19:59:40 #startmeeting Infrastructure 19:59:40 Meeting started Thu Jan 20 19:59:40 2011 UTC. The chair is smooge. Information about MeetBot at http://wiki.debian.org/MeetBot. 19:59:40 Useful Commands: #action #agreed #halp #info #idea #link #topic. 20:00:05 #meetingname infrastructure 20:00:05 The meeting name has been set to 'infrastructure' 20:00:16 #chairs skvidal CodeBlock mmcgrath 20:00:21 ok 20:00:23 #topic here 20:00:29 * nirik is lurking around 20:00:42 * CodeBlock was a tad off, thought the meeting was at 4 EST. whoops. ;) 20:01:01 CodeBlock: 3 eastern 20:01:03 :) 20:01:39 ok, so ... who's here? :) 20:01:49 #topic Roll Call 20:01:59 I am here. 20:02:08 present 20:02:33 Alright then, looks like this is going to be a short meeting ;D 20:02:56 #topic Jan 16 iscsi outage 20:03:03 or not. 20:03:18 smooge needs to make you a chair methings 20:03:22 if that didn't happen 20:03:26 he did 20:03:43 well then 20:03:44 anyway 20:03:53 there was an outage on jan 16th at midnight EST...lasted two hours 20:03:57 the command is '#chair' 20:04:23 iscsi/netapp crashed, RHIT was called 20:04:33 #chair skvidal CodeBlock 20:04:33 Current chairs: CodeBlock skvidal smooge 20:04:41 fricking plurrals 20:04:47 problem was fixed by 2AM EST ... affected were db02 (thus FAS and all services which depend on it), ns04, and a few other boxes 20:05:10 #topic what CodeBlock said 20:05:11 #topic Jan 16th iscsi outage 20:05:24 hehe 20:05:53 so that was a fun morning... 20:06:00 any comments on any of that? 20:06:03 anything other than iscsi? 20:06:11 or was it the entire netapp 20:06:12 one additional thing to note is that that takes out the package signing server too... has to be restarted after a db outage it seems. 20:06:17 or does that netapp do just iscsi 20:06:29 the entire netapp was down for a bit. 20:06:38 system is single headed so there was no failover 20:06:49 we are moving to dual headed in the next 10-14 dyas 20:07:02 nirik, ah I didn't realize that. 20:07:06 smooge: 10-14? Over fudcon? 20:07:11 I think we will want to move 20:07:17 CodeBlock, either before or after 20:07:22 but not during 20:07:34 no sleep for the wicked it would seem 20:07:40 haha 20:07:41 so I have better be a lot more wicked 20:08:10 where do we put the information on system dependancies? 20:08:26 such as the one nirik mentioned? 20:09:07 We should probably put that on whatever SOP talks about the package signing server 20:09:52 If someone wants to do that, feel free 20:10:00 Otherwise, I can do that a bit later 20:10:08 Any other comments on this? 20:10:13 there's also something to be said to put that on the SOP for the db server 20:10:15 * nirik could, but hopefully netapp won't die. 20:10:32 both directions of the dependandcy should be mentioned 20:10:42 nirik: it's just a good thing to note, probably. Things happen 20:11:03 sorry, I misspoke... it's after the iscsi outage, not db. 20:11:27 yeah 20:11:40 currently I think we need to look at several changes. 20:11:46 1) Do we need to be on iscsi? 20:12:01 2) What services should be/remain on iscsi? 20:12:46 3) Longer term: how do we break up single points of failure better. 20:13:04 Those are things I hope to tackle with others during the FUDcon talks 20:13:20 that way we can draw out a picture, agree we are all seeing the same thing and move forward. 20:14:43 4) Asset management needs to be dealt with so we can find these things easily. 20:15:12 something that writes out static pages because if the box we are looking relies on the db wiki wont work 20:15:28 anything else from people? 20:18:33 CodeBlock, sorry I took over didn't I :). 20:19:49 * smooge wonders if he has been netsplit and not realized it. 20:20:07 #topic FUDCon 20:20:49 smooge: Good thing you did..sorry my boss had called me into his office to ask me to do something 20:20:54 >.> I'm back now ;D 20:21:03 ok FUDcon NA is going to happen in 9 days. I will be in Az from Thursday to Monday playing santa claus and giving out rpms to good developers (and debs to bad ones) 20:21:21 what plans do people want to cover over Sunday/Monday hackfests? 20:22:13 time permitting, it would be nice to try and drive the ticket count down again. 20:22:19 it keeps going up. ;( 20:22:38 hm 20:23:44 ok that sounds good. I would like to have a "Is this still be worked on?" for old ones, and kill those that dont have an answer in 30 days. If there is already one in the ticket... close it. 20:23:55 they can be reopened later. 20:24:12 yeah, there are some old ones with no response for a while. 20:24:35 Also I would like people to generate logins for all fedora machines from the last 60 days 20:24:52 it will be used to winnow down our sysadmin count 20:25:17 hmm 20:27:38 I have to head to a budget talk for Fedora FY 2011->2012 so I am going to cut this short. 20:27:44 #topic Open Floor 20:28:19 alright...doesn't look like anyone is here to go over tickets (most of the meeting tickets are assigned to ianweller, ricky, skvidal) 20:28:39 I've not touched our nagios 2 -> nagios 3 stuff yet, so ... no update there 20:28:46 I'll do another ticket triage this week 20:28:53 like I've done in the past 20:29:09 and I'll make sure the meeting reminder/agenda is posted weds of next week 20:30:49 Alright, anyone have anything else? 20:31:48 Alright then.. smooge are we having a meeting next week? 20:32:57 yes. 20:33:15 alright 20:33:18 thanks 20:33:23 agenda item for next week: the ecualiptys(sp) stuff 20:33:25 no? 20:33:30 or is that going to be a seperate meeting? 20:33:31 yes that why we have meeting 20:33:39 top of the list then 20:33:41 rbergeron will run it for us :) 20:33:41 got it 20:33:44 Ah, ok 20:33:51 Just wasn't sure because it's so close to fudcon 20:34:07 Anyway - I'm out of here to do some work-related work. ;) 20:34:08 #endmeeting