19:03:59 #startmeeting 19:04:00 Meeting started Tue Jun 19 19:03:59 2012 UTC. The chair is jeblair. Information about MeetBot at http://wiki.debian.org/MeetBot. 19:04:01 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 19:04:21 #topic recent work 19:04:57 i have a proposal for a backup system 19:04:59 #link https://review.openstack.org/#/c/8623/ 19:05:24 so we'll start having off-site backups for our important servers 19:05:36 (at least, as soon as i can spin up an hpcloud machine for that purpose) 19:05:50 I read it. Looks good. 19:06:00 is the plan to backup all servers to both RS and HP? 19:06:14 yep 19:06:32 or at least, all the 'important' servers. 19:06:54 (which is most of them) 19:07:21 also, a number of us have changes submitted upstream to gerrit 19:07:37 mine are 'add username to json query output' (merged) 19:07:59 and 'openid sso', not merged yet at last check 19:08:16 and i've been making a few changes to the jenkins job filler 19:08:16 mine upstreams the 'status:reviewable' query and the "important changes" page 19:08:29 jeblair: sounds doubtful mine (WIP) will get accepted 19:09:06 Shrews: it's an important first step though, and having WIP even if it's not merged upstream as-is, is way better than not having anything. 19:09:18 jeblair: true 19:09:39 my current thinking is that assign to owner is a pretty good match for WIP, but.... 19:09:39 jeblair: we should mention for anyone listening that we have 2.4.1 ready to go which has the email fix 19:09:55 what do do with the other half of the assignment functionality is still a question... 19:10:27 it's useless or possibly actively harmful to us as currently designed. if it could be assigned to a group though, that could be useful to us. 19:10:50 Shrews: great, is there a change proposed to productionize that? 19:11:17 jeblair: yes https://review.openstack.org/#/c/8699/ 19:11:25 Oh, i just remembered, there's one thing i want to test out on 2.4.1 before we put it in... 19:11:34 (it's not something that would affect gerrit itself) 19:11:46 k 19:12:00 spearce did respond to the comments about drafts, WIP, and private changes. https://groups.google.com/forum/?fromgroups#!searchin/repo-discuss/drafts/repo-discuss/YuLup-ELrP8/GOHurEUcA8IJ 19:12:11 but rather, I changed zuul to set verified=0 when it starts testing a change, and i also had it leave a message 19:12:34 i bet the message will get emailed; i need to check if setting verified=0 without leaving a message creates an email 19:13:02 basically, i'd like zuul to be as quiet as possible. but if it sends an email regardless, may as well leave the message there. 19:13:32 #link https://groups.google.com/forum/?fromgroups#!searchin/repo-discuss/drafts/repo-discuss/YuLup-ELrP8/GOHurEUcA8IJ 19:13:38 doesn't sound like upstream has much interest in "fixing" drafts 19:14:00 the solution of having a second repository for private changes seems like a lot of extra work 19:15:25 U( 19:15:26 :( 19:15:53 anybody else been working on something interesting? 19:16:18 in tree docs are now uploaded to docs.openstack.org/developer/$project 19:16:37 #link http://docs.openstack.org/developer/nova 19:17:13 cool! has there been discussion about moving ci docs there as well? 19:18:00 I think annegentle didn't want the CI docs there, but if we setup ci.openstack.org to be an ftp server we could use the same jenkins jobs for ci.openstack.org 19:18:48 jeblair: i have pygerrit working in dev mode, but it's not useful to us unless we upload it to Google Apps. I sent mtaylor a list of changes that I think it needs if we want to explore using it instead of java gerrit. 19:19:27 Shrews: is pygerrit v1.0 of gerrit? or some fork? 19:19:31 short of it: it needs a lot of work 19:19:31 Shrews: neat. i'm guessing that's a substantial list of changes? 19:19:46 clarkb: yes, v1.0 of gerrit 19:20:09 Shrews: ok, well it's still good to know what our options are(n't). :) 19:20:21 jeblair: yes. we first need to divorce it from GAE as a first step 19:20:58 i mean, it's doable though 19:21:25 #topic etherpad / docs changes 19:21:29 if anyone wants to play with the code, my modified version is: https://github.com/Shrews/PyGerrit 19:22:02 clarkb: want to summarize etherpad-lite, and where that's going? 19:22:48 sure. Etherpad Lite is light weight implementation of etherpad basically. Lite on resource use, not on features. 19:23:23 The idea is to migrate the existing etherpad install to etherpad lite. I have a puppet module written to do that and the change to apply it to the future etherpad.openstack.org host is in gerrit 19:23:40 https://review.openstack.org/#/c/8706/ 19:23:54 before that gets approved I need to copy a few files over to the new host (ssl certs) 19:24:27 the ssl certs are currently self signed by me, but I imagine at some point we may want trusted certs? 19:25:01 yeah, we'll get real certs 19:25:24 the migration of data from etherpad to etherpad-lite is blocking on access to the old server, but the process looks straightforward and is documented at https://github.com/Pita/etherpad-lite/wiki/How-to-migrate-the-database-from-Etherpad-to-Etherpad-Lite 19:25:43 so once we have access to the data I don't expect any major issues migrating and testing 19:27:09 I know annegentle is hoping to use etherpad-lite as a place where folks can edit code then push to gerrit for review. etherpad-lite is pluggable through node.js and has an API (with python bindings) so in theory this is doable 19:27:44 cool. my hope is that if that project gets off the ground, we'll have an etherpad server ready for it. 19:28:47 #topic testing 19:29:27 I'm going to continue working on the jenkins job filler with the goal of being able to succinctly describe the devstack jobs 19:29:49 I'd like to check in with the tempest folks and see if they think it'll be ready for gating soon 19:30:13 and we should start looking into how we can test client library backwards compatibility 19:31:27 #topic open discussion 19:31:39 anything else anyone wants to talk about? 19:32:06 I did eventually manage to update the github pull request closing script 19:33:15 yes, thank you. that source of cronspam is dealt with. now onto the next! :) 19:33:23 I do 19:33:42 last week there was some discussion about gating swift on devstack (or something along those lines) 19:34:10 adding swift to the devstack gate (may be a better way to say that) 19:34:46 I'm trying to catch up from being out all last week, and just want to know the status on that conversation 19:34:56 it seems a number of people would like to see that added 19:35:09 i have not done any work on that, nor proposed a change 19:35:38 to gate swift on this or to add swift to devstack defaults? 19:35:48 (I'm actually not sure of the differences) 19:35:57 so the process would be: 19:36:21 1) make sure devstack can configure swift correctly (whether enabled by default or not is orthogonal) 19:36:48 2) propose a change to the devstack-gate scripts to enable swift for the gating tests 19:37:14 what is currently gated by devstack? 19:37:39 (that change is actually self-limiting; in that if it doesn't work, it wont pass it's own gate test.) 19:38:17 (1 sec) 19:39:11 https://review.openstack.org/#/c/8443/ is a move to add quantum 19:39:33 openstack-dev/devstack openstack/nova openstack/glance openstack/keystone openstack/python-novaclient openstack/python-keystoneclient openstack/python-quantumclient openstack/python-glanceclient openstack/horizon openstack/tempest" 19:39:48 notmyname: those are the projects currently gated on devstack ^ 19:40:15 ok 19:40:21 clarkb: yep, and it still needs some devstack configuration work, so it won't go in yet 19:42:09 er, except tempest; it's running silently right now, not part of the gate. 19:43:07 notmyname: do you happen to know if swift-in-devstack works currently? 19:43:20 jeblair: no idea. I've never used devstack 19:43:54 I certainly support getting it to work, but devstack isn't something I've looked at yet 19:45:05 ok. it worked at some point, i believe, so if it doesn't currently, i doubt it will be too hard to get it up to date. 19:45:24 great :-) 19:46:00 cool; other items? 19:47:19 oh, er, mtaylor has started a significant mailing list thread about python client library versioning. all the details have not yet been resolved. 19:47:33 not sure we're equipped to talk about it in depth here, but i thought i'd mention it for the record. 19:48:14 thanks everyone! 19:48:16 #endmeeting