Web   ·   Wiki   ·   Activities   ·   Blog   ·   Lists   ·   Chat   ·   Meeting   ·   Bugs   ·   Git   ·   Translate   ·   Archive   ·   People   ·   Donate

#olpc-admin, 2009-01-20

Index | Today     Channels | Search | Join

All times shown according to UTC.

Time Nick Message
16:03 hhardy agenda follows:
16:04    *  Status of teach.laptop.org
16:04    * Identify owners for critical existing services
16:04          o Status of RT ticket handling
16:04          o Mission-critical EXISTING services requiring care
16:04          o Migration to W91
16:04    * New business
16:04 dogi hi all
16:04 hhardy lol I'm still here
16:04 either that was short enough or Dogi fixed the flooding parameters
16:04 dogi :)
16:04 edmcnierney Maybe that's our incentive to keep the agenda short!
16:04 hhardy #topic Status of teach.laptop.org
16:04 edmcnierney (Actually, I did trim it down quite a bit)
16:05 OK, I know nothing about this topic - anyone?
16:05 dogi no that ticket is still open ,,, hhardy :P
16:05 hhardy I can say some things
16:05 teach was a workstation operated by Michael Stone
16:06 it was officially a personal workstaton but over time a number of orphan functions accreted onto it
16:06 such as sftp sites for communicating with some of our partners and vendors
16:07 checking wiki docs on it
16:08 hmmm doc says it is a stub
16:08 cjb and also allowing volunteers without Fedora installs to perform package builds there
16:08 hhardy doesn't descibe the users or uses
16:08 I can't be specific without logging into it but there weem to have been at least a dozen people with root on it last time I was there
16:09 plus people from Brightstar and patriot and others with nonprivileged accts
16:09 edmcnierney lfaraone was asking over the weekend if it was OK to upgrade the system.  I suggested it would be fine with appropriate notice to everyone, warnings, etc.
16:09 hhardy the login banner warned that it was unsupptrted use at own risk etc
16:09 thats the high level summary
16:10 cjb edmcnierney: let m_stone know and go ahead if no complaints, I guess
16:10 (he was the previous admin for the machine)
16:10 dogi i did that already
16:11 edmcnierney cjb: I think I suggested m_stone was a required notification-person, but the MOTD should announce it widely for a while to others.
16:11 dogi Michael Stone
16:11 to Stefan
16:11 show details 12:17 PM (3 hours ago)
16:11 Reply
16:11 I believe Luke will do a fine job fixing things up.
16:11 hhardy lol the OLPC bandito
16:11 dogi :P
16:11 edmcnierney Whew - I'm glad we finally got the flood-test taken care of for this meeting
16:13 It sounds like we're OK on the upgrade Luke wants to do then, but I think dogi was also finding a better location for the machine, too.
16:13 dogi yes
16:13 hhardy +1 manage it as a server
16:13 dogi it can be that i have to reboot the machine
16:14 edmcnierney We shouldn't have servers sitting out in the garden - there are too many orphan machines out there, and it will get unplugged or walk away (or both).
16:14 dogi: Are you going to shut it down, move it to the server closet, and start it up again?
16:14 dogi no ed this machine is already in the serverroom
16:15 but not on a good place
16:15 edmcnierney dogi: Too fast for me, then.  Thanks.  But in there is much better than outside.
16:15 hhardy next item?
16:15 edmcnierney Good
16:16 hhardy #topic Identify owners for critical existing services
16:16 1st subitem is: Status of RT ticket handling
16:17 I have given over all but one ticket to Nobody
16:17 I need he help Seth get into the google laptop.;org acct
16:17 I got into it once for Sharon but not sure if I remember the password
16:17 there is a procedure to recover it if need be
16:18 edmcnierney Thanks - so (while I realize there's a lot to be done) does this seem to be in good shape?  That is, anything we (VIG) need to do?
16:18 hhardy Make sure all vig members can see the sysamdin queue
16:19 let them volunteer themselves for tickets and ed and dogi decide if its appropriate
16:19 dogi adding the server/services to bigsister?
16:19 hhardy thats a suggestion
16:19 edmcnierney Sorry, which server - rt?
16:19 hhardy things like setting up git repositories and mailing lists could well be delegated
16:19 to VIG
16:20 edmcnierney On the rt topic, I am now getting the emails from "logcheck system account" on rt, notifying me of "rt.laptop.org 2009-01-20 16:02 Security Events", etc.
16:21 hhardy thats on the rt server on solar
16:21 I changed me to you
16:21 me or adric can change that if you care to nominate a new owner
16:21 edmcnierney These appear to be extremely verbose but mainly benign, so it's not obvious to me how to scan them to find something nasty.  Would it be wise to change me to someone else who's more adept at interpreting them?
16:21 hhardy probably
16:21 nominate Adric
16:22 edmcnierney adric seems to be absent, so that sounds like a good nomination
16:22 hhardy lol
16:22 dogi :P
16:22 edmcnierney hhardy: Can you switch the owner to adric yourself?
16:23 hhardy yes we were going to have me keep root there anyway to halp with rt
16:23 edmcnierney That would be good, then.  He can't possibly do less with those messages than I do, so it will probably be an improvement.
16:23 hhardy *help lol
16:24 edmcnierney Great - move on to next item?
16:24 hhardy #topic Mission-critical EXISTING services requiring care
16:25 teamwiki?
16:25 edmcnierney This is just a placeholder to see if there's anything that's being critically overlooked.
16:25 teamwiki's a good thought.  I need to check with wad on that (he just left the office, so I'll do it tomorrow) and I can come up with a plan for it (will check with smithbone, too)
16:26 CanoeBerry cjb was looking into bringing teamwiki back online.
16:26 would be lovely!
16:28 hhardy restarting exim4 on rt
16:28 edmcnierney That sounds even better - it seems like it was mainly an issue of coming up with arbitrarily different URLs to avoid googlization
16:28 cjl getting teamwiki would be nice, I was working on an RT guide for SG'ers there
16:28 hhardy adric is now happy receiver of voluminous mysterious log emails
16:29 edmcnierney hhardy: Thanks
16:29 hhardy +1 rt guide
16:29 edmcnierney hhardy: Does someone in VIG understand the issues about exim4 and what to do about it?
16:29 hhardy exim4 is the Mail Transport Agent on the rt vserver
16:30 its a debian which is unusual for us
16:30 I restart it whenever I change the rt settings because its so closely tied to rt
16:30 probably doeesn't have to be done
16:31 edmcnierney Oh, I see - thanks.
16:31 hhardy just an excessof caution
16:31 edmcnierney If there's nothing else urgent, let's move on to the W91 equipment move.
16:31 hhardy restarted from init so: /etc/init.d/exim4 restart
16:32 #topic W91 equipment move
16:32 edmcnierney dogi, can you report on the status at W91?  dogi went for the tour on Friday (I had to stay home for an NN meeting)
16:33 dogi yes
16:33 i was on that tour ...
16:34 edmcnierney Sorry - my machine has lost its mind - brb.
16:34 dogi and after that i set owl on rails
16:34 hhardy edmcnierney: a likely story ;)
16:35 dogi and i set up an Sugar/XO with wifi back entrence to W91
16:35 hhardy are you certain this might not create a routing loop?
16:36 cjl W91 is a closet at 1CC or MIT hosting?
16:36 hhardy cjl: the latter
16:36 cjl k
16:36 dogi so now if W91 line to medialab is down i can try remotly to find/fix the problem
16:36 sorry
16:37 W91 is one of many Mit hosting buildings
16:37 hhardy yes what did they have to say at tour?
16:37 dogi too many things
16:38 cjl Highlights?
16:38 dogi but it was more a showing what we are able to do tuor
16:38 hhardy is there any change as far as w91?
16:38 dogi no not for now ...
16:39 hhardy ok good
16:39 so what is the migration plan for w91 which machines are going?
16:39 cjl How many machines does OLPC have at W91 now?
16:39 dogi first awk
16:39 hhardy that is a 1u ibm box with 2 gb ram
16:40 dogi that is to change to 8gb
16:40 edmcnierney Hmm.  Sorry - I'm back, but seem to have lost my backlog - carry on.
16:40 dogi and the ilo is still not working
16:41 hhardy #LINK http://meeting.laptop.org/olpc[…]0090120_1603.html
16:41 edmcnierney Oh, right - thanks.
16:41 dogi has somebody experience with ibm redbook stuff?
16:42 the next machine is weka
16:42 but bevor moving weka we have to move the proxy
16:43 cjb do we have control of the laptop.org DNS yet?
16:43 edmcnierney cjb: I asked mbletsas about that and he replied that MIT manages it and he only relays requests on to the network team there.
16:44 cjl edmcnierney: what about all the doamins for which mbletsas is listed as admin contact in a whois?
16:44 edmcnierney This seemed very slightly implausible, but it was clear that that was the answer I was going to get that day.  I need to do more investigation.
16:45 hhardy [meeting minutes are posted now]
16:45 edmcnierney cjb: Yes, those are all on my to-do list.  We have quite a number of odd domains floating about with only one person listed as both technical and admin contacts
16:45 cjl Hate to see one come up for renewal and get squatted
16:45 edmcnierney I will follow up with mbletsas if I need to get more info after I complete the list of contacts that need updating.  I don't think that's a problem.
16:45 cjl speaking with British accent, many people make that mistake
16:46 cjb cjl: hm?
16:46 hhardy especially cjb :)
16:46 edmcnierney I did a quick renewal check - something's due in March, I think, so it's not critical but needs to be handled promptly.
16:46 hhardy not sure its a mistake tho
16:46 cjb hhardy: One would certainly hope not ;-)
16:47 hhardy cjb: Surely not sir!
16:47 cjb yeah, laptop.org expires in two months
16:47 edmcnierney That would be it, then.
16:47 cjl then there is laptopfoundation, etc. etc. etc.
16:47 edmcnierney Presuming, of course, you mean the domain registration for laptop.org expires in two months, and not something more sinister.
16:47 dogi xogiving
16:48 cjb :)
16:48 cjl I had hunted a bunch down at one point when I was doing my investigation of who you shady characters were. . .
16:48 hhardy beware the Ides of March
16:49 edmcnierney So I own the DNS cleanup tasks.  I'll report back if I need help.
16:49 cjl edmcnierney: If I find my notes, I'll send them to you.
16:49 edmcnierney cjl: Thanks.
16:50 hhardy any new business?
16:50 edmcnierney None from me!
16:50 cjl some continuing points I'd like to touch on
16:50 hhardy #topic New Business
16:51 cjl Pootle.  Bernie has unfortunately fled the country ahead of the INS agents.
16:51 hhardy it would be more unfortunate if he were not ahead of them...
16:52 edmcnierney I did speak to Bernie about what we could/should do to support Pootle and other sl infrastructure.
16:52 cjl I'm not sure how in touch he will be, but there is not a solid future (final) home for Pootle at SL.  Dogi has kindly built a little sandbox, whcih will have it's uses, but not be a fianl resting place.
16:52 dogi dogi     pts/1        kuku.laptop.org  Tue Jan 20 16:49   still logged in
16:52 bernie   pts/1   Tue Jan 20 01:26 - 01:48  (00:21)
16:52 sayamind pts/1   Sun Jan 18 06:19 - 13:23 (1+07:04)
16:52 cjl      pts/1        ool-45735eb0.dyn Sat Jan 17 22:56 - 23:41  (00:44)
16:53 that is "last" on pootle VM
16:53 cjb perhaps OLPC can help SL find a machine to host the migrated Pootle on
16:53 hhardy awk?
16:53 cjb which pootle VM?
16:53 oh, the new sandbox thing
16:53 dogi bingo
16:53 edmcnierney I've asked dogi to start the work to (a) move all public-facing services to W91 and (b) help normalize what runs on what machines there.  All of that is in support of finding a machine to host the Pootle services for SL
16:53 cjb VM's aren't very good for pootle, it uses as much CPU as it can get
16:54 cjl cjb I think SL would be interested in such a possibility
16:54 hhardy cjb: +1 its a hog
16:54 cjl and RAM
16:54 edmcnierney I pushed back on Bernie a bit, as he was quite eager to get a server set up.  I would very much like to do that, but I want to understand (a) all our services and (b) all our servers so we can make a global allocation decision rather than one at a time.
16:54 cjb so I wouldn't put it on a machine that hosts VMs without serious cpu and memory limits
16:54 cjl Ther are some anticipated improvements in the 1.3 Django rebase expected soon, but it's not oput yet.
16:55 hhardy it uses at least up to 1.7 gb of ram
16:55 cjl edmcnierney: That is an understandable point of view
16:55 edmcnierney I'm sure we can help - we've got an awful lot of hardware - we just need to get a comprehensive plan together.  I don't think that's incredibly hard; it's just a bit of work.
16:56 cjl and quite reasonable in the face of a continued willingness to host the existing Pootle instance until everything shakes out.
16:56 dogi i had to restart pootle on crank twice last 10 days
16:56 hhardy #topic Pootle
16:57 cjl I beleive the urgency that is felt about a Pootle migration is everyone on all sides acknowledging it's importance.
16:58 hhardy 5 minutes to 5
16:58 edmcnierney Oh, thanks - I do need to leave at 5 to catch the train.
16:59 hhardy is there an action item regarding Pootle?
16:59 cjb cjl: I don't think I understand why it would be urgent
16:59 for one thing, it's not appropriate for the migration to be needed "urgently" while SL does not have a machine to put it on.  That doesn't make any sense.
16:59 cjl edmcnierney: I believe that some messign around on the sandbox VM tha dogi has provided will be productive and that as long as Pootle has a home at OLPC until a proper fostering can be arranged, then I think that will be great.
17:00 edmcnierney cjl: Yup, I think that's the plan.
17:00 cjb agreed
17:00 dogi +1
17:00 hhardy +1
17:00 cjl cjb I belevie with hte layoffs at OLCP, the loss of contractors and admins, there was room for some concern about the infrastructure.  Those fears can be somewhat allayed.
17:01 hhardy better safe than sorry tho
17:01 cjl Best done wit hcelar discussion amongst all stakeholders and in a carefully planned manner.
17:01 cjb sure.  we don't want to run that infrastructure anymore, so it's in everyone's interest to hand it off to someone like SL
17:02 let's wait until they're ready to take it, though :)
17:02 edmcnierney cjl: Actually, although we're certainly short-staffed, it's even MORE important for us to maintain infrastructure so a broader OLPC-friendly community can take advantage of it.
17:02 hhardy cjl: is that a euphemism for "blind panic"? :p
17:02 edmcnierney That means keeping servers/bandwidth/etc. available so folks can do their own things on them.
17:02 cjl cjb, I think some urgency (on bernie's part) may have to do with his leaving the country and wanting things squared away.
17:02 cjb oh.  sure.
17:02 I wonder how SL sysadminning will happen with Bernie out of the country
17:03 maybe I should volunteer for some of the physical access stuff or something.
17:03 edmcnierney cjl: Yes, and that's what I resisted a bit, as we really weren't quite ready to do that.  Bernie needs to learn to schedule his travel around layoffs better.
17:03 cjl cjb: he has identified back ups for various functions, but I'm sure his presence will be missed
17:03 hhardy #LINK http://lists.sugarlabs.org/listinfo/systems
17:04 edmcnierney I need to take off, folks - I'll read the log tonight, so send any tasks my way if needed.  Thanks.
17:04 cjl cjb: I think additional hands to contribute to tha work would be welcome.
17:04 evening ed
17:04 hhardy now is the time go give ed work! bwahaha
17:04 L)
17:05 anything further?
17:05 cjl nope
17:05 Oh, did the SSH ticket get closed for adric?
17:05 hhardy #endmeeting

Index | Today     Channels | Search | Join

Powered by ilbot/Modified.