Friday, 2017-03-17

*** svenkat has joined #openstack-powervm00:35
*** k0da has quit IRC00:53
*** svenkat has quit IRC01:45
*** jwcroppe has joined #openstack-powervm04:55
*** k0da has joined #openstack-powervm07:03
*** k0da has quit IRC08:34
*** k0da has joined #openstack-powervm09:46
*** smatzek has joined #openstack-powervm10:52
*** edmondsw has quit IRC11:14
*** edmondsw has joined #openstack-powervm12:09
*** tblakes has joined #openstack-powervm12:09
*** svenkat has joined #openstack-powervm12:49
*** apearson has joined #openstack-powervm12:56
*** mdrabe has joined #openstack-powervm13:08
*** esberglu has joined #openstack-powervm13:13
*** tjakobs has joined #openstack-powervm13:16
*** tjakobs has quit IRC13:22
*** smatzek has quit IRC13:32
*** jpasqualetto has joined #openstack-powervm13:55
*** smatzek has joined #openstack-powervm13:57
*** kriskend_ has joined #openstack-powervm14:14
*** tjakobs has joined #openstack-powervm14:36
*** burgerk has joined #openstack-powervm15:05
openstackgerritTaylor Jakobson proposed openstack/nova-powervm master: Add iSER support
*** jwcroppe has quit IRC15:28
*** jwcroppe has joined #openstack-powervm15:29
*** jwcroppe has quit IRC15:33
efriedtjakobs yt?15:38
efriedWhere (what releases) is this iscsi stuff?15:39
efriedTrying to figure out if proposed pypowervm change 4985 will break backward compatibility.  I think it will, if someone uses a new pypowervm against an older REST.15:39
tjakobsi think the first iscsi stuff was in newton15:40
efriedOh, then definitely.15:40
efriedCommented accordingly - please revise.15:40
*** jwcroppe has joined #openstack-powervm15:43
efriedtjakobs ...and test accordingly (sorry)15:49
*** mdrabe has quit IRC16:26
*** mdrabe has joined #openstack-powervm16:26
*** k0da has quit IRC16:26
*** jpasqualetto has quit IRC16:45
*** jpasqualetto has joined #openstack-powervm16:57
*** jpasqualetto has quit IRC17:50
*** jpasqualetto has joined #openstack-powervm18:09
*** jayasankar_ has joined #openstack-powervm18:11
jayasankar_Hi All, While I deploy a VM on SDN host with SSP, I'm getting the error
*** mdrabe has quit IRC18:22
esberglujayansankar_: This is using the in-tree driver correct?18:22
*** burgerk_ has joined #openstack-powervm18:26
*** burgerk has quit IRC18:29
esberglussp isn't supported yet in-tree unless you have the in-flight changeset for it18:29
*** kriskend_ has quit IRC18:32
*** mdrabe has joined #openstack-powervm18:34
*** kriskend has joined #openstack-powervm18:34
jayasankar_esberglu: Yes I applied that change set18:35
*** burgerk_ has quit IRC19:01
*** apearson has quit IRC19:01
*** apearson has joined #openstack-powervm19:03
*** apearson has quit IRC19:19
efriedesberglu - this looks like the failure we thought we were circumventing with the change sets you dropped earlier.  Can you debug?19:53
efriedjayasankar_ Looking...19:53
efriedjayasankar_ Looks like you don't have your cluster set up.  If you reinstalled, you'll need to recreate (or rediscover) your SSP on your VIOS.19:54
efriedI should put in a better error message for this.19:54
efriedCan you leave a -1 with a comment to that effect on this change set?19:54
jayasankar_efried: Yes, I recreated the cluster as well,  and I was able to see it in pvmctl cluster list on the host19:55
efriedBefore or after this error?19:56
jayasankar_before the error19:56
efriedhmph.  Okay, I'll log into your system and poke around some.19:56
jayasankar_before starting the first deploy19:56
jayasankar_The other observation is that, though the VM's are going into error state those are holding the cores and the latest deploy gave me the below error19:57
jayasankar_neo@neo34:/opt/stack/logs$ openstack server create B18_VM --image Ubuntu_16-04-01_2g --flavor ds4G19:57
jayasankar_ Quota exceeded for cores: Requested 4, but already used 20 of 20 cores (HTTP 403) (Request-ID: req-bc20b77d-685d-4ccc-81c1-b14b3f3f10b4)19:57
efriedeek, you're using a 20-CPU flavor?19:58
jayasankar_No, I was using the flavor ds4G  with 4cpu's19:59
efriedHow many times did you try to deploy?19:59
jayasankar_4 got error and 5th got the above msg at the command line.20:00
jayasankar_4 error VM's are already there in the Instances list.20:00
efriedBut you've cleaned up since then?20:01
jayasankar_These are all recent deploys, just kept them for debugging..20:01
*** apearson has joined #openstack-powervm20:01
efriedAre you still on neo34?20:02
efriedI'm looking at the n-cpu log and only seeing that one error.20:02
jayasankar_Yeah.. that is the pblm..20:03
efriedThe first attempt to start the driver failed with the error you pastebin'd.  The second appears to have succeeded.  But I don't see any deploys after that.  And I don't see any nova instances.  And I don't see any LPARs.20:03
jayasankar_Yeah.. that is the behavior even I observed that is why I didn't delete those Err vm's from the instance list.20:04
efriedThere's no evidence that any VMs were ever created20:05
jayasankar_None went through after the reconfiguration20:06
efriedWhat reconfiguration?  Did you restack at some point?20:06
jayasankar_Yeah.. reinstalled Neo for getting additional file storage20:06
efriedOkay, but the spawn failures were _before_ the reinstall?20:07
jayasankar_I don't think so. bcz after the scratch install only I installed Devstack and created SSP and started fresh deploy20:08
efriedjayasankar_ Okay, I think the problem is more fundamental.20:12
efriedNo host-to-cell mapping found for selected host neo34. Setup is incomplete.20:12
efried^^ from the conductor20:12
efriedI don't understand cells.20:12
openstackgerritTaylor Jakobson proposed openstack/nova-powervm master: Add iSER support
efriedjayasankar_ Okay, this is interesting.  I found this launchpad bug:
openstackLaunchpad bug 1673305 in Packstack "No host-to-cell mapping found for selected host packstack. Setup is incomplete." [Undecided,New]20:15
efriedjayasankar_ After I ran the nova-manage command suggested by amoralej, I was able to get a deploy started.20:19
*** k0da has joined #openstack-powervm20:20
jayasankar_it got failed again..20:21
efriedDamn, it's that same failure you were seeing before.  I can't suss this.20:24
efriedOkay, so here's what I'd like you to do.20:25
efriedUpload the logs somewhere we can get at 'em.  Maybe box or something?  Preferably a public spot so you can legally paste the URL into the change set.20:26
efried-1 the change set, with two comments:20:26
efriedFirst, need a better error message (i.e. better than "list index out of range") when there's no cluster.20:27
efriedSecond, need to nail down this glance error.20:27
efriedAnd actually, also put a link to the above launchpad bug, noting that we had to run that nova-manage command before we could even get started.20:27
efriedI'm going to have to rebase the world anyway here, which I probably won't finish up until next week.  Once that's done, we can try again, and maybe those latter two issues will magically resolve themselves.20:28
efriedI just rebased the bottom change set ( so once you get done with the above, you could re-test that guy.20:30
efriedFor that one, you're just verifying that the driver starts successfully, and spawn/delete go through (though they don't actually create LPARs)20:31
efriedtjakobs I think you've got some more work to do on
efriedtjakobs Did you get it passing locally with the pypowervm patch loaded into your tox env?20:32
efriedI suspect not.20:32
efriedI think the autospec isn't quite right.20:33
tjakobsi didn't put the whole patch on, but i added the 1 arg to discover_iscsi20:33
efriedThere's an easier way; I'll show you in a minute.20:34
*** smatzek has quit IRC20:34
efriedBut having done that, did it pass locally?20:34
tjakobsi added it to py34 and got pep8/py34 passing and py27 failing20:34
efriedhmph.  That surprises me, given the error I'm seeing in the change set.20:34
jayasankar_ Adam: Can you tell how to upload logs ?20:34
efriedAnyway, the easiest way is to cd .tox/pyXX/src/pypowervm and then git review -d <number>.  That'll pull down the change set as it sits on gerrit.  That way you know you're consistent.20:35
efriedadreznec yt?20:35
efriedesberglu Maybe you can help too20:35
efriedBueller?  Bueller?20:36
jayasankar_esberglu: where and how I can keep the logs into ?20:37
tjakobsefried: tried the git review -d from .tox/pyXX/src/pypowervm and got this `ValueError: No JSON object could be decoded`. I'm just going to cp the files in and try20:38
adreznecefried: jayasankar_ Sorry, was moving between locations20:38
adreznecUpload which logs exactly? devstack logs?20:38
efriedtjakobs You'll be better off getting set up properly.20:38
efriedUnless that directory didn't exist.20:39
efriedCause maybe we're using global reqs now.20:39
efriedNoooo, we shouldn't be.20:39
efriedDid that directory exist?20:39
efriedFirst, if you had made manual changes, you would have to reset or stash before you could download.20:39
jayasankar_adreznec: Yes devstack logs20:39
efriedSecond, you'll have to set up your git remote.20:39
efriedadreznec Yeah, we want to upload some logs from a local test to a public place so jayasankar_ can link them from a review comment on openstack gerrit.20:40
efriedA launchpad bug isn't appropriate because the change set is in flight.20:40
efriedDo we use box for that?20:40
esbergluefried: Sorry I was afk for a while, catching up20:40
adreznecjayasankar_: Your best bet is going to be to upload them to since that's what most of upstream uses20:40
adreznecThere are CLI utilities for pushing a file to paste20:41
adreznecLike pastebinit20:41
adreznecSo you could do like "pastebinit -i your.log -b"20:42
jayasankar_ah..ok, I can point the whole whole logs folder ?20:42
adreznecI don't know if it does recursion20:42
adreznecI've never used it that way20:42
jayasankar_efried: you want me to add only n-cpu or any more ?20:42
efriedjayasankar_ n-cpu, n-sch, and n-cond20:43
esbergluefried: Regarding the failure you linked earlier. The nodes just hadn't finished clearing out so that typo fix wasn't in yet20:45
efriedesberglu Okay, cool.  We ready to recheck yet?20:46
esbergluYeah. It might take a little while to go through. A bunch of the CI neos got shut off for memory upgrades so the queue got a little backed up20:47
esbergluThose are coming back online now20:47
esbergluCI should catch back up pretty quickly once they are all back up20:47
*** svenkat has quit IRC20:47
efriedesberglu Okay, posted recheck.  Thanks.20:48
*** tblakes has quit IRC20:48
jayasankar_adreznec: I uploaded using that command but I'm wondering I haven't got notification on the paste link ?20:50
adreznecjayasankar_: That's weird, it should have given you links back in the terminal20:51
*** k0da has quit IRC20:58
*** apearson has quit IRC20:58
*** apearson has joined #openstack-powervm21:08
*** apearson has quit IRC21:21
*** jpasqualetto has quit IRC21:23
openstackgerritTaylor Jakobson proposed openstack/nova-powervm master: Add iSER support
*** kriskend has quit IRC21:24
*** edmondsw has quit IRC21:24
*** edmondsw has joined #openstack-powervm21:29
*** edmondsw has quit IRC21:31
*** edmondsw has joined #openstack-powervm21:34
*** jpasqualetto has joined #openstack-powervm21:35
*** edmondsw has quit IRC21:39
*** jwcroppe has quit IRC21:40
*** esberglu has quit IRC21:41
*** esberglu has joined #openstack-powervm21:41
*** esberglu has quit IRC21:46
*** tjakobs has quit IRC21:48
*** jpasqualetto has quit IRC21:58
*** mdrabe has quit IRC22:15
*** k0da has joined #openstack-powervm22:17
efriedadreznec esberglu, in case we ever feel down about the reliability of our CI, we need look no further than xenserver for an ego boost.22:47
efriedThat POS fails like 50% of the time.22:47
*** jayasankar_ has quit IRC22:50

Generated by 2.14.0 by Marius Gedminas - find it at!