[00:00:23] ah [00:00:29] my policy.json is bad [00:00:39] PolicyNotAuthorized: Policy doesn't allow network:allocate_for_instance to be performed. [00:00:49] heh [00:00:58] Assume you where trying to limit public ip assignments? [00:01:13] Chrome really sucks when it tries to re-open 20 tabs :( [00:01:50] yeah. seems allocate for instance and deallocate for instance aren't what I thought [00:02:02] I was trying to limit floating ip management to netadmin [00:03:18] figuring out the permissions in this json aren't easy [00:03:35] Gah labsconsole is so slow pulling the instance list. [00:03:49] yes [00:03:51] because of nova [00:04:04] Updating chrome totally doesn't fix that css issue, happens on some boxes on the current console too though. Guess it's a theme/browser issue/ [00:04:05] it's pulling every instance for every project every time [00:04:35] it also doesn't properly filter the deleted instances either [00:04:48] Awesome [00:05:01] Was that the whole cloudinit lag issue a while back thing? [00:08:03] yes. we did a patch for that [00:08:10] it only fixes that one query, though [00:15:53] now that the token cache invalidation issue is out of the way, I should be able to tackle the rest of the issues by removing myself from project groups. heh [00:16:34] :D [00:16:51] Lets just hope you don't end up rolling back because that would be a huge pita. [00:17:00] well, this isn't the upgrade yet [00:17:04] it's next week [00:17:07] I pushed it back [00:17:15] Ah [00:17:16] Hopefully as openstack matures migrations betwean versions aren't going to be hellish. [00:17:44] well, this one is really difficult because I need to change os versions, switch apis and add keystone support [00:17:52] going to folsom from essex should be relatively easy [00:18:04] You going to 12 and using the packages from ubuntu? [00:18:08] yeah [00:18:12] Cool [00:18:49] Wasn't there an upgrade a few versions ago that had a backward incompatibility change in one of the itnernal apis that required pretty much stopping all compute nodes? Or is that another piece of software I'm thinking of. [00:19:04] I think that's another piece of software [00:19:12] gluster needed that for storage [00:22:14] I know gluster's a little fussy, don't use it much due to the write overhead on small files though. Just hoping Cpeh starts packaging rpms and supporting RHEL somewhat more now. It will be interesting if Redhat push gluster to succeed now they are moving to support openstack. [00:22:41] there we go [00:22:43] fixed that issue [00:23:13] they say not to use gluster for instance storage [00:23:22] for the exact reasons that I mentioned [00:23:41] I have a feeling I influenced that [00:23:57] after the constant complaints about outages we were having. heh [00:24:20] lol [00:24:42] I think Ceph are pushing to become standard now with Dreamhost backing development of openstack somewhat. [00:24:50] yeah [00:25:02] ceph is separating into another organization [00:25:13] they were a conference sponsor last summit, though [00:25:37] I read about some enterprise side of ceph a while back which sounded interesting. I remember when it was literally just a subdomain on dreamhost's dev site with 2 pages heh [00:25:55] Might actually build an awesome community and do what gluster promised. [00:26:07] well, ceph's design is *way* more interesting than gluster's [00:26:22] Yeah, splitting the services makes sense [00:26:45] well, also that it treats everything like an object [00:27:04] The only thing I dislike is having the 'api' server betwean storage and openstack, it makes sense for swift style usage but not really vms as you want low latency as possible imo. But that probably works better when thought about longer. [00:27:13] so that you can have object, block or posix access [00:27:55] Ah yeah I know what you mean [00:28:10] I'm actually willing to take a small performance hit for that [00:28:42] I think we'll be sticking with local storage for a while, though [00:28:45] From my understanding that was a core decission so they could support serving to usermode like gluster/nfs and serving up object/vm storage. [00:28:52] Though I could be totally wrong [00:29:06] I think it's a basic design decision of the filesystem [00:30:41] hm. there's some spots in the interface I need to fix instance ids [00:30:58] since the switch to the nova api the ids are uuids and not ec2 style ids [00:31:05] so I need to translate all over the place [00:31:20] Probably easier than moving hundreds of pages in the wiki :) [00:31:33] yeah [00:31:37] and dns issues and such too [00:31:42] Also I don't believe there hasn't been any vandalism for 3hours... wth is the bot doing -.- [00:33:15] o.O [00:33:18] I don't suppose there's any chance of building a general labs graphite setup where any projects can just pipe data to it? [00:33:21] PHP Warning: mysql_pconnect(): Can't connect to MySQL server on 'bots-sql2' (4) in /mnt/share/cluebot/cluebotng/bot/mysql_functions.php on line 4 [00:33:30] two instances have the same floating IP associated? [00:33:32] how's that possible? [00:33:39] magic? [00:33:48] wtf is bots-sql2 unresponsive [00:33:50] * Damianz finds a hammer [00:33:51] Damianz: I don't see why not [00:34:00] Damianz: make a proposal for doing so [00:34:06] think about how to make it per-project as well [00:34:26] ah [00:34:28] there we go [00:34:30] It would be interesting to build a general monitoring/graphing setup - gangalia is awesome for hosts but for stuff like bots there's so much data the community could use. [00:34:41] it took a while for it to list the other instance as disassociated [00:35:27] !log bots bots-sql2 oom, rebooting [00:35:29] Logged the message, Master [00:35:32] Need to bring nagios bot back [00:36:22] I need to add error handling so if the bot can't reconnect it outputs debugging info it seems. [00:36:38] * Damianz is highly tempted to rewrite this in python using threadpools and magic but meh, not worth the effort. [00:36:40] I need to make ircecho reconnect properly [00:38:56] [[List of Dragon Ball Z episodes (season 1)]] by "76.119.104.85" ( http://en.wikipedia.org/w/index.php?diff=508374803&oldid=498605046 ) 0.857058 (Reverted) (Default revert) (4.7305338382721 s) < and vandal slapping re-commences [00:39:10] heh [00:39:31] it's going to confuse the hell out of users that security groups are per-region [00:40:23] I suppose in theory regions can be totally isolated, we're just lucky to have connectivity betwean both which is awesome for internal stuff. [00:41:00] yeah [00:41:25] well, it's also more flexible to be able to handle both regions independently [00:41:49] but in most cases people are going to want the same rules in both places [00:42:08] mhm [00:42:10] thankfully for some things (like sudo rules) it's not per-region [00:42:16] same with puppet groups [00:42:18] Now that would be annoying [00:42:40] dns is per-region [00:42:44] kind of [00:42:45] heh [00:43:08] As we use ldap and powerdns for dns does it actually matter how openstack stores it? [00:43:15] no [00:43:28] but we're going to let openstack manage that soon [00:43:35] since we added dns support to openstack [00:44:41] Totally makes sense for openstack to handle dns [00:44:53] yep [00:50:47] Hmm 2am, I probably should go to bed considering I have an interview at 9.30, listening to random music and reading hacker news is far too entertaining. [00:51:55] heh [00:52:03] Damianz: well, go to bed! [00:52:04] ttyl [00:52:09] good luck on your interview [00:52:37] Cheers :) [01:32:10] Change on 12mediawiki a page Developer access was modified, changed by Sumanah link https://www.mediawiki.org/w/index.php?diff=574146 edit summary: /* User: */ [01:47:47] Change on 12mediawiki a page Developer access was modified, changed by Russetburbank link https://www.mediawiki.org/w/index.php?diff=574147 edit summary: [04:05:51] 08/21/2012 - 04:05:51 - Updating keys for junaidpv at /export/keys/junaidpv [10:58:56] Hi there. Could someone help me with gerrit, please? I am a beginner as to this. I have an extension Extension:URNames and a repository was created for this. Unfortunately I don't know what to do now. Tutorials are not as much specific. I don't know paths, branches, where the repo is located ... I would be greatful for some help with getting my code to repository. Thanks a lot. [11:02:47] https://labsconsole.wikimedia.org/wiki/Git [11:02:55] I don't really see an extension repo for that though. [11:03:13] You'll need to add your ssh key in gerrit if not done already also [11:03:32] To see all repos hit https://gerrit.wikimedia.org/r/#/admin/projects/ (need to be logged in) [11:11:19] You're right, the repo is not listed here https://gerrit.wikimedia.org/r/#/admin/projects/. But it should be somewhere by the page http://www.mediawiki.org/wiki/Git/New_repositories/Requests (ctrl-f urnames). Maybe I should ask directly ^demon, who made it. Thanks. [11:13:12] I'd ask chad when he's around as I don't see it and don't have access to make it. Probably be a few hours though. [11:13:21] Once it's there you can just git clone as per the labs docs [11:14:09] Weird, the group is there, but there's no repo with the group assigned to it [11:18:06] * Damianz pokes paravoid, he's in europe IIRC not sure if he has super gerrit access though [11:22:34] Yes, I see the group. I think I can proceed with the help of tutorials if I can find the repo details. [11:23:10] hey [11:23:23] I probably do have sufficient access, but I'm not very experienced in that sort of thing [11:24:04] Heh, it is somewhat a ^demon area. [11:24:36] jossmart: You can create a git repo and commit to it then add a remote later, just can't push up/review stuff until the repo exists in gerrit ;) [11:26:09] We probably should update the docs as "Push access is currently limited to staff developers and operations engineers" isn't 100% accurate, it's just not as open as it might be. [11:38:02] If everything's ok I should find the repo information here https://gerrit.wikimedia.org, right? So I can try it later. [11:39:05] Yeah, not sure if it shows on your dashboard but if you hit admin you can see the projects. I'll poke ^demon when he appears if I remember [11:46:59] Thanks for your help [15:11:14] Change on 12mediawiki a page Developer access was modified, changed by Sumanah link https://www.mediawiki.org/w/index.php?diff=574295 edit summary: /* Notes */ [16:47:17] * Damianz pokes ^demon [16:47:28] <^demon> What's up? [16:48:16] Some guy was asking about his urnames repo, looks like you made the group but not the actual repo. Guy was jossmart, listed as done on http://www.mediawiki.org/wiki/Git/New_repositories/Requests but no repo under admin. [16:48:42] <^demon> Um, should be [16:48:44] <^demon> Lemme double check [16:50:27] <^demon> Oh what do you know, it's there now ;-) [16:50:46] It's like magic [17:11:16] apmon: is there a JIRA ticket for the PostGIS update? I cannot find any [17:11:35] yay, wrong channel! [17:13:28] Change on 12mediawiki a page Developer access was modified, changed by Iggyvolz link https://www.mediawiki.org/w/index.php?diff=574340 edit summary: [17:14:02] PostGIS is cool [17:14:31] Change on 12mediawiki a page Developer access was modified, changed by Iggyvolz link https://www.mediawiki.org/w/index.php?diff=574341 edit summary: /* User:Iggyvolz */ [18:07:31] andrewbogott: can you follow up on https://review.openstack.org/#/c/10365/ and https://review.openstack.org/#/c/10370/ ? [18:07:54] getting changes merged into stable is a pain in the ass [18:08:13] Ryan_Lane: Yep, looking now. [18:08:17] thanks [18:08:21] they want tests written [18:08:35] unfortunately, the ldap backend is almost totally lacking in tests [18:09:06] it's bullshit that I need to report the bug, fix it, then write the tests too [18:09:45] You're just awesome like that [18:10:02] Surly having someone report it and write the tests and another write the code would make more sense [18:10:36] -_- [18:10:47] ideally I report the bug and someone fixes it [18:10:50] and writes the tests [18:11:27] I know the real world doesn't work that way, especially in a project like openstack where people don't give too much of a shti [18:11:28] *shit [18:11:47] self-motivation is the name of the game there [18:13:24] Ryan_Lane: Do you agree that https://review.openstack.org/#/c/10544/ accomplishes what you're after in 10370? [18:14:38] oh [18:14:45] yes [18:14:50] that's in master, though [18:14:58] the other change is in stable [18:15:06] but yes, that fixes it in master [18:15:37] Change on 12mediawiki a page Developer access was modified, changed by Jeremyb link https://www.mediawiki.org/w/index.php?diff=574370 edit summary: /* User:Iggyvolz */ done [18:15:37] yep, ok. [18:21:21] The only thing is when you've motivated yourself to do it merging shouldn't be that hard :P [18:22:35] Damianz: indeed [18:28:04] "code review expired after 1 week of no activity after a negative review, it can be restored using the `Restore Change` button under the Patch Set on the web interface" wtf, as if that's a thing [19:17:38] Ryan_Lane: can you help me with this https://gerrit.wikimedia.org/r/#/c/20925/ [19:17:59] Ryan_Lane: I need to merge that