14:01:32 <bauwser> #startmeeting nova_scheduler
14:01:33 <openstack> Meeting started Mon Aug 17 14:01:32 2015 UTC and is due to finish in 60 minutes.  The chair is bauwser. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:01:34 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:01:38 <openstack> The meeting name has been set to 'nova_scheduler'
14:01:52 <bauwser> sorry folks, a bit late
14:02:03 <bauwser> anyone here to talk about the scheduler?
14:02:08 <PaulMurray> o/
14:02:10 <edleafe> o/
14:03:03 <bauwser> #chair bauzas
14:03:03 <openstack> Warning: Nick not in channel: bauzas
14:03:05 <openstack> Current chairs: bauwser bauzas
14:03:11 <lxsli> o/
14:03:28 <jaypipes> hola
14:03:50 <bauzas> ohayo gaimazu folks
14:04:11 <bauzas> so, n0ano was unable to be here
14:04:14 <bauzas> let's statr
14:04:24 <lxsli> btw I have to go in 30mins
14:04:29 <bauzas> #topic status updates
14:04:39 <edleafe> lxsli: should be a quick meeting
14:04:54 <bauzas> #link https://etherpad.openstack.org/p/liberty-nova-priorities-tracking
14:04:57 <bauzas> anything to report ?
14:05:12 <lxsli> alaski's persist-request-spec is up, which is cool
14:05:26 <lxsli> we're working on longest patch chain ever
14:05:29 <bauzas> the ReqSpec object chain is updated, I need to verify the CI
14:05:40 <lxsli> I have an initial patch up but it needs a lot of work yet
14:05:50 <bauzas> lxsli: yeah, we need to be very careful with git
14:06:04 <edleafe> I pushed a new version of the NoValidHosts logging change: https://review.openstack.org/#/c/192184/
14:06:22 <lxsli> Enqueued
14:06:25 <jaypipes> I could use everyone's reviews on my PCI cleanup patch series (it's a precursor for resource objects): https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:pci-cleanup,n,z
14:06:32 <edleafe> It was a change suggested by johnthetubaguy - would like some feedback on the approach
14:06:39 <bauzas> edleafe: sure
14:06:51 <bauzas> edleafe: CC'ing myself
14:07:17 <edleafe> jaypipes: added to my TODO list
14:07:34 <jaypipes> cheers
14:07:47 <bauzas> jaypipes: as well
14:07:49 <jaypipes> edleafe: I will review the NoValidHosts shortly.
14:07:56 <edleafe> kewl
14:08:09 <lxsli> edleafe: is there other work to log (in the scheduler log) the errors encountered when a retry happens?
14:08:27 <edleafe> lxsli: not that I'm aware of
14:08:38 <edleafe> lxsli: I thought all retry failures were already logged
14:08:56 <lxsli> edleafe: you get the classic "NoValidHost" because all hosts become ignored
14:08:59 <bauzas> lxsli: well, there is a thing
14:09:10 <lxsli> edleafe: then you have to check the compute logs of each machine it tried
14:09:13 <edleafe> lxsli: not any more
14:09:14 <bauzas> lxsli: mostly unimplemented yet
14:09:30 <edleafe> lxsli: you get MaxRetriesExceeded exception
14:09:30 <bauzas> lxsli: that says, NoValidHosts for retries is silly
14:09:41 <bauzas> edleafe: oh that is merged?
14:09:42 <johnthetubaguy> I thought we were adding a new exception for the retry case?
14:09:51 <lxsli> edleafe: OK, that's better but getting the exceptions from compute hosts would be better
14:09:52 <johnthetubaguy> thats the one
14:09:54 <lxsli> even better
14:10:15 <bauzas> meh
14:10:26 <lxsli> anyway straying out of topic
14:10:30 <bauzas> 2 weeks in the weeds left me with a big backlog
14:10:37 <edleafe> lxsli: wait - are you saying that when a build fails, there is no logging? Or that you want additional reporting?
14:11:04 <lxsli> I'm not sure the instance error table is working as it should
14:11:21 <bauzas> lxsli: the instance actions table ?
14:11:21 <lxsli> also it'd be nice to get the compute error which caused the retry copied into the scheduler log
14:11:36 <lxsli> that could be it?
14:11:41 <edleafe> lxsli: here's the new exception: https://review.openstack.org/194204
14:12:03 <bauzas> edleafe: is that implemented ?
14:12:20 <edleafe> bauzas: yes - merged a few weeks ago
14:12:26 <bauzas> edleafe: okay
14:12:37 <bauzas> good
14:12:39 <bauzas> moving on ?
14:12:56 <bauzas> let's clarify that specific compute error problem off-topic
14:13:20 <bauzas> on the bp side, I'm working on allocation-ratios-to-rt
14:13:34 <bauzas> nothing yet uploaded
14:13:42 <bauzas> so we can move on
14:13:51 <lxsli> oh that's merged
14:13:55 <lxsli> oh well ^^
14:14:16 <jaypipes> bauzas: you are already pretty overloaded, IMHO. Any chance someone from Mirantis' Nova team can assist you on the allocation-ratios bp?
14:14:44 <jaypipes> bauzas: Timofei I know has some bandwidth he can help you with.
14:15:30 <bauzas> jaypipes: I've done a few things already, we could maybe take that decision at the end of the week if you wish :)
14:15:49 <bauzas> jaypipes: I really want to upload something, so timofei could look at
14:17:28 <bauzas> ok, moving on ?
14:17:32 <bauzas> #topic opens
14:17:36 <bauzas> #undo
14:17:36 <openstack> Removing item from minutes: <ircmeeting.items.Topic object at 0xa20cc90>
14:17:41 <bauzas> #topic open discussion
14:18:22 <bauzas> crickets ?
14:18:34 <jaypipes> bauzas: sure, that's fine :) just let us know
14:18:45 <bauzas> jaypipes: I appreciate the offer :)
14:19:00 * jaypipes suggests we use the remainder of the meeting time to do reviews.
14:19:12 <bauzas> coolness
14:19:12 * edleafe seconds the motion
14:19:24 * jaypipes seconds that emotion
14:19:26 <bauzas> 2 weeks before FF ?
14:19:35 <bauzas> just for verifying dates
14:19:44 * edleafe wants to hear jaypipes' falsetto
14:19:57 <bauzas> #link https://wiki.openstack.org/wiki/Nova/Liberty_Release_Schedule#Spec_and_Blueprint_Approval_Freeze
14:20:21 <bauzas> okay, let's call the wrap then
14:20:35 <bauzas> #endmeeting