15:05:27 <scottda> #startmeeting CinderTestWG 15:05:28 <openstack> Meeting started Wed Jun 22 15:05:27 2016 UTC and is due to finish in 60 minutes. The chair is scottda. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:05:29 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:05:31 <openstack> The meeting name has been set to 'cindertestwg' 15:05:37 <erlon> Hey 15:05:48 <DuncanT> Do we have an agenda 15:05:50 <DuncanT> ? 15:05:53 <e0ne> #link https://etherpad.openstack.org/p/Cinder-testing 15:06:04 <eantyshev1> mriedem, e0ne: Hello! Can I ask here more attention on https://review.openstack.org/#/c/276465 ? Its prerequisites got merged last week 15:06:04 <scottda> Not much agenda. No DuncanT , not really 15:06:22 <e0ne> DuncanT: we've got only two items now 15:06:35 <akerr> i'm in a meeting but I can glance here a bit 15:06:51 <scottda> With help from mriedem I found a good pattern to test in-flight patches with infra dependencies... 15:07:09 <scottda> https://review.openstack.org/331840 15:07:13 <scottda> #link https://review.openstack.org/331840 15:07:45 <scottda> simple in retrospect, but I was having trouble figuring out how to CI these tests, with config stuff that needed to be set in Devstack.... 15:08:13 <scottda> I'm also wondering about the idea we had last week to make an experimental job for multi-backend testing. 15:08:22 <scottda> (Which I've put up a patch for) 15:08:45 <e0ne> scottda: I like to hr nowave this tests as experimental fo 15:08:53 <scottda> Because the question on a dependant patch was: Does anyone actually use CINDER_ENABLED_BACKENDS in a test? 15:09:48 <scottda> e0ne: Sorry, couldn't parse that? 15:09:55 <e0ne> scottda: oops 15:10:24 <e0ne> scottda: I like the idea to have these tests as experimental for now 15:10:50 <scottda> e0ne: That's OK, I do as well. It was brought up here: 15:11:10 <scottda> #link https://review.openstack.org/#/c/325895 15:11:38 <scottda> #link https://review.openstack.org/#/c/325895 15:12:09 <scottda> "I also would like to see something actually using this option, because clearly nothing has been testing it, otherwise when you blindly rmeoved the deprecated options from tempest nothing was using it in the gate to block that from landing. I'd like to avoid repeating the mistakes of the past" 15:13:03 <scottda> But we can keep as is for now. Just trying to figure out what it will take to get all the patches merged. 15:13:36 <scottda> dulek: Have you had time to work on any multi-node stuff? I can help with some of that if you are busy? 15:14:41 <dulek> scottda: Help would be fine. I wanted to get that this week. 15:15:11 <dulek> scottda: So as you have some more experience than me, I would welcome assistance. :) 15:15:19 <scottda> dulek: OK. I'll see if I can test some patches and I'll ping you with any progress... 15:15:50 <scottda> I don't want to take up too much time here...Anyone have anything specific to discuss? Or help needed? 15:16:09 <akerr> user messages tests merged :) 15:16:22 <scottda> akerr: Yeah, nice! 15:16:29 <akerr> so we have a base class for v3 microversions tests 15:16:37 <e0ne> scottda: I've one question to Cinder's team 15:16:53 <scottda> akerr: Yes, that's a bonus. Thanks for that. 15:17:07 <scottda> e0ne: Ask away... 15:17:16 <e0ne> what do you think about using fake driver in functional tests? 15:17:36 <erlon> scottda: I need someone to help get this merged: https://review.openstack.org/#/c/310061/ 15:18:09 <erlon> scottda: I talked to people on tempest channel, but still need a +2 15:18:09 <xyang1> e0ne: we need to use the FakeGateDriver for CG functionsl tests 15:18:20 <e0ne> xyang1: +1 15:18:23 <scottda> e0ne: Seems like a good idea. That will give us a lot of freedom for things that are not fully implemented, like CG 15:18:45 <scottda> Yeah, what xyang1 said. 15:18:45 <erlon> scottda: this test, should break a few backends that merged a fix in the clone feature, some time ago 15:19:02 <openstackgerrit> Alex Meade proposed openstack/cinder: Add prefix to user message event ids https://review.openstack.org/313140 15:19:04 <e0ne> we've got good tempest coverage for backends tests, so I'm going to switch functional job to use fake driver 15:19:13 <scottda> erlon: YOu mean the test will fail on those backends? 15:19:38 <ameade> ^^ dulek really appreciate the reviews 15:19:48 <erlon> scottda: yes 15:19:50 <xyang1> e0ne: which tests? backup? 15:20:11 <xyang1> e0ne: you said backends 15:20:26 <erlon> scottda: some patches that was meat to fix that where merged, others are waiting on this test to be on tempest to move forward 15:20:37 <erlon> s/meat/meant/ 15:20:42 <e0ne> xyang1: all tests with this job gate-cinder-tox-db-functional 15:20:54 <xyang1> ok 15:21:30 <scottda> e0ne: Seems like a good idea, as long as people can override to test their own backends with functional tests if they want. 15:21:59 <erlon> scottda: 15:21:59 <erlon> https://bugs.launchpad.net/cinder?field.searchtext=honor&search=Search&field.status%3Alist=NEW&field.status%3Alist=INCOMPLETE_WITH_RESPONSE&field.status%3Alist=INCOMPLETE_WITHOUT_RESPONSE&field.status%3Alist=CONFIRMED&field.status%3Alist=TRIAGED&field.status%3Alist=INPROGRESS&field.status%3Alist=FIXCOMMITTED&field.assignee=&field.bug_reporter=&field.omit_dup 15:21:59 <erlon> es=on&field.has_patch=&field.has_no_package= 15:22:23 <erlon> scottda: pretty much a lot 15:23:24 <scottda> erlon: So, is this one of those which-should-merge-first issues? The tempest test needs the fixes in the drivers, but the driver patches are waiting on the tempest test? 15:23:52 <scottda> Seems like the drivers should just Depends-On the tempest test, as the other way around would be very complicated and ugly. 15:25:03 <erlon> scottda: I don't think so, at least no one in tempest mentioned that 15:25:31 <scottda> erlon: Oh, I see, those driver patches have already merged..and the tempest test will fail for them. Is that right? 15:25:58 <erlon> scottda: yes, exactly 15:26:14 <erlon> scottda: but not all were merged 15:26:36 <erlon> scottda: some I was testing in my backend and found that issue 15:26:51 <scottda> erlon: Well, if the merged patches will fail the tempest test, they have a bug, right? Seems those bugs should be filed. 15:27:17 <erlon> scottda: the fix seem very trivial, so cores where approving then before testing from real testing 15:27:25 <scottda> erlon: And then the new patches, to fix that bug, should Depends-On 15:27:45 <scottda> ...should Depends-On your in-flight Tempest test. 15:27:58 <erlon> scottda: mhm ok 15:28:13 <scottda> Does that make sense to everyone? I could be wrong here.... 15:28:24 <erlon> scottda: does the gate honor those inter project dependencies? 15:28:32 <scottda> erlon: Yes. 15:29:00 <scottda> That's *supposed to be* the point of Depends-On. For cross repo dependencies. 15:29:33 <erlon> scottda: good, but I should be good to have that test merged so we can fix the BE that merged the fix 15:29:57 <scottda> erlon: For details see #link http://docs.openstack.org/infra/zuul/gating.html#dependent-pipeline 15:31:41 <erlon> scottda: I'll add that to the patches that are pending 15:31:55 <jungleboyj> doffm: I started looking at that to bring over to Cinder. Needs some work that I haven't gotten done yet. 15:32:06 <scottda> erlon: But the backends with the bad fix merged can still fix the bug, and use your test as a Depends-On to verify. At least, it *should* work that way. We should keep an eye on one of the patches to make sure it runs your test. 15:32:33 <erlon> scottda: mhm 15:32:38 <scottda> erlon: and you saw that -1 from Jenkins for gate-tempest-dsvm-neutron-full-liberty? Not sure what that's about. 15:33:06 <openstackgerrit> Aleksey Ruban proposed openstack/cinder: Nexenta: Add NBD driver for NexentaEdge. https://review.openstack.org/320060 15:33:27 <erlon> scottda: yep, don't know as well. I was working in previous rebases 15:34:03 <scottda> erlon: Let's find one of those BE's which needs the fix, get the patch up with Depends-On, and have a look at the tempest logs to verify that it is working. Then we can go from there, if that seems OK with people. 15:35:10 <scottda> erlon: Can you also put an entry for all of this into the Cinder-testing etherpad? 15:35:11 <erlon> scottda: ok, Ill do it in my BE: https://review.openstack.org/#/c/296083/ 15:36:25 <scottda> erlon: Cool. If there are issues, or push back, we can bring up at next week's Cinder meeting. Thanks for dealing with this. 15:37:06 <erlon> scottda: ok, thanks for helping. 15:37:28 <scottda> Anyone else have something to talk about? 15:38:06 <scottda> Going twice.... 15:38:31 <scottda> Thanks Everyone! 15:38:34 <scottda> #endmeeting