Friday, 2019-05-17

*** lseki has quit IRC00:20
*** carloss has quit IRC00:28
*** lixiaoy1 has joined #openstack-cinder00:38
*** mmethot has quit IRC01:12
*** gkadam has joined #openstack-cinder01:16
*** baojg has joined #openstack-cinder01:22
*** gkadam has quit IRC01:40
*** brinzhang has joined #openstack-cinder01:44
*** mmethot has joined #openstack-cinder01:55
*** ianychoi has joined #openstack-cinder02:00
*** whoami-rajat has joined #openstack-cinder02:12
*** enriquetaso has quit IRC02:31
openstackgerritYong Huang proposed openstack/cinder master: [VNX] Fix test case issue
*** minmin has joined #openstack-cinder03:00
*** awalende has joined #openstack-cinder03:13
*** awalende has quit IRC03:17
*** anks2k has joined #openstack-cinder03:20
anks2kwhoami-rajat: jungleboyj: hemna_: Would appreciate a review on
*** gkadam has joined #openstack-cinder03:23
*** gkadam has quit IRC03:23
*** gaoyan has joined #openstack-cinder03:28
*** anks2k has quit IRC03:36
*** anks2k has joined #openstack-cinder03:39
*** psachin has joined #openstack-cinder03:41
openstackgerritSimon Dodsley proposed openstack/cinder master: Pure Storage FlashArray iSCSI driver target CIDR support
*** ianychoi has quit IRC03:49
*** ianychoi has joined #openstack-cinder03:49
*** ianychoi has quit IRC03:52
*** ianychoi has joined #openstack-cinder03:52
*** gaoyan has quit IRC04:31
*** udesale has joined #openstack-cinder04:34
*** bhagyashris has joined #openstack-cinder04:55
*** gaoyan has joined #openstack-cinder05:14
*** pcaruana has joined #openstack-cinder05:20
*** brinzhang has quit IRC05:21
*** brinzhang has joined #openstack-cinder05:22
*** gaoyan has quit IRC05:29
openstackgerritRajat Dhasmana proposed openstack/cinder master: Fix volume type quota defaults in quotas tests
*** radeks has joined #openstack-cinder05:50
*** Luzi has joined #openstack-cinder05:52
*** vishalmanchanda has joined #openstack-cinder05:54
*** ruffian_sheep has joined #openstack-cinder05:54
ruffian_sheepjungleboyj|smcginnis|whoami-rajat:I modified my original environment and now I can generate an error-free html file. This is the content of the execution script volume_test and HTMLTestRunner files used. Does this meet the requirements for joining the main line of cinder?05:59
ruffian_sheepjungleboyj|smcginnis|whoami-rajat: volume_test  HTMLTestRunner
*** baojg has quit IRC06:01
whoami-rajatruffian_sheep: Hi, could you provide the link to your driver patch?06:02
ruffian_sheepjungleboyj|smcginnis|whoami-rajat: This result has to trouble you to get a text to open the view, thank you very much. result
ruffian_sheepYes, but I have not updated the submission yet. Can it be before?06:04
*** lpetrut has joined #openstack-cinder06:05
whoami-rajatruffian_sheep: not updated? you've pushed a patch for your driver right?06:05
whoami-rajatruffian_sheep: also i see some of the tests skipped like test_reset_group_snapshot_status , may i know the reason?06:06
*** faizy98 has quit IRC06:07
ruffian_sheep <whoami-rajat>:Yes, I have submitted a driver before, on the branch.
*** sapd1_x has joined #openstack-cinder06:10
ruffian_sheep <whoami-rajat>:Later, I learned more about the CI process, set up the CI and debug the errors. This is an earlier one, and has not been updated after debugging.06:10
whoami-rajatruffian_sheep: okay, i remember we discussed regarding this in the PTG, IMO updating the driver constantly to make it work should be the first priority else setting up the CI for a "unready" driver isn't worth06:12
*** raghavendrat has quit IRC06:14
ruffian_sheep <whoami-rajat>:Because we are already using our own developed cinder driver, we provide it directly for user. So we went directly to build CI06:15
ruffian_sheep <whoami-rajat>:  This is my configuration file, I did not find the settings for the group. In addition, the driver does not support group operations for the time being, and I see that the use cases are used to do group related operations.06:17
whoami-rajatruffian_sheep: so you've a ready driver in your environment, then i would recommend to update the driver patch ASAP06:19
whoami-rajatruffian_sheep: hmm, so consistency group isn't supported by your driver?06:20
ruffian_sheep<whoami-rajat>:Yes, I am currently planning to debug the use case after passing the tempest and then update the branch. So I want to confirm if there are any shortages.06:20
ruffian_sheep<whoami-rajat>:Yes, it is not supported now. It was tested first based on most of the features required by the driver. Is the consistency group necessarily done? I see the documentation showing that this is optional.06:22
whoami-rajatruffian_sheep: okay, but you can just update the driver and CI and then we can review both on gerrit, easier than sending html's for the result :)06:24
whoami-rajatruffian_sheep: yes, it is optional, i was just querying.06:25
*** sahid has joined #openstack-cinder06:25
whoami-rajatruffian_sheep: also that will allow reviewers in different TZ's to view your driver progress, i think i'm only one currently aware of the state of it.06:26
ruffian_sheepEmmm my company let me first pass the use case and then provide the CI environment, we have to get a separate network to the community. There is an extra charge, so let me start the pre-work and then open the network.06:27
ruffian_sheep<whoami-rajat>: I also think: (But the work environment cannot be directly connected to the community, so the implementation plan needs to consider the cost.06:29
*** baojg has joined #openstack-cinder06:36
ruffian_sheepjungleboyj|smcginnis|whoami-rajat:I am not very clear about the last problem remaining in debugging. This is my profile confidence. I didn't set up a keypair here, and the settings used to create the network are private or shared. But the way he authenticates is keypair, and the network used to build the instance is public, which I am not very clear about.06:36
ruffian_sheepjungleboyj|smcginnis|whoami-rajat:  /opt/stack/tempest/etc/tempest.conf  log
ruffian_sheepjungleboyj|smcginnis|whoami-rajat:  log from the cmd : tox -e all -- volume --concurrency=1|06:39
whoami-rajatruffian_sheep: are your referring to test_volume_boot_pattern failing?06:44
ruffian_sheep<whoami-rajat>: yes,it didn't follow my setting ing the configration to use the password.06:46
whoami-rajatruffian_sheep: i think that test has also caused problems  in other CI's. most of the CI's currently are ignoring that test AFAIK06:54
whoami-rajatruffian_sheep: also your driver doesn't currently support retyping right?06:55
ruffian_sheep<whoami-rajat>: support retyping {0} tempest.api.volume.admin.test_volume_retype.VolumeRetypeWithoutMigrationTest.test_available_volume_retype [8.479916s] ... ok06:56
whoami-rajatruffian_sheep: i mean retype with migration06:57
whoami-rajatruffian_sheep: anyways that's also not important. if you skip/disable the test_volume_boot_pattern test, your CI would be functioning correctly right?06:59
openstackgerritpengyuesheng proposed openstack/python-cinderclient master: Add Python 3.7 to tox
ruffian_sheep<whoami-rajat>: I think this is the case, I want to confirm whether the passed use cases meet the criteria for joining the main line.07:00
ruffian_sheepIf it is not enough, I will debug and add the required test cases. The retype with migration you said should be that I only used one backend to operate. I can try to turn this feature on. Try it again.07:01
whoami-rajatruffian_sheep: I won't be able to decide that alone. will need smcginnis  and jungleboyj  to take a look. but as far as i can tell, it looks ready to me.07:03
ruffian_sheep<whoami-rajat>:That is really great! (づ ̄3 ̄)づ╭❤~Love you07:04
whoami-rajatruffian_sheep: oh yeah, i saw multi-backend was commented in tempest.conf so associated tests skipped. i don't think that is much needed so you can leave it.07:05
whoami-rajatruffian_sheep:  oh i missed some cases07:08
*** tesseract has joined #openstack-cinder07:08
whoami-rajatruffian_sheep: the backup tests are being skipped07:09
*** awalende has joined #openstack-cinder07:11
*** awalende has quit IRC07:13
ruffian_sheep<whoami-rajat>:Backup is what needs to be added ?07:15
whoami-rajatruffian_sheep: no, multi-attach, backup and multi-backend functionality isn't required, i'm finding some cases missing in image_test ... still checking07:16
*** awalende has joined #openstack-cinder07:17
whoami-rajatruffian_sheep: these are some tests in compute directory that are missed by your regex07:19
*** rcernin has quit IRC07:19
whoami-rajatruffian_sheep:  there are 2 tests in tempest.scenarios but they are going to be skipped anyway. so adding the above tests i mentioned will cover all cases.07:36
*** mcgigglier has joined #openstack-cinder07:50
openstackgerritSimon O'Donovan proposed openstack/cinder master: PowerMax driver -  changing from 9.0 to 9.1 REST endpoints
*** lpetrut has quit IRC08:09
openstackgerritpengyuesheng proposed openstack/python-cinderclient master: Add Python 3.7 to tox
ruffian_sheep<whoami-rajat>:Maybe I should change the path to /opt/stack/tempest/tempest08:15
whoami-rajatruffian_sheep: and what will be the regex then?08:16
*** sapd1_x has quit IRC08:21
*** awalende has quit IRC08:24
*** wxy-xiyuan has left #openstack-cinder08:27
*** lixiaoy1 has quit IRC08:28
ruffian_sheep<whoami-rajat>:what about use the regex of test*?I use it to run the cmd tox -e all -- volume --concurrency=108:29
*** lpetrut has joined #openstack-cinder08:29
*** awalende has joined #openstack-cinder08:37
ruffian_sheep<whoami-rajat>:tempest.api.compute.servers.test_device_tagging&&tempest.scenario.test_volume_boot_pattern&&tempest.scenario.test_volume_migrate_attached right?08:40
*** tkajinam has quit IRC08:44
openstackgerritpengyuesheng proposed openstack/python-cinderclient master: Add Python 3.7 to tox
*** baojg has quit IRC08:45
openstackgerritLiang Fang proposed openstack/python-cinderclient master: upload-to-image returns NoneType error
ruffian_sheep<whoami-rajat>:I think the log of executing the tox output should include what you said. I have one use case left and I don't know how to solve it. When he wants to ssh my instance, it doesn't depend on my configuration file, it uses the keypair which I have not configured.08:46
*** baojg has joined #openstack-cinder08:50
*** davidsha has joined #openstack-cinder08:52
openstackgerrityenai proposed openstack/cinder master: Speed up starting cinder-backup
openstackgerritMerged openstack/os-brick stable/queens: Fix FC case sensitive scanning
*** baojg has quit IRC09:06
*** baojg has joined #openstack-cinder09:07
*** sapd1_x has joined #openstack-cinder09:13
*** raghavendrat has joined #openstack-cinder09:14
ruffian_sheep<whoami-rajat>:Is it that tox can only use keypair to ssh instance?09:14
openstackgerritSimon O'Donovan proposed openstack/cinder master: PowerMax driver - add 'target_array_model' to metadata
*** e0ne has joined #openstack-cinder09:17
openstackgerritSimon O'Donovan proposed openstack/cinder master: PowerMax driver - clone improvements
openstackgerritTrent Lloyd proposed openstack/cinder master: RBD: Add 512e/4k disk geometry configuration
openstackgerritTrent Lloyd proposed openstack/cinder master: RBD: Add 512e/4k disk geometry configuration
openstackgerritSimon O'Donovan proposed openstack/cinder master: PowerMax driver - locking improvements
openstackgerritSimon O'Donovan proposed openstack/cinder master: PowerMax driver - Rapid TDEV Deallocation
*** baojg has quit IRC09:32
*** bhagyashris has quit IRC09:36
*** sapd1_x has quit IRC09:36
openstackgerritSimon O'Donovan proposed openstack/cinder master: PowerMax driver - Rapid TDEV Deallocation
*** awalende has quit IRC09:49
*** sapd1_x has joined #openstack-cinder09:49
*** awalende has joined #openstack-cinder09:49
*** rcernin has joined #openstack-cinder09:50
*** minmin has quit IRC09:50
*** awalende has quit IRC09:50
*** awalende has joined #openstack-cinder09:50
*** minmin has joined #openstack-cinder09:51
*** minmin has quit IRC09:52
*** sapd1_x has quit IRC09:59 proposed openstack/cinder master: WIP: Ceph: Cinder volume revert to snapshot support
openstackgerritSneha Rai proposed openstack/cinder master: HPE 3PAR-Added support for multiattach feature in hpe3par cinder driver
*** awalende has quit IRC10:24
*** brinzhang has quit IRC10:28
*** whoami-rajat has quit IRC10:42
*** vishalmanchanda has quit IRC10:44
*** awalende has joined #openstack-cinder10:48
*** baojg has joined #openstack-cinder11:02
*** henriqueof has joined #openstack-cinder11:06
*** jcosmao has joined #openstack-cinder11:18
*** awalende has quit IRC11:23
*** rcernin has quit IRC11:36
*** awalende has joined #openstack-cinder11:39
*** awalende has quit IRC12:03
*** awalende has joined #openstack-cinder12:05
*** dviroel has joined #openstack-cinder12:09
pcaruanaHi, this one is needing some workflow love
*** mchlumsky has joined #openstack-cinder12:24
*** udesale has quit IRC12:30
*** udesale has joined #openstack-cinder12:31
*** dklyle has joined #openstack-cinder12:50
*** enriquetaso has joined #openstack-cinder13:03
raghavendratHi core reviewers, It would be great if anyone can have a look at and review13:07
*** jmlowe has quit IRC13:11
*** awalende has quit IRC13:15
*** enriquetaso has quit IRC13:17
*** mriedem has joined #openstack-cinder13:21
*** lseki has joined #openstack-cinder13:22
*** raghavendrat has quit IRC13:45
*** mmethot has quit IRC13:50
*** jmlowe has joined #openstack-cinder14:00
*** enriquetaso has joined #openstack-cinder14:05
*** lpetrut has quit IRC14:06
*** mmethot has joined #openstack-cinder14:13
*** baojg has quit IRC14:16
*** baojg has joined #openstack-cinder14:16
*** baojg has quit IRC14:16
*** baojg has joined #openstack-cinder14:17
*** baojg has quit IRC14:17
*** baojg has joined #openstack-cinder14:18
*** baojg has quit IRC14:19
*** baojg has joined #openstack-cinder14:19
*** baojg has quit IRC14:19
*** baojg has joined #openstack-cinder14:21
*** baojg has quit IRC14:21
*** baojg has joined #openstack-cinder14:22
*** baojg has quit IRC14:23
*** baojg has joined #openstack-cinder14:23
*** baojg has quit IRC14:23
*** baojg has joined #openstack-cinder14:24
*** baojg has quit IRC14:24
*** dpawlik has quit IRC14:24
*** baojg has joined #openstack-cinder14:25
*** Luzi has quit IRC14:25
*** baojg has quit IRC14:25
*** baojg has joined #openstack-cinder14:25
*** baojg has quit IRC14:26
*** baojg has joined #openstack-cinder14:26
*** baojg has quit IRC14:27
*** baojg has joined #openstack-cinder14:27
*** baojg has quit IRC14:27
*** baojg has joined #openstack-cinder14:28
*** baojg has quit IRC14:28
*** baojg has joined #openstack-cinder14:28
*** baojg has quit IRC14:29
*** baojg has joined #openstack-cinder14:29
*** baojg has quit IRC14:30
*** baojg has joined #openstack-cinder14:31
*** baojg has quit IRC14:31
*** baojg has joined #openstack-cinder14:32
*** baojg has quit IRC14:32
*** baojg has joined #openstack-cinder14:32
*** baojg has quit IRC14:33
*** baojg has joined #openstack-cinder14:33
*** baojg has quit IRC14:34
*** baojg has joined #openstack-cinder14:34
*** baojg has quit IRC14:34
*** baojg has joined #openstack-cinder14:35
*** baojg has quit IRC14:35
*** baojg has joined #openstack-cinder14:36
*** baojg has quit IRC14:36
*** baojg has joined #openstack-cinder14:36
*** baojg has quit IRC14:37
*** baojg has joined #openstack-cinder14:37
*** baojg has quit IRC14:38
*** baojg has joined #openstack-cinder14:38
*** baojg has quit IRC14:38
*** baojg has joined #openstack-cinder14:39
*** baojg has quit IRC14:39
openstackgerritMerged openstack/cinder stable/queens: NetApp: Return all iSCSI targets-portals
*** baojg has joined #openstack-cinder14:40
*** baojg has quit IRC14:41
*** baojg has joined #openstack-cinder14:41
*** jmlowe has quit IRC14:41
*** baojg has quit IRC14:42
*** baojg has joined #openstack-cinder14:42
*** baojg has quit IRC14:42
*** baojg has joined #openstack-cinder14:43
*** baojg has quit IRC14:43
*** kmalloc is now known as kmalloc_away14:47
*** carloss has joined #openstack-cinder14:54
*** jmlowe has joined #openstack-cinder14:58
*** mcgigglier has quit IRC15:05
smcginnisjungleboyj, hemna_: I've noticed we have a lot more little things that new driver developers are just supposed to somehow know.15:14
jungleboyjsmcginnis:  Morning.  Was just about to ping you.15:14
smcginnisI plan on adding docs to our actually docs instead of our wiki, but started putting together a sort of checklist of things they can go through and make sure they are ready.15:14
jungleboyjLooking at the Quantastor driver.15:14
*** qqmber has joined #openstack-cinder15:14
smcginnisWould like to group source that into a good list15:14
smcginnisStarted here:
jungleboyjsmcginnis:  ++  Cool.15:15
qqmbernope.. there is errors.. but I don't understand them... i guess15:15
qqmberforget last line15:15
qqmberhi everyone... I'm using Cinder, no Ceph for anything... everything is "Dockerized" (kolla-ansible). In tgtd log appears a LOOOOOOT this:15:16
qqmber{"log":"tgtd: iscsi_noop_out_rx_start(1694) ffffffff 31 0\n","stream":"stderr","time":"2019-05-17T15:05:33.718926452Z"}15:16
qqmber600GB of logs with this in 3 days15:16
jungleboyjhemna_:  ^^^15:16
woojaysmcginnis: I'll be happy to add details to your guide.15:18
jungleboyjwoojay: ++15:18
smcginnisjungleboyj: What does "A failed run should include the information for rerunning." mean?15:18
jungleboyjThat the CI should post the information for doing a recheck.15:19
smcginnisWe have a required third party recheck trigger, why would we need that?15:20
jungleboyjI really appreciate CIs that do that so I can just copy and paste it and not worry about getting it right.15:20
jungleboyjAnd because I always forget the the recheck trigger is and have to go look it up.15:20
smcginnisPrepending "run-" to a copy paste of their name is so hard. :{15:20
jungleboyjYou have to go look up the name as well.15:21
*** sapd1_x has joined #openstack-cinder15:21
smcginnisYou can read text that includes the recheck trigger in the comment but not read the name on the comment?15:21
jungleboyjLenovo Storage CI is not the name to do the retrigger.15:22
smcginnisIt better be, that's are documented required recheck trigger.15:22
* smcginnis gets unsupported patch ready. 15:22
smcginnishehe :)15:22
jungleboyjNo.  Am I stupid here?15:23
jungleboyjFor example Dell EMC ScaleIO CI15:23
jungleboyjTo re run you have to do run-dell-emc-vxflexos to retrigger.15:23
smcginnisAccording to what we documented as required, that should be "run-Dell EMC ScaleIO CI"15:23
smcginnisAnd even though I keep commented on their patches to stop using that form, they still do.15:23
jungleboyjThe name of the CI Account is not the same as the CI Account.15:23
smcginnisHaven't checked if they support both, which would be fine.15:23
smcginnisWe say the name of the CI, not the user id.15:24
smcginnisAnd documented an example showing so.15:24
jungleboyjSo you are saying that run-Lenovo Storage CI should trigger it?15:24
smcginnisAnd if we find any that don't, we should make them fix that. Which is why I added it to the checklist.15:25
jungleboyjHuh.  I didn't know that was the case.15:25
jungleboyjSo many CIs are not doing that.15:25
jungleboyjI don't think.15:25
jungleboyjOk.  I retract my comment then.15:25
smcginnisIt's only been like three years that we declared and documented that that was the requirement.15:26
jungleboyjI bet if we go and test it a lot of CIs aren't going to work.15:26
smcginnisProbably. Which is an issue, but not one big enough that's motivated me to start really pushing.15:26
smcginnisI just want to make sure new drivers are correct.15:26
jungleboyjOk.  Now I know.  :-)15:26
jungleboyjYeah, something to go back and look at.15:26
smcginnisReally wish more CIs would follow that last part of the documentation too about not running unless zuul gives +1.15:27
woojayI totally missed that until you told me about it...  8-)15:29
*** jmlowe has quit IRC15:38
qqmberwhat this log line means? {"log":"tgtd: iscsi_noop_out_rx_start(1694) ffffffff 31 0\n","stream":"stderr","time":"2019-05-17T15:05:33.718926452Z"}15:42
qqmberI'm not iscsi expert...15:43
smcginnisqqmber: Not sure what that would be from. It doesn't happen in non-containerized deployments. Have you tried asking the kolla folks if that's a know issue?15:46
qqmberyes, I'm there15:47
qqmbersmcginnis: thanks15:47
jungleboyjSounds like a timeout might be getting exceeded?15:48
qqmberjungleboyj: the thing is.. I'm not having hanging... weird...15:49
jungleboyjOh, so things are working ok but you are getting tons of logging?15:50
qqmberjungleboyj: exactly15:50
smcginnisHow's performance? Maybe it is needing to continually reset the session?15:51
jungleboyjThat is odd.  I wonder if they just have the logging set too high?15:51
qqmbersmcginnis: I don't know, because this is a new installation... I don't have anything to compare to...15:56
*** sahid has quit IRC15:57
qqmberjungleboyj: I thought the same.. but.. that's an error message15:57
jungleboyjAre you seeing dropped packets on the interface that iSCSI is using?15:57
qqmberthis is only localhost traffic (not all of them, but 90%+++)15:58
qqmberI configure it to have in the sam ehost the volumes and the instances..15:59
qqmberbecause this are very read/write intensive15:59
jungleboyjInteresting.  That would not be a standard configuration.16:02
jungleboyjWonder if you are hitting a bug in tgtd then?16:03
qqmber{"log":"tgtd: iscsi_noop_out_rx_start(1694) ffffffff 31 0\n","stream":"stderr","time":"2019-05-17T15:05:33.718926452Z"}16:06
qqmberbut that line... tells you anything?16:06
qqmberbesides there is a stderr?16:06
jungleboyjqqmber:  Trying to get a clue from the tgtd code but don't really understand.16:15
jungleboyjIt is one of the Opcode values .  It is in iscsi_task_rx_start .16:16
jungleboyjIt gets send to iscsi_noop_out_rx_start16:16
jungleboyjThat is where the printf is coming from.16:16
*** enriquetaso has quit IRC16:17
*** tesseract has quit IRC16:17
*** psachin has quit IRC16:17
*** tesseract has joined #openstack-cinder16:18
jungleboyjI wonder if it is because you are using localhost it is going down the noop path.16:18
jungleboyjNot sure though.16:19
*** whoami-rajat has joined #openstack-cinder16:20
*** jcosmao has left #openstack-cinder16:31
qqmberjungleboyj: thanks! I'll investigate that...16:43
*** davidsha has quit IRC16:48
*** sapd1_x has quit IRC16:48
*** mriedem is now known as mriedem_away16:49
qqmberjungleboyj: I don't think is because I'm using localhost (is not configured to but the local em1 IP!), because I have this exact issue in an older deployment, where not everything is in localhost..16:55
*** tesseract has quit IRC16:58
hemna_what's up16:59
*** tesseract has joined #openstack-cinder16:59
*** carloss has quit IRC17:03
*** enriquetaso has joined #openstack-cinder17:14
*** udesale has quit IRC17:15
*** tesseract has quit IRC17:17
*** e0ne has quit IRC17:17
jungleboyjqqmber:  Hmmm, strange.17:20
jungleboyjhemna_:  qqmber is seeing lots and lots of logging from iSCSI.  You have more experience with iSCSI under the covers than I.  So thought you might have thoughts.17:20
hemna_I've never seen that before17:24
hemna_seems related17:24
jungleboyjhemna_:  :-)  I shared that earlier too.17:24
hemna_maybe a networking issue17:25
hemna_containerd networking...17:25
hemna_it's not an error17:30
hemna_it's just from the dprintf on line 169617:30
jungleboyjYeah, I didn't think so either.17:30
*** enriquetaso has quit IRC17:32
*** enriquetaso has joined #openstack-cinder17:34
hemna_qqmber:  I'd check to see how tgtd is getting launched.  My guess is that there is a -d <number> on the command line enabling debug logging17:34
hemna_which is causing that to get constantly logged17:34
*** jdillaman has quit IRC17:56
smcginnishemna_: If you didn't see it, would love your input on
*** mriedem_away is now known as mriedem18:23
hemna_smcginnis:ok checking18:35
smcginnisCool, thanks.18:35
*** itlinux has joined #openstack-cinder18:38
*** whoami-rajat has quit IRC18:39
*** mvkr has joined #openstack-cinder18:49
hemna_so all drivers have to be apache 2 license?19:00
hemna_not lgpl or other OSS based licenses?19:00
*** swegener has quit IRC19:01
smcginnishemna_: Yeah, libs can be different OSI licenses, but any code we accept in tree absolutely has to be Apache 2.19:02
*** jmlowe has joined #openstack-cinder19:03
hemna_I think drivers should also be required to have a version specified in them19:03
hemna_and the version history in the driver comment block too19:04
smcginnisLine 18?19:04
smcginnisVersion history hasn't been a requirement, but I do like having that too.19:04
hemna_ah yah line 1819:04
hemna_it's not much effort to have that version history19:05
hemna_and it's really helpful to see19:05
smcginnisFeel free to throw that in there.19:05
hemna_ok added19:05
smcginnisI'm mainly trying to capture all the little things we end up downvoting new driver patches on that really the contributors don't have much chance of knowing they even need to do. So whatever we can put in here that helps, hopefully that will result in new drivers needing so many revisions.19:06
hemna_yah it's a great idea19:06
smcginnis*not needing19:06
hemna_and this should get converted into documentation for driver devs?19:06
hemna_not just captured in the etherpad?19:06
smcginnisYeah, after we iterate on this, then I want to add something to our actual docs, not just our wiki.19:07
hemna_should state what minimum CI tests should be run19:09
hemna_and how to document any that are skipped19:09
hemna_and why19:09
hemna_I added a few things19:09
hemna_looks good19:09
smcginnisYeah, that would be good. And including our tempest plugin.19:09
hemna_how about, driver required libraries need to be python 3 compliant, and in global requirements19:10
hemna_any the pypi based driver requirements need to be in setup.cfg (once my patch lands)
hemna_as well as any binary dependencies need to be OSS licensed and placed in bindep19:13
smcginnisAdded a few more things.19:14
smcginnisWow, there really is a lot we expect new driver developers to know.19:15
hemna_yah, it's a lot19:16
hemna_should we add some info about supporting older release, and how to do that?19:17
hemna_backporting a fix the entire chain of releases back to the target release...19:17
hemna_what's allowed to backport, and what's not19:18
smcginnisI was thinking this was targeted just at new drivers, so backporting wouldn't apply yet.19:18
smcginnisBut we could add a note so they know what to expect going forward I guess.19:18
smcginnisOr explicitly state that the new driver can't be backported, since that question has been asked a few times.19:18
smcginnisrosmaita: If you're around, it might be nice to get eyes on from someone new to needing to review new drivers.19:19
hemna_also, when a driver reports it's usage stats it needs to be associated with a pool, even if they only have 1 pool (or no pools)19:20
rosmaitasmcginnis: ack19:21
smcginnisrosmaita: Thanks19:21
smcginnishemna_: I didn't think that was the case if they didn't support multiple pools. Been a long time since I've thought about that area though. :)19:21
hemna_I think they just need to report up via some poolname19:22
hemna_because it's part of the hostname from what I remember19:22
smcginnisHave we had new drivers doing that part wrong?19:22
hemna_no idea...19:22
smcginnisEh, add a note with some details if you think it's something important to watch out for.19:23
hemna_also, we should raise the _init_vendor_properties() usage as well19:23
smcginnisI started by looking at things we've actually downvoted on recent driver patches to try to get the most common issues.19:23
smcginnisOh right.19:23
hemna_I think I might put an info log message in the base's implementation of _init_vendor_properties()19:24
hemna_saying that driver isn't reporting properties....but should19:24
*** enriquetaso has quit IRC19:24
smcginnisThat could actually be very useful.19:24
hemna_ok I'll put that up19:24
hemna_I see why it's a _init_vendor_properties() and not init_vendor_properties call now19:25
hemna_it's called from init_capabilities19:25
smcginnisSo it's "private" to within the driver?19:25
hemna_but I think we should all be implementing that now19:25
smcginnisYeah, that makes sense.19:27
openstackgerritWalter A. Boring IV (hemna) proposed openstack/cinder master: LOG _init_vendor_properites
hemna_it's not required now, but it should be "soon"19:31
hemna_maybe even warning, but info is probably ok for now?19:32
woojaysmcginnis: I'm adding some URL references right now but obviously feel free to edit them for readability.19:32
smcginnisIt is possible a driver doesn't have anything to add. I think there's a few that don't have anything.19:32
smcginniswoojay: Thanks!19:32
smcginnishemna_: Looks like you accidentally picked up some RBD driver exception move things with that patch.19:33
hemna_git stash.....19:33
hemna_always have too much going on19:34
openstackgerritWalter A. Boring IV (hemna) proposed openstack/cinder master: LOG _init_vendor_properites
qqmberhemna_: I jussy come back from lunch and just read your messages.19:45
*** imacdonn has quit IRC19:45
qqmberhemna_: Is not only a lot of messages.. is this one that bothers me the most: {"log":"tgtd: iscsi_noop_out_rx_start(1694) ffffffff 31 0\n","stream":"stderr","time":"2019-05-17T15:05:33.718926452Z"}19:45
qqmberthat appears... a LOOOOTTTTTT19:45
qqmber600GB in three days19:45
hemna_qqmber: not sure why it's dumping that out that much19:46
hemna_it's not an error19:46
hemna_it's just an info message that they left in for whatever reason19:46
hemna_can you track down how tgtd is getting launched?19:46
hemna_maybe dial down the -d <number> if it's on19:46
qqmberhemna_: even though there is a "stderr" in there?19:46
hemna_yah I don't think that's an error19:47
qqmberI can ask kolla-ansible people...19:47
hemna_can you get in the container and ps -ef |grep tgtd ?19:47
qqmberthe question would be "tgtd is being launch with the -d flag?"19:47
*** mchlumsky has quit IRC19:47
qqmberhold on19:47
hemna_yah I think it might be19:47
* hemna_ is just guessing though from looking at the tgtd code19:48
qqmberhemna_: this is in the host (not container): tgtd -d 1 -f --iscsi portal=
qqmbernow I'll go to the container19:48
hemna_nuke the -d option19:49
qqmberwhat is the -d option?19:49
hemna_debug logging19:49
qqmberholly crap19:49
smcginnisDoesn't seem like something you would want on by default.19:49
qqmbernot... at ALL19:49
qqmberif it is an error... it should appears in the logs without the -d option, right?19:50
*** ircuser-1 has quit IRC19:50
*** radeks has quit IRC19:51
*** radeks has joined #openstack-cinder19:51
qqmberok, i'll go for it19:52
*** imacdonn has joined #openstack-cinder19:52
*** ircuser-1 has joined #openstack-cinder19:53
qqmber"command": "tgtd -d 1 -f --iscsi portal=",19:55
qqmberthat's the command... should I take away the "-d 1", right?19:55
qqmbertgtd -f --iscsi portal= <--- that's the answer of "ps -ef |grep tgtd"19:58
qqmberthank you very much19:58
qqmberI'll check at the logs now in order to detect any errors.19:58
*** pcaruana has quit IRC20:02
*** itlinux has quit IRC20:06
*** jsquare has quit IRC20:26
*** jsquare has joined #openstack-cinder20:26
*** radeks has quit IRC20:30
jungleboyjqqmber: Yay!20:39
jungleboyjhemna_:  Thanks for the additional help.20:39
_alastor_jungleboyj: My CI should be reporting now.  I still have to migrate it to a permanent setup, but at least it works :)20:49
jungleboyj_alastor_:  \o/20:52
openstackgerritMerged openstack/cinder master: Handle deprecations
openstackgerritMerged openstack/cinder master: [Unity] Update doc for compressed volume support
*** KeithMnemonic has quit IRC21:08
*** whoami-rajat has joined #openstack-cinder21:30
*** swegener has joined #openstack-cinder21:53
*** henriqueof has quit IRC21:55
*** itlinux has joined #openstack-cinder21:58
*** qqmber has quit IRC22:00
*** rchurch_ has joined #openstack-cinder22:01
*** rchurch has quit IRC22:02
*** dviroel has quit IRC22:19
*** mriedem has quit IRC22:20
*** rosmaita has left #openstack-cinder22:35
*** jcosmao has joined #openstack-cinder22:36
*** jcosmao has left #openstack-cinder22:37
*** dave-mccowan has quit IRC22:50
*** itlinux_ has joined #openstack-cinder22:51
*** itlinux has quit IRC22:52
openstackgerritWalter A. Boring IV (hemna) proposed openstack/cinder master: Move RBD driver exception to driver
*** mvkr has quit IRC23:21
*** whoami-rajat has quit IRC23:40
*** mmethot has quit IRC23:53
*** lseki has quit IRC23:57

Generated by 2.15.3 by Marius Gedminas - find it at!