14:00:01 <dulek> #startmeeting Kuryr
14:00:02 <openstack> Meeting started Mon Apr 16 14:00:01 2018 UTC and is due to finish in 60 minutes.  The chair is dulek. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:04 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:00:06 <openstack> The meeting name has been set to 'kuryr'
14:00:19 <dulek> #chair irenab dmellado
14:00:20 <openstack> Current chairs: dmellado dulek irenab
14:00:25 <irenab> hi
14:00:30 <danil> hello
14:01:06 <yboaron> Hi
14:01:06 <ltomasbo> o/
14:01:25 <dulek> Welcome, everyone! Let's wait a bit for apuimedo as he's not in the channel.
14:01:25 <garyloug> o/
14:03:23 <irenab> dulek, can you please update regarding Octavia issues?
14:03:29 <apuimedo> \o/
14:03:40 <apuimedo> irenab: which of them?
14:03:44 <dulek> :D
14:03:54 <irenab> choose the order :-)
14:03:55 <dulek> So last week was about gate failures.
14:04:28 <dulek> There were 2 of them that we've fixed/workaround. Plus one breaking containerized gate and fixed in psutil now.
14:04:50 <dulek> Now we're suffering with Amphorae image not being built in OpenStack.
14:05:05 <dulek> Coincidentally it can be fixed with dmellado patch that uses tarballed Octavia.
14:05:12 <dulek> So we need it to unblock the gate.
14:05:22 <dulek> Let me take a look on that patch status.
14:05:42 <irenab> seems it got merged
14:05:52 <dulek> Yup, great, gate should work again!
14:06:13 <apuimedo> dulek: should we recheck the caps patch?
14:06:20 <dulek> I don't really know why those gate issues always come in batches. :(
14:06:45 <dulek> apuimedo: Which one? Sorry, there were multiple of them lately.
14:06:51 <irenab> dulek, fixing one bug ususally unblocks few more
14:07:00 <apuimedo> https://review.openstack.org/#/c/560885/
14:07:22 <dulek> irenab: But why unrelated issues came out in a single day? :D
14:07:42 <dulek> apuimedo: You're right, I've rechecked it.
14:07:58 <dulek> apuimedo: Though that patch doesn't block anything - the container build issue was fixed in psutil.
14:08:01 <apuimedo> dulek: cause friday
14:08:07 <dulek> See - https://github.com/giampaolo/psutil/issues/1268
14:08:21 <apuimedo> dulek: I know it does not. It's just to use it to check the rest is okay
14:08:30 <dulek> :)
14:08:39 <dulek> It's going through gate again now.
14:09:44 <apuimedo> ok
14:09:55 <dulek> Okay, so we can expect gate is fixed now, if anyone spots another issue - ping me, I've became quite good in debugging this lately. ;)
14:10:11 <dulek> #topic kuryr-kubernetes
14:10:22 <dulek> Anyone wants to do an update on kuryr-kubernetes?
14:10:36 <danil> hi
14:10:42 <danil> currently I refactor CNI part according to Michal's remarks (https://review.openstack.org/#/c/471012/). As he told earlier that callback should called from CNIHandlerBase.on_present directly and maybe_callback should just check the possibility of callback
14:11:16 <danil> also I've made some changes in controller part, but haven't pushed yet
14:11:59 <dulek> Uh, we all owe danil reviews, aren't we?
14:12:01 <danil> also as I can see (https://review.openstack.org/#/c/512281/) falls with some problem in opts.py ( AttributeError: 'module' object has no attribute 'subnet_caching_opts'). It is strange because the following pathces don't fall with it
14:13:21 <danil> so I can focus my work now on refactoring you remarks
14:13:41 <dulek> danil: Nothing strikes me as wrong in that patch. Is the error in the gate?
14:13:46 <apuimedo> dulek: yup
14:14:04 <danil> just recheck?
14:15:03 <dulek> danil: Can you link to that failure? Is it the one in Octavia gate?
14:16:25 <danil> error at kuryr-kubernetes-tempest-daemon-octavia. http://logs.openstack.org/81/512281/11/check/kuryr-kubernetes-tempest-daemon-octavia/d70f096/job-output.txt.gz
14:16:40 <danil> "AttributeError: 'module' object has no attribute 'subnet_caching_opts'"
14:17:03 <dulek> Ah, nice one.
14:17:55 <dulek> Very strange. I'll look at that after the meeting. You can try rechecking.
14:18:17 <danil> ok, sure. Thanks
14:18:32 <dulek> Anyone else have something to share?
14:18:39 <apuimedo> nope
14:18:47 <irenab> will appreciate reviews on Network Policy patches as well
14:18:57 <apuimedo> trying to fix all the gate bugs only
14:19:02 <apuimedo> (on my side)
14:19:18 <apuimedo> oh yes
14:19:39 <apuimedo> dulek: irenab: I wonder whether we should backport all the zuul improvements that dmellado did to stable/queens
14:20:36 <dulek> Hm, are the fact they're not backported blocking you on something?
14:20:40 <irenab> apuimedo, lets ask dmellado later
14:21:20 <irenab> dulek, it should be mostly to impove gates, so stable release can benefit from it
14:21:27 <apuimedo> dulek: well, they make backports that change things like which octavia lbaas stuff gets installed differ
14:21:29 <apuimedo> so I wonder
14:21:41 <apuimedo> not difficult to modify though
14:21:48 <apuimedo> I just wonder if it's worth it to have it differ
14:22:28 <dulek> I guess we can judge them on one-by-one basis.
14:22:56 <dulek> E.g. I do need this iptables addition before backporting `docker exec` patch.
14:24:49 <apuimedo> is it not backported yet?
14:24:58 <dulek> It is.
14:25:40 <apuimedo> ok
14:26:14 <dulek> Anything else anyone? :)
14:26:28 <garyloug> I can, quickly!
14:26:28 <apuimedo> not here
14:26:32 <apuimedo> sure
14:26:42 <garyloug> on nested DPDK, I'm following up on feedback from Irena and Yossi around the dpdk-devbind and driver stuff
14:26:44 <apuimedo> garyloug: sorry I didn't get to the review with all the gate breakage
14:27:00 <garyloug> Yossi had some really nice questions on the context of the patch which I've replied to earlier
14:27:09 <garyloug> apuimedo, np :)
14:27:10 <yboaron> garyloug, 10x!
14:27:27 <garyloug> One question I had, if setting the location of dpdk-devbind in kuryr.conf, I guess it makes sense to put this under the binding section, right?
14:27:40 <garyloug> But I think this would require changes to Kuryr-Lib, right?
14:27:49 <garyloug> Am I best to keep everything in one patch for now?
14:28:41 <garyloug> maybe we can follow up in openstack-kuryr if people haven't reviewed the patch, just something I was wondering
14:30:05 <dulek> garyloug: Yeah, we probably should. Sorry for not looking at the patch, but -1 from PEP8 test makes me a bit reluctant. :P
14:30:29 <apuimedo> garyloug: yeah. We should find a way for it to not need changes to kurryr-lib
14:30:29 <irenab> keeping discussion on the patch will be great
14:30:30 <garyloug> dulek, sure ok, I'll fix that :P
14:30:44 <dulek> :)
14:30:57 <garyloug> irenab, good idea, I'll ask this on the patch, thx
14:32:41 <dulek> Should we move on from kuryr-kubernetes?
14:33:09 * irenab need to leave now, will follow up later
14:33:58 <dulek> #topic Other Kuryr components
14:34:12 <dulek> I guess we can aggregate rest of our repos like that. :)
14:34:44 <apuimedo> yes
14:34:48 <dulek> I've seen that some of stable/queens housekeeping patches were merged into openstack/kuryr.
14:35:10 <dulek> Which is good. Is anyone aware of other ones that need attention?
14:35:45 <dulek> Hm, looks like this one needs +W: https://review.openstack.org/#/c/537499/
14:36:02 <dulek> And this: https://review.openstack.org/#/c/537501/
14:36:49 <dulek> And for kuryr-libnetwork there seems to be nothing related to housekeeping in queue.
14:36:50 <apuimedo> done
14:37:00 <apuimedo> right
14:37:11 <dulek> :)
14:37:47 <dulek> Okay, anything else here?
14:38:01 <apuimedo> not here
14:38:44 <dulek> #topic Open discussion
14:38:51 <dulek> apuimedo: And here? ;)
14:39:02 <apuimedo> neither
14:39:06 <ltomasbo> lol
14:40:02 <dulek> apuimedo: Anywhere else? :D
14:42:05 <dulek> I guess not. Okay, thanks for attending folks!
14:42:09 <dulek> #endmeeting