13:31:16 <esberglu> #startmeeting powervm_ci_meeting
13:31:17 <openstack> Meeting started Thu Dec 15 13:31:16 2016 UTC and is due to finish in 60 minutes.  The chair is esberglu. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:31:18 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
13:31:20 <openstack> The meeting name has been set to 'powervm_ci_meeting'
13:31:24 <thorst_> o/
13:31:36 <adreznec> \o
13:32:36 <openstackgerrit> ymadhavi@in.ibm.com proposed openstack/nova-powervm: Set MTU for the tap device  https://review.openstack.org/411317
13:32:53 <esberglu> #topic openstack ci
13:32:56 <thorst_> so I guess - note that we have a hard stop in 28 minutes for the nova meeting
13:33:04 <thorst_> :-)
13:33:58 <esberglu> Yep. And you want the reviews done by then
13:34:20 <thorst_> if possible.  I'd like the WIP removed ASAP on the first patch, but I think its held back by reviews
13:34:37 <thorst_> I also asked wangqsh to run it through a devstack on neo-4 or 50
13:34:44 <thorst_> which I know is currently earmarked for CI.
13:35:10 <esberglu> Okay
13:35:18 <esberglu> Tell him neo 4 I still need to reinstall 50
13:35:28 <thorst_> k
13:35:39 <adreznec> Did anyone look at why the first patch failed our CI?
13:35:48 <thorst_> nope.
13:35:56 <thorst_> probably because of namespace overlap
13:36:22 <esberglu> The earlier patches were failing because the driver failed to initialize
13:36:30 <esberglu> I haven't looked at the newer ones yet
13:36:38 <adreznec> k
13:37:39 <thorst_> it actually raises an interesting problem...
13:38:01 <thorst_> the nova proper namespace is not being overridden by the out of tree driver
13:38:11 <thorst_> so it is using the in tree driver.
13:38:26 <thorst_> and of course, we're missing 99% of the function there.
13:38:42 <thorst_> so the patch still needs to be WIP...maybe we need to change the namespace or something.
13:39:35 <esberglu> Link to the logs from the CI run
13:39:37 <esberglu> http://184.172.12.213/88/391288/12/check/nova-pvm-dsvm-tempest-full/f8be969
13:39:37 <thorst_> anyway, besides that...ci looks like it is going well?  This other topic needs more discussion.
13:39:53 <esberglu> Yeah CI is great
13:40:24 <esberglu> #topic nova driver
13:41:04 <thorst_> ahh, well if we're talking nova driver...
13:41:12 <thorst_> so any ideas on what we do there?
13:42:53 <adreznec> Yeah, that's a hard one...
13:43:07 <thorst_> we could change the name space in tree to pvm.
13:43:20 <adreznec> I think it leads to a bigger question - how do we decide when a patch should be tested against nova-powervm vs in-tree powervm
13:43:37 <thorst_> it should be tested with both.
13:43:43 <adreznec> Right
13:43:48 <adreznec> Just doubles our CI load
13:43:54 <thorst_> and the series of tests we run against each changes
13:43:59 <adreznec> Mhm
13:44:08 <adreznec> Two lists of tempest tests
13:44:30 <thorst_> yeah...that doesn't worry me too much.  One test should be very quick to start...
13:45:43 <adreznec> Right
13:46:23 <adreznec> I kind of don't like the solution of changing the in-tree namespace to pvm from a technical standpoint... but we can't break all our current consumers either
13:46:31 <adreznec> We should look at what the hyper-v guys are doing
13:46:56 <thorst_> adreznec: agree...and I bet their out of tree has a different name.
13:47:03 <thorst_> we could do something silly and make it power_vm
13:47:53 <thorst_> adreznec: mind doing some research on that and see what you think is best?
13:47:54 <adreznec> Right
13:47:56 <adreznec> Sure
13:48:34 <thorst_> cool.
13:49:15 <thorst_> wangqwsh: did you see the note about testing with devstack?
13:49:25 <wangqwsh> yes, I read it
13:49:47 <thorst_> does it make sense?  esberglu was recommending we use neo-dev-4 for that instead of 50
13:49:55 <wangqwsh> install AIO or AIT?
13:50:08 <wangqwsh> ok
13:50:26 <thorst_> AIT is easier probably
13:50:31 <thorst_> because the novalinks have small disks
13:50:47 <thorst_> and devstack can consume more than all of it if it is also running the controller
13:50:47 <wangqwsh> I do not have VMs
13:50:53 <thorst_> lets get you one on jupiter
13:51:08 <thorst_> we can send you link in private message
13:51:34 <wangqwsh> ok
13:52:57 <thorst_> the idea here is that we just test the function in the WIP patch set
13:54:28 <thorst_> alright.
13:54:32 <thorst_> any other topics?
13:54:35 <thorst_> t-minus 5 min
13:55:05 <wangqwsh> i am fine.
13:55:13 <esberglu> I don't have anything
13:56:05 <adreznec> I have to drop here to transition to the nova meeting
13:56:07 <adreznec> Back online then
13:56:26 <thorst_> k
13:56:30 <thorst_> end meeting time?
13:57:28 <esberglu> #endmeeting