with a Cinder driver at the gate?16:55
potsonly in my 3rd party CI runs for (e.g.) the Lenovo driver16:56
*** carloss has joined #openstack-cinder16:56
potsit affects anything that subclasses the dothill driver.16:56
potsi can't see how the current code code be correct, though, because iscsiadm_update() clearly expects a string, not a list?16:57
openstackgerritJay Bryant proposed openstack/cinder master: Update Veritas Access in Support Matrix  https://review.opendev.org/67465417:01
*** markvoelker has joined #openstack-cinder17:01
geguileopots: I'm not saying you are wrong, but a list should not render as ('iqn.2015-11.com.hpe:storage.msa2050.120114328c',)17:02
geguileopots: and I want to make sure there's nothing else at play here17:02
potsi didn't trace it all the way down through rootwrap, but you can see how whatever is in the target_iqn property gets passed as an argument to -T, and I'm guessing none of the code past that point protects itself from a non-string argument.17:04
pots(iscsiadm -T, that is)17:04
geguileopots: but why a list would render as a tuple?17:06
*** udesale has quit IRC17:07
potsbut it works if the iqn is not wrapped in a list.  i would say that the behavior of the code being called is undefined if the arguments are not valid.17:10
geguileopots: unlikely, it's more likely that this is caused by the communication between os-brick and privsep, that a list is converted to a tuple17:12
geguileopots: by the looks of the code that's probably it17:13
geguileopots: I'm surprised that drivers are actually working...17:13
potsi'm not sure how many drivers are testing multiattach with iscsi17:14
potser, sorry multipath17:14
potsi am testing multiattach but turned on multipath at the same time just to make things more complicated :(17:15
potsi think non-multipath iSCSI connections don't go through this code path17:16
geguileopots: I think most drivers try not to use discovery, since it's less resilient to faulty paths17:16
geguileopots: and most multipaths don't either, only if they use discovery17:16
potswhen you say discovery do you mean sendtargets?  I'm not sure how I would modify the driver to avoid this code path.17:18
geguileopots: yes, sendtargets happen when the Cinder driver is asking os-brick to use discovery17:20
potsgeguileo: so that would just involve setting 'disable_discovery = true' in cinder.conf?17:21
geguileopots: the problem using discovery is that if the IP that the driver returns for discovery is down, then the connections will not be stablished, even if the other N-1 paths are OK17:21
potsgeguileo: that's sort of a defect, isn't it?17:22
geguileopots: the disable_discovery is the conf option I added for Kaminario's K217:22
geguileopots: K2 only supported discovery, but it was a problem for me when doing tests, so I added support for not doing discovery even when I don't work for that company17:23
geguileopots: This is the patch that introduced that feature https://review.opendev.org/#/c/50814117:24
geguileopots: you can see that it's pretty easy to support non discovery, you just need to return target_iqns, target_portals, and target_luns17:24
geguileopots: in any case, give me a minute and I'll try to confirm the bug17:25
*** psachin has quit IRC17:28
*** ociuhandu has quit IRC17:28
*** ociuhandu has joined #openstack-cinder17:29
potsgeguileo: thanks.  i'm kind of reluctant to turn off discovery, in case it might break something else.  but i think i already return all the data in initialize_connection()17:30
*** ociuhandu has quit IRC17:31
*** sahid has quit IRC17:31
*** ociuhandu has joined #openstack-cinder17:31
openstackgerritMerged openstack/cinder master: Fix get_driver_options  https://review.opendev.org/67346717:32
openstackgerritJay Rubenstein proposed openstack/cinder master: SF remove deprecated sf_allow_template_caching  https://review.opendev.org/66473917:35
*** ociuhandu has quit IRC17:35
geguileopots: do you have a log on the CI where I can see what you are returning?17:37
geguileopots: I just tested Kaminario using discovery without privsep and it works fine17:38
geguileowhich is not all that meaningful17:39
potsLooks like most recent log is http://os-logs.tristero.net/seagate-ci/656724-13/HPMSA-iSCSI/17:39
potsthis would have my os-brick patch applied17:39
geguileopots: I'm in favor of your solution17:44
geguileopots: do you have a patch up for review?17:44
geguileopots: though it bothers me that it works without privsep :-(17:44
potsgeguileo: (sorry, i think that link was to a bad example, i didn't upload the logs when i reran that test).17:45
geguileopots: but I could see there the connection info17:45
potsgeguileo: i don't have a patch up, i've only ever submitted cinder patches and I wasn't sure if os-brick had a different core team, etc?17:45
geguileopots: same people in Cinder and OS-Brick17:45
geguileopots: if you submit it I can help you get it in, if you don't want to I can do it myself17:46
potsi can try, i just wasn't sure if there was a different procedure to follow, e.g. devstack installs brick with pip by default?17:47
geguileopots: yeah, devstack installs from pip, not master by default17:48
geguileopots: You should be able to ask it to use git with: LIBS_FROM_GIT=os-brick17:49
geguileoin your config file17:50
potsgeguileo: i'll give that a try, thanks!17:50
geguileopots: thank you !  :-)17:50
*** e0ne has joined #openstack-cinder18:05
*** dviroel has quit IRC18:16
openstackgerritChris M proposed openstack/os-brick master: Fix bad argument to iscsiadm in iSCSI discovery  https://review.opendev.org/67466918:20
*** tosky has quit IRC18:32
*** e0ne has quit IRC18:45
*** spsurya has quit IRC18:49
*** lpetrut has quit IRC19:05
*** e0ne has joined #openstack-cinder19:16
*** carloss has quit IRC19:23
*** e0ne has quit IRC19:31
*** lemko has quit IRC20:54
openstackgerritEric Harney proposed openstack/cinder master: Refactor API utilities into api_utils.py  https://review.opendev.org/67469821:01
*** vishalmanchanda has quit IRC21:03
*** eharney has quit IRC21:07
*** ircuser-1 has quit IRC21:10
*** mriedem has quit IRC21:19
*** mriedem has joined #openstack-cinder21:21
*** markvoelker has quit IRC21:29
*** whoami-rajat has quit IRC21:58
*** rcernin has joined #openstack-cinder22:06
*** spatel has joined #openstack-cinder22:12
*** lixiaoy1 has joined #openstack-cinder22:23
*** tosky has joined #openstack-cinder22:26
*** markvoelker has joined #openstack-cinder22:46
*** tosky has quit IRC22:47
*** tkajinam has joined #openstack-cinder22:56
*** markvoelker has quit IRC23:20
*** enriquetaso has quit IRC23:23
*** spatel has quit IRC23:33
*** lixiaoy1 has quit IRC23:40
*** enriquetaso has joined #openstack-cinder23:52
« Sunday, 2019-08-04 Index Tuesday, 2019-08-06 »

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