Monday, 2019-07-29

« Sunday, 2019-07-28 Index Tuesday, 2019-07-30 »
*** tejdeep has quit IRC00:06
*** brinzhang has joined #openstack-cinder00:24
*** brinzhang has quit IRC00:32
*** gemini_117 has joined #openstack-cinder00:40
*** brinzhang has joined #openstack-cinder01:00
*** brinzhang has quit IRC01:05
*** baojg has joined #openstack-cinder02:03
*** brinzhang has joined #openstack-cinder02:08
*** brinzhang_ has joined #openstack-cinder02:09
*** brinzhang_ has quit IRC02:10
*** brinzhang_ has joined #openstack-cinder02:10
*** brinzhang_ has quit IRC02:11
*** brinzhang_ has joined #openstack-cinder02:12
*** brinzhang has quit IRC02:13
*** eharney has quit IRC02:23
*** markvoelker has joined #openstack-cinder02:29
*** rcernin has quit IRC02:45
*** rcernin has joined #openstack-cinder02:47
*** bhagyashris has joined #openstack-cinder02:49
*** m75abrams has joined #openstack-cinder03:09
*** psachin has joined #openstack-cinder03:31
*** psachin has quit IRC03:34
*** psachin has joined #openstack-cinder03:37
*** psachin has quit IRC03:55
*** udesale has joined #openstack-cinder03:59
*** brinzhang has joined #openstack-cinder04:04
*** brinzhang_ has quit IRC04:07
*** gemini_117 has quit IRC04:11
openstackgerritfengjiankui proposed openstack/cinder master: query volume2  https://review.opendev.org/67320404:38
*** Kuirong has joined #openstack-cinder04:43
*** davee_ has quit IRC05:03
*** davee_ has joined #openstack-cinder05:03
*** Luzi has joined #openstack-cinder05:06
*** whoami-rajat has joined #openstack-cinder05:22
*** e0ne has joined #openstack-cinder05:41
*** gemini_117 has joined #openstack-cinder05:45
*** sahid has joined #openstack-cinder06:00
*** jc_ has quit IRC06:06
*** e0ne has quit IRC06:11
*** gemini_117 has quit IRC06:13
*** gemini_117 has joined #openstack-cinder06:15
*** brinzhang_ has joined #openstack-cinder06:16
*** brinzhang has quit IRC06:19
*** dpawlik has joined #openstack-cinder06:27
*** brinzhang_ has quit IRC06:28
*** jc_ has joined #openstack-cinder06:42
*** brinzhang has joined #openstack-cinder07:05
*** tesseract has joined #openstack-cinder07:08
*** rcernin has quit IRC07:11
*** tosky has joined #openstack-cinder07:16
*** bhagyashris has quit IRC07:19
*** pcaruana has joined #openstack-cinder07:31
*** lpetrut has joined #openstack-cinder07:51
*** ircuser-1 has joined #openstack-cinder07:57
*** bhagyashris has joined #openstack-cinder07:57
openstackgerritNaoki Saito proposed openstack/cinder master: NEC Driver: allow more than 4 iSCSI portals  https://review.opendev.org/66808808:21
openstackgerritNaoki Saito proposed openstack/cinder master: NEC Driver: allow more than 4 iSCSI portals  https://review.opendev.org/66808808:30
*** tkajinam has quit IRC08:30
*** lemko has joined #openstack-cinder08:42
*** gkadam has joined #openstack-cinder08:44
*** lpetrut has quit IRC08:58
*** tesseract has quit IRC08:58
*** gemini_118 has joined #openstack-cinder09:08
*** e0ne has joined #openstack-cinder09:08
*** gemini_117 has quit IRC09:11
*** gemini_118 is now known as gemini_11709:11
*** trident has quit IRC09:16
*** trident has joined #openstack-cinder09:17
*** tesseract has joined #openstack-cinder09:20
*** altlogbot_2 has quit IRC09:24
*** altlogbot_2 has joined #openstack-cinder09:24
openstackgerritAleksandr Podrepnyi proposed openstack/cinder master: Fix small typo in releasenotes about iscsi_helper.  https://review.opendev.org/67324509:35
openstackgerritPawel Kaminski proposed openstack/os-brick master: connectors/nvme: Wait until nvme device shows up in kernel  https://review.opendev.org/67203109:40
*** m75abrams has quit IRC09:49
*** bhagyashris has quit IRC10:02
openstackgerritMichael McAleer proposed openstack/cinder master: PowerMax Driver - Volume & Snapshot Metadata  https://review.opendev.org/67301310:09
openstackgerritRajat Dhasmana proposed openstack/cinderlib master: [DOC]:Correct variable name kv  https://review.opendev.org/67326210:19
*** brinzhang_ has joined #openstack-cinder10:27
*** brinzhang has quit IRC10:30
*** kaisers has quit IRC10:35
openstackgerritLuigi Toscano proposed openstack/python-cinderclient master: Migrate the functional job to Zuul v3  https://review.opendev.org/67278410:37
*** jawad_axd has joined #openstack-cinder10:38
*** gemini_117 has quit IRC10:39
*** yaawang has quit IRC10:40
openstackgerritGhanshyam Mann proposed openstack/cinder master: Run 'tempest-ipv6-only' job in gate  https://review.opendev.org/67190110:45
openstackgerritGhanshyam Mann proposed openstack/cinder master: Remove .zuul.yaml from irrelevant-files list  https://review.opendev.org/67326810:55
gmannwhoami-rajat: jungleboyj making tests run for  the .zuul.yaml changes - https://review.opendev.org/#/c/673268/10:57
rosmaitagmann: do you know that answer to Luigi's question at the bottom of https://review.opendev.org/#/c/671947/ ?11:00
gmannrosmaita: oh, there is already fix up. I will check11:01
rosmaitagmann: thanks11:01
*** gkadam has quit IRC11:02
gmannrosmaita: replied. i will say irrelevant-files only changes are very rare so it is ok to skip .zuul.yaml for everything11:07
*** lpetrut has joined #openstack-cinder11:07
rosmaitagmann: yes, i agree the changes are infrequent, but they seem to occur mostly at RC-time, when you notice some irrelevant change is taking 6 hours to merge ... and then when you edit the irrelevant files, if .zuul.yaml is not in there, that also takes 6 hours to merge11:10
gmannrosmaita: every RC time ? i think it should be one time changes. once project define the irrelevant-files then why we need to change.11:11
rosmaitawell, maybe they'll be all sorted out by the Z release !11:11
gmannbut it is more about you will have more risk to merge the breaking changes of CI than saving time for infrequent cases.11:13
rosmaitai guess, but you'd think someone revising a job would keep it a WIP until it had actually run ... likewise, the reviewers would check to see that the job ran before approving the change11:14
rosmaitabut anyway, i did not put a negative score on the patch, so my opinion isn't blocking anything11:15
gmannyeah sure. IMO manual check of such thing is always risky and error prone in term of what it can break.11:16
*** gkadam has joined #openstack-cinder11:17
rosmaitagmann: yeah, you're right, automation is better (though i wish it were faster in this case!)11:19
*** udesale has quit IRC11:26
openstackgerritLuigi Toscano proposed openstack/python-cinderclient master: Migrate the functional job to Zuul v3  https://review.opendev.org/67278411:37
*** n-saito has quit IRC11:41
*** gkadam has quit IRC11:42
*** gkadam has joined #openstack-cinder11:44
*** n-saito has joined #openstack-cinder11:49
*** dviroel has joined #openstack-cinder11:53
*** eharney has joined #openstack-cinder12:00
openstackgerritNaoki Saito proposed openstack/cinder master: NEC Driver: allow more than 4 iSCSI portals  https://review.opendev.org/66808812:05
*** Kuirong has quit IRC12:17
*** Kuirong has joined #openstack-cinder12:22
*** mvkr has quit IRC12:22
*** jrubenst has joined #openstack-cinder12:26
*** deiter has joined #openstack-cinder12:28
*** Conqueror has joined #openstack-cinder12:31
*** deiter has quit IRC12:38
*** n-saito has quit IRC12:40
*** mriedem has joined #openstack-cinder12:47
*** raghavendrat has joined #openstack-cinder12:49
raghavendrathi hemna: this is regarding https://review.opendev.org/#/c/657585/; HPE Storage CI [ISCSI and FC] have passed. lefthand CI is no longer supported, so its non-voting. Request you to please provide your feedback.12:51
openstackgerritLuigi Toscano proposed openstack/python-cinderclient master: Migrate the functional job to Zuul v3  https://review.opendev.org/67278412:55
openstackgerritLuigi Toscano proposed openstack/python-cinderclient master: Migrate the functional job to Zuul v3  https://review.opendev.org/67278412:56
*** carloss has joined #openstack-cinder12:57
*** jroll has quit IRC13:07
*** jawad_axd has quit IRC13:08
*** jroll has joined #openstack-cinder13:08
*** jmlowe has joined #openstack-cinder13:13
toskygeguileo: hi, ahem, another question on cinderlib and its current set of jobs13:19
BLZbubbais anyone here good with lvm & mdadm block size issues?13:19
toskygeguileo: if I'm not mistaken, the functional jobs inside the cinderlib repository are based on the "functional" tox environment, which uses stestr13:19
*** mvkr has joined #openstack-cinder13:20
toskybut the runner used by the devstack jobs in cinder uses unit2; would it make sense to use the 'functional' tox environment everywhere?13:20
*** enriquetaso has joined #openstack-cinder13:21
*** jdillaman has joined #openstack-cinder13:23
openstackgerritPawel Kaminski proposed openstack/cinder master: target/spdknvmf: Add max_queue_depth configuration parameter  https://review.opendev.org/67206413:27
geguileotosky: I think that's a good idea13:28
geguileotosky: I designed the cinderlib jobs trying to run the same thing one runs locally13:28
geguileotosky: didn't know that Cinder functional jobs were running using unit213:29
openstackgerritHelen Walsh proposed openstack/cinder master: PowerMax Driver - Miscellaneous improvements to delete  https://review.opendev.org/67329213:29
toskygeguileo: see https://opendev.org/openstack/cinder/src/branch/master/playbooks/tempest-and-cinderlib-run.yaml#L23 which calls13:30
toskygeguileo: https://opendev.org/openstack/cinder/src/branch/master/playbooks/cinderlib-run.yaml#L4413:30
geguileotosky: ooooooh, now I see what you mean13:30
geguileotosky: I think what we have makes sense13:31
geguileotosky: because in devstack you have everything installed on the host13:32
walshh_Is there anyone available to take a look at https://review.opendev.org/#/c/664860 and https://review.opendev.org/#/c/669835.  Thank you13:32
geguileotosky: whereas on the cinderlib jobs you install stuff running tox, right?13:32
toskygeguileo: but the functional tox environment uses sitepackages=True13:32
toskyso it should be the same13:32
geguileotosky: mmmmm, true, true13:33
geguileotosky: then I don't really have any preference, as long as the logs are saved.13:34
*** mvkr has quit IRC13:34
geguileotosky: probably using the tox env like you suggest is the proper way to do it13:35
toskygeguileo: ack, let see if I can simplify the code; right now the unit2 runner is not tested by cinderlib, but only outside cinderlib, so it's not the most desirable scenario13:36
toskyas it may break13:36
geguileothanks13:37
toskyalso, the default functional job should be python3-based according the on-going python3 community goal13:41
toskylet's see how to minimize the changes13:41
*** whoami-rajat has quit IRC13:42
*** dpawlik has quit IRC13:45
*** mvkr has joined #openstack-cinder13:47
*** usr2033 has joined #openstack-cinder13:52
*** dpawlik has joined #openstack-cinder13:52
*** dklyle has quit IRC14:00
*** dklyle has joined #openstack-cinder14:00
usr2033hi, i have a problem about cinder loadbalancer mysql connection. when active connections unuseable because of broken mysql backend, cinder continue to use those broken connections. Is there a setting or workaround?14:04
*** deiter has joined #openstack-cinder14:06
*** jmlowe has quit IRC14:08
openstackgerritEric Harney proposed openstack/cinder master: Prevent double-attachment race in attachment_reserve  https://review.opendev.org/67137014:21
*** smrcascao has joined #openstack-cinder14:24
*** smrcascao has quit IRC14:30
*** dpawlik has quit IRC14:31
*** ykarel has joined #openstack-cinder14:32
ykareleharney, smcginnis can u check https://bugs.launchpad.net/cinder/+bug/183822514:35
openstackLaunchpad bug 1838225 in Cinder "tools/generate_driver_list.py broken with python2" [Medium,Triaged]14:35
*** lpetrut has quit IRC14:39
*** raghavendrat94 has joined #openstack-cinder14:42
raghavendrat94is Walter Boring online?14:44
*** raghavendrat94 has quit IRC14:54
smcginnisykarel: We only need to run with py3, so that bug is invalid.14:54
smcginnisykarel: I think hemna_ will want those driver_options fixed though since he was using that.14:56
hemna_sup14:56
smcginnishemna_: See bug mentioned above. We let a few drivers (back) in withouth implementing the get_driver_options.14:57
hemna_no worky?14:57
hemna_oh :(14:57
hemna_ok14:57
smcginnisShould be an easy fix I would assume.14:57
hemna_yah14:57
hemna_want me to handle it?14:57
smcginnisSure, you did all of the other ones, so you can probably crank that out fastest/14:58
hemna_sounds like we might need a job to run that tool during CI ?14:58
smcginnisWe do with the docs build, but apparently it only blows up with py2 which we don't actually support for those things anymore.14:59
smcginnisNot sure if we can add it to the interface since it's static, but tox -e compliance should fail on that if it handles static calls right.14:59
ykarelsmcginnis, cinder still supports python2, right?15:00
smcginnisykarel: Not for any tooling.15:00
smcginnisService runtime and client runtime only.15:00
smcginnisAll docs builds and other jobs are py3 only.15:01
ykarelsmcginnis, okk got it, possibly than that part(where issue occured) is not tested/reached without python2 in any other function job15:01
ykarels/without/with15:01
ykareli just shared that to get a reproducer15:02
smcginnisykarel: Thanks for letting us know.15:02
ykarelnp, Thanks for looking into it15:03
*** deiter has quit IRC15:04
*** Luzi has quit IRC15:06
*** zul has quit IRC15:21
*** udesale has joined #openstack-cinder15:22
hemna_so I ran the generate_driver_list.py with python3 and it worked15:24
smcginnishemna_: Does MacroSAN and Infortrend have that method you added to get the config options?15:24
hemna_nope15:26
hemna_but the tool didn't fail to run15:27
hemna_it generated a list of options for those drivers too15:27
hemna_.....15:27
openstackgerritLuigi Toscano proposed openstack/python-cinderclient master: Migrate the functional job to Zuul v3  https://review.opendev.org/67278415:27
hemna_which are the options from VolumeDriver15:27
hemna_heh15:27
smcginnisWeird that py2 has a problem with it and py3 doesn't.15:27
*** e0ne has quit IRC15:30
openstackgerritLuigi Toscano proposed openstack/cinder-tempest-plugin master: WIP Update cinder-tempest-plugin-lvm-lio (revamp)  https://review.opendev.org/67280415:30
jungleboyjhemna_:  Ooops.15:31
hemna_yah, I'm working on it15:31
jungleboyjGood that that was caught.15:31
openstackgerritEric Harney proposed openstack/cinder master: Prevent double-attachment race in attachment_reserve  https://review.opendev.org/67137015:32
*** udesale has quit IRC15:41
*** udesale has joined #openstack-cinder15:41
*** ykarel has quit IRC15:42
*** ykarel has joined #openstack-cinder15:42
*** mvkr has quit IRC16:01
*** raghavendrat has quit IRC16:02
openstackgerritWalter A. Boring IV (hemna) proposed openstack/cinder master: Add missing get_driver_options to new drivers  https://review.opendev.org/67332616:03
*** nikeshm has joined #openstack-cinder16:04
nikeshmHi kaminario ci is passing on py37 tests.  But we are using third party ppa16:06
nikeshmFor installing python 3.716:07
nikeshmas part of ci job before running tests16:08
jungleboyjnikeshm:  Sounds good!16:08
openstackgerritLuigi Toscano proposed openstack/python-cinderclient master: Migrate the functional job to Zuul v3  https://review.opendev.org/67278416:09
nikeshmAre we safe from unsupported flag or we have to use Ubuntu 18.04 in nodepool16:12
nikeshmWhich needs some time16:13
smcginnis18.04 isn't a requirement. Just running with Python 3.7.16:13
jungleboyjnikeshm:  No, we didn't set any requirements on the OS environment.  It was just required to have Python 3.7.16:13
jungleboyj:-)16:15
nikeshmOk.  Tests fail some times if third party ppa is down  or not responded16:15
nikeshmBut tests passed when python 3.7 installed correctly using dead snake ppa16:16
nikeshmEven i tried to build python 3.7 from source code but did not find any way in devstack to not use apt get  to install python 3.716:20
nikeshmBut use the build from source code while running job16:21
*** ganso has quit IRC16:24
*** sahid has quit IRC16:24
smcginnisHmm, I wonder what marker devstack looks for to determine it needs to try to do an apt install.16:24
smcginnisMight be possible to get it to stop trying that if you can track down where it makes that determination.16:25
jungleboyjnikeshm:  How often does the PPA fail?16:25
jungleboyjsmcginnis: If the 3rd party package doesn't work reliably it might be worth figuring out how to do that.16:25
smcginnisAnother option would be to set up a local mirror so even if the PPA goes down, you still have access to it via your local mirror.16:27
*** tesseract has quit IRC16:28
*** nikeshm has quit IRC16:30
openstackgerritEric Harney proposed openstack/cinder master: Prevent double-attachment race in attachment_reserve  https://review.opendev.org/67137016:39
*** N3l1x has joined #openstack-cinder16:43
*** ganso has joined #openstack-cinder16:56
*** udesale has quit IRC16:57
openstackgerritEric Harney proposed openstack/cinder master: Prevent double-attachment race in attachment_reserve  https://review.opendev.org/67137016:59
*** ykarel is now known as ykarel|away17:10
*** e0ne has joined #openstack-cinder17:14
*** Kuirong has quit IRC17:18
*** e0ne has quit IRC17:20
*** dpawlik has joined #openstack-cinder17:22
*** whoami-rajat has joined #openstack-cinder17:39
*** deiter_ has joined #openstack-cinder17:40
*** trident has quit IRC17:47
*** ykarel|away has quit IRC17:47
*** trident has joined #openstack-cinder17:51
*** nikeshm has joined #openstack-cinder18:13
*** lemko has quit IRC18:14
nikeshmjungleboyj smcginnis : ok, i will check the local mirror thing and frequency of failure in case of  third party ppa18:16
jungleboyjnikeshm:  Sounds like a good plan.18:16
nikeshmOne more thing we are using krest Library in kaminario driver.  Are we checking any thing else apart from py37 tests18:17
nikeshmDo we need to check any thing else also even if py37 tests pass18:18
nikeshmIn driver code18:19
nikeshmOr dependent library like krest18:19
*** nikeshm has quit IRC18:25
*** nikeshm has joined #openstack-cinder18:31
*** nikeshm has quit IRC18:36
jungleboyjhemna_:  You around?18:57
hemna_yah18:57
jungleboyjFinally getting back to playing with my Ceph iSCSI env.18:58
hemna_ah cool18:58
hemna_any luck? :P18:58
jungleboyjOh wait, give me a second.  Just had an idea.18:58
jungleboyjNope.18:59
jungleboyjhemna_:  So, there were a bunch of python packages I had to install beyond what was documented.19:00
jungleboyjNow I am at the point where it just fails to start with this:19:00
jungleboyjhttps://www.irccloud.com/pastebin/q3HWzk6j/19:00
hemna_yuhk19:01
jungleboyjIf I try to run it by hand I get:19:01
hemna_what kernel version are you on?19:01
jungleboyjhttps://www.irccloud.com/pastebin/T4fFMX62/19:01
jungleboyjhemna_: Linux ceph-iscsi-test 4.18.0-25-generic #26~18.04.1-Ubuntu SMP Thu Jun 27 07:28:31 UTC 2019 x86_64 x86_64 x86_64 GNU/Linux19:09
jungleboyjSo, it is greater than 4.1519:10
hemna_yah that should be ok19:10
hemna_that error you are seeing seems familiar19:11
jungleboyjI was hoping it might.19:11
hemna_gonna grab some lunch...bbiab19:12
jungleboyjOk.  Just found a bug online.  Ping me when you get back.19:13
*** e0ne has joined #openstack-cinder19:22
jungleboyjhemna_:  Did you need to hack things up like this?  https://bugzilla.redhat.com/show_bug.cgi?id=150364419:29
openstackbugzilla.redhat.com bug 1503644 in iSCSI "rbd-target-gw fails to start" [Unspecified,Closed: notabug] - Assigned to mchristi19:29
hemna_no I don't remember being required to hack the files up19:41
*** deiter_ has quit IRC19:45
*** jmlowe has joined #openstack-cinder19:57
*** jmlowe has quit IRC20:14
*** lemko has joined #openstack-cinder20:30
*** whoami-rajat has quit IRC20:39
BLZbubbaok this is a weird problem, my lvm backend always starts with free_capacity_gb set to 0.020:42
BLZbubbaeven though i have like 100TB available20:43
*** e0ne has quit IRC20:46
*** trident has quit IRC20:49
*** trident has joined #openstack-cinder20:52
BLZbubbaah it's the lvm_mirror option what is hosing me20:59
BLZbubbaam i the only person on the planet using this feature?21:20
*** dpawlik has quit IRC21:21
*** jrubenst has quit IRC21:21
BLZbubbamust be.  it improved performance over mdadm+lvm by a factor of like 20 for me21:38
BLZbubbabut it's just too buggy to use for real21:38
*** baojg has quit IRC21:53
*** mvkr has joined #openstack-cinder22:06
*** mriedem has quit IRC22:11
BLZbubbai think the answer may be to disable anything that is "thin"22:24
*** davee_ has quit IRC22:25
BLZbubbaotherwise it seems to try creating a raid10 LV on top of the thin pool's LV22:25
*** davee_ has joined #openstack-cinder22:26
BLZbubbawhich causes serious mayhem22:26
*** tosky has quit IRC22:32
*** eharney has quit IRC22:51
*** tkajinam has joined #openstack-cinder22:54
*** tejdeep has joined #openstack-cinder22:57
*** rcernin has joined #openstack-cinder23:02
*** lemko has quit IRC23:13
openstackgerritRafael Weingärtner proposed openstack/cinder master: Add mv to return volume type name and ID with volumes  https://review.opendev.org/66688623:21
*** n-saito has joined #openstack-cinder23:30
« Sunday, 2019-07-28 Index Tuesday, 2019-07-30 »

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