infrastructure
LOGS
19:00:02 <nirik> #startmeeting Infrastructure (2011-03-24)
19:00:02 <zodbot> Meeting started Thu Mar 24 19:00:02 2011 UTC.  The chair is nirik. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:00:02 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
19:00:03 <nirik> #meetingname infrastructure
19:00:03 <zodbot> The meeting name has been set to 'infrastructure'
19:00:03 <nirik> #topic Robot Roll Call
19:00:03 <nirik> #chair goozbach smooge skvidal codeblock ricky nirik
19:00:03 <zodbot> Current chairs: codeblock goozbach nirik ricky skvidal smooge
19:00:21 * CodeBlock is short-time/will not be here for long.
19:01:02 <smooge> my head feels like puffy sausages so will be a bit slow
19:01:39 * lmacken 
19:01:43 * jsmith lurks
19:01:43 <nirik> I didn't have much for agenda today... so hopefully a shortish meeting. ;)
19:02:13 <smooge> yeah
19:02:17 * waltJ is here.
19:02:27 <nirik> ok... lets dive in...
19:02:50 <nirik> #topic new folks introduction
19:03:01 <nirik> Any new folks around who would like to say hi and get more involved?
19:04:04 <nirik> ok, we have had a few chime in on list. I would encourage them to show up to meetings and say hi in #fedora-admin
19:04:25 <nirik> #topic pkgs01 iscsi issues
19:04:44 <nirik> so pkgs.fedoraproject.org has had periodic issues with it's lookaside storage of late.
19:05:00 <nirik> dgilmore suspects it might be related to it being a rhel6 guest on a rhel5 xen host.
19:05:07 * skvidal is here
19:05:19 <nirik> so, we might try and move it to a rhel6 kvm host if we can find a place.
19:05:48 <nirik> or open to other ideas.
19:06:17 <nirik> we might want to do this move at the same time as the possible db03 move in april, since that might be downtime for packagers anyhow.
19:06:17 <skvidal> do we have any new kernels
19:06:23 <skvidal> we can try on rhel6 under rhel5 xen?
19:06:32 <nirik> skvidal: it's running the newest I think.
19:06:41 <skvidal> nirik: 6.1b got announced the other day
19:06:41 <nirik> I rebooted it last time this happened to get it up to the latest.
19:06:49 <skvidal> dir drjones give us anything more recently?
19:06:52 <nirik> true... we could look for fixes in that kernel.
19:06:53 <skvidal> s/dir/dir/
19:07:12 <smooge> the problem is still outstanding
19:07:23 <smooge> it is only seen here on our stuff
19:07:29 <skvidal> what are the reqs for pkgs01?
19:07:38 <nirik> also, a quick task for someone: nagios check for read-only happening there, so we notice it fast instead of having user reports. ;)
19:07:40 <skvidal> it looks like it is using a lot of disk space in /?
19:07:42 <smooge> but bvirthost01 might be able to take it
19:08:29 <skvidal> ah - I see /srv/git is mounted locally
19:08:36 <skvidal> for some reason I figured that would be elsewhere
19:08:40 <nirik> skvidal: yeah...
19:08:55 <skvidal> but even so it's only 60GB
19:09:01 <skvidal> (for all of /)
19:09:16 <skvidal> smooge: is bvirthost01 hw we can rely on?
19:09:23 <smooge> it is new ahrdware
19:09:25 <skvidal> yah - it's virthost13 I'm thinking of which is 'odd'
19:09:28 <dgilmore> skvidal: i believe its new hardware
19:09:29 <nirik> smooge: did you have a bug on the rhel6 on rhel5 xen thing?
19:09:31 <smooge> it has only run 1 server
19:09:47 <smooge> nirik, yes. I have several bugs on that
19:10:11 <nirik> ok, if you get a chance to shoot me the #'s, I can cc myself...
19:10:30 <smooge> nirik, I am at the point of when we can decommision something in PHX2 we are shipping the IBM hardware to a kernel admin with a blood sacrifice
19:10:48 <nirik> :)
19:11:10 <nirik> which db's are on db03?
19:11:51 <skvidal> koji I think
19:11:57 <skvidal> that's all iirc
19:11:59 <skvidal> lemme find my notes
19:12:33 <smooge> .bug 632802
19:12:34 <zodbot> smooge: Access Denied - https://bugzilla.redhat.com/show_bug.cgi?id=632802
19:12:41 <skvidal> nirik: yes, just koji
19:12:43 <dgilmore> only koji is on db02
19:12:50 <skvidal> dgilmore: db02?
19:12:52 <dgilmore> 03
19:12:55 <skvidal> ah, whew
19:12:55 <dgilmore> typo
19:12:56 <nirik> yeah, so, if db03 is koji, packagers would have an outage there anyhow for moving db03... so we could do pkgs01 in the same window.
19:13:02 <skvidal> nod
19:13:10 <dgilmore> its only there because we had rpmfiles and one other table
19:13:16 <smooge> nirik, it is 632802. Not sure if you can join it
19:13:19 <dgilmore> they were massive
19:13:29 <dgilmore> millions of rows
19:13:36 <dgilmore> and frequently hit
19:13:38 <nirik> smooge: ok, not sure either, but will note it.
19:13:51 <dgilmore> but kojis db needs have changed since
19:14:02 <dgilmore> it doesnt need gobs of ram and cpu like it did
19:14:19 <smooge> nirik, once I get an idea of what we would like to replace db03 with even if its a virtual box to run it and other stuff.. I will get a quote from CDW next week
19:14:40 <smooge> that could become bvirthost02 and that would allow for all kinds of moves/cleans etc
19:14:48 <nirik> ok.
19:15:16 <dgilmore> it could be a virt host
19:15:28 <nirik> we may want to look at our current db's and needs. I know we wanted to think about how to replicate/move SPOF for them if possible.
19:15:40 <nirik> so, for now I just wanted to get the pkgs01 thing on everyones radar.
19:16:35 <nirik> I think moving it would be good when we have a good window and place.
19:17:15 <skvidal> nod
19:17:17 <nirik> #idea possibly move pkgs01 when there's a db03 move outage.
19:17:38 <nirik> #action need to add a check for detecting the r/o condition.
19:17:40 * skvidal is curious
19:18:02 <skvidal> if we had one piece of hw could db03 and pkgs01 run on the same virt host?
19:18:05 <skvidal> seems like they could
19:18:05 <nirik> I can add that check if no one beats me to it.
19:18:26 <dgilmore> skvidal: probably
19:19:17 <nirik> ok, anything more on this? or shall we move on?
19:20:26 <nirik> #topic puppet reorg
19:20:51 <nirik> skvidal has been redoing puppet stuff. ;) Care to give an overview of the changes for folks?
19:20:59 <skvidal> sure
19:21:11 <skvidal> first: puppet is no longer running as a daemon for us on our systems
19:21:25 <skvidal> it runs out of cron every half hour - at 23 minutes after the hour and 53 minutes after the hour
19:21:32 <skvidal> it randomly waits up to 10m
19:21:33 <skvidal> before running
19:21:45 <skvidal> so we don't get the stampeding herd problem
19:22:06 <skvidal> so
19:22:16 <skvidal> we have gotten rid of the 'is puppet running' nagios check
19:22:26 <smooge> actually the pkgs item is pretty 'simple' it is on iscsi and bvirthost01 should have a mount for it (if not I will put it in the queue). Just undefine on one box and define on another and we are golden
19:22:29 <skvidal> I've written a script to check to see if puppet has checked in to the puppetmaster
19:22:31 <smooge> oh sorry
19:23:01 <skvidal> and it will emit a notice to the nagios server using func + nsca to let us know a box has not run puppet in N amount of time
19:23:08 <skvidal> finally we've been working on the puppet error reports
19:23:24 <skvidal> and we've trimmed them down - but it's still an issue
19:23:30 <skvidal> b/c some of them won't go away for quite a while
19:23:52 <skvidal> I'm actively working on a new script to parse the reports, per host, and generate a checksum of the errors/warnings
19:24:00 <skvidal> if the checksum changes, then it will send us a notice about that host
19:24:07 <skvidal> if it doesn't then it won't so we should see errors only once
19:24:14 <skvidal> to make sure we don't forget about them
19:24:22 <skvidal> I'm going to have the checksums nulled out once a month
19:24:22 <smooge> skvidal, thank you very much for doing all that work.
19:24:24 <skvidal> so we get notices
19:24:36 <skvidal> it's interesting - parsing the puppet report yaml
19:24:39 <nirik> skvidal: yeah, thanks a lot. It's great stuff to get fixed up.
19:24:53 <skvidal> it'll be nice to trim out a lot of crap
19:24:58 <skvidal> something that someone could take on
19:25:01 <skvidal> if they wanted to
19:25:21 <dgilmore> smooge: its iscsi is on the equalogic not netapp
19:25:22 <nirik> smooge: we still would need a outage to sync over the / data, but yeah... I guess we can do it sooner if people think. It's not been super often it does tho
19:25:24 <skvidal> 1. we need someone to add aliases to our nagios hosts to correspond with their real host name, their vpn hostname and others
19:25:55 <skvidal> 2. it would be worthwhile, I think, to go through our puppet config and remove projects which are never coming back - Istr seeing zabbix in there
19:26:13 <skvidal> 3. I would love to have some volunteers for audits on systems and the pkgs/files they have installed
19:26:20 <smooge> dgilmore, on bxen03 it is running from the netapp. It mounts other stuff from the equalogic via iscsi
19:26:20 <skvidal> sorry that took so long to get out
19:26:33 <dgilmore> smooge: huh
19:26:48 <nirik> skvidal: no worries.
19:26:48 <marchant> i volunteer to do audits, if appropriate
19:26:49 <dgilmore> smooge: there is 2 volumes on the equalogics units
19:26:55 <dgilmore> /mnt/koji and lookaside cache
19:26:58 <nirik> marchant: excellent!
19:27:21 <marchant> i do not have appropriate access priv's though, more than likely
19:27:27 <skvidal> marchant: good - a lot of the auditing is going through sets of rpms on machines and assuming nothing belongs on there
19:27:29 <smooge> dgilmore, all I am saying is on bxen03 doing an lvs shows that the virtual machine pkgs01 is on xenGuests and open which is fromt he netapp
19:27:45 <skvidal> marchant: for what I've just suggested for pkg auditing - you don't need any other privs than you have
19:28:05 <smooge> skvidal, I really appreicate the work. it was needed
19:28:07 <dgilmore> smooge: pkgs01 and nfs01 are the only things accessing the equalogics
19:28:08 <marchant> that sounds like a good fit then
19:28:22 <skvidal> marchant: I can generate a list of every pkg on every system for you
19:28:31 <dgilmore> smooge: or are we confusing things
19:28:33 <skvidal> marchant: and we can weed out the crazy
19:28:52 <marchant> OK, tell me what you need and I will be happy to help
19:28:53 <skvidal> marchant: another important thing (to me) is going through machines to find places where machinetype01 doesn't match machinetype02
19:28:55 <dgilmore> smooge: are you saying that pkgs raw disk is a netapp iscsi volume?
19:29:07 <smooge> dgilmore, yes. sorry for not being clear
19:29:23 <nirik> hey folks, I have to step away for a bit. would someone be able to take over running things? we have meeting tickets if anyone has any they want to address and open floor...
19:29:26 <marchant> skvidal: do you mean as far as the installed packages
19:29:30 <skvidal> marchant: yes
19:29:31 <smooge> dgilmore, so it should be very simple to get it to work on something else.
19:29:36 <skvidal> marchant: I'll email you with more
19:29:38 <nirik> smooge: ah, that makes it easier. I see what you are saying.
19:29:44 <marchant> skvidal; totally makes sense
19:29:54 <marchant> skvidal: thanks
19:29:59 <skvidal> marchant: nod
19:30:16 <nirik> smooge: so, outage would be minutes, not long...
19:30:25 <smooge> yes.
19:30:36 <skvidal> just an idea - let's assume the outage involves rebuilding the box
19:30:36 <skvidal> :)
19:30:38 <smooge> and could go back quickly if it didn't wrok at all
19:30:41 <skvidal> and plan for 2 hours
19:30:44 <skvidal> just b/c
19:31:00 <smooge> skvidal, I always plan via the Scotty method.
19:31:06 <nirik> skvidal: always good to plan for disaster, and deliver only minutes of downtime. ;)
19:31:14 <skvidal> smooge: I'm givin' her all I can capt.
19:31:21 * nirik has to go visit sick down now.
19:31:27 <skvidal> dog?
19:31:29 <smooge> by
19:31:37 <skvidal> unless your dog is named 'down' which would be funny
19:31:42 <skvidal> 'come, down'
19:31:46 <skvidal> 'down, heel'
19:31:49 <smooge> up, down
19:31:50 <nirik> ha. dog, yes.
19:31:54 * nirik can't type
19:31:58 <smooge> down down
19:32:12 <nirik> if someone could take over that would be great.
19:32:15 * nirik leaves for a bit.
19:32:16 <smooge> ok skvidal what is your view of what to happen on puppet
19:32:24 <skvidal> 'happen'?
19:32:26 <dgilmore> skvidal: i think that once we declare beta gold
19:32:26 <smooge> for the next week? And what can we help with
19:32:36 <dgilmore> ill take a buildsys outage
19:32:39 <dgilmore> do db03
19:32:44 <dgilmore> and pkgs01
19:33:12 <skvidal> puppet is working as normal right now
19:33:25 <skvidal> I should be able to finish the script to handle the error reports today
19:33:26 <skvidal> if I do
19:33:30 <skvidal> I'll disable tagmail for everyone
19:33:38 <skvidal> and start sending out this mail instead
19:33:56 <skvidal> one thing to keep in mind
19:34:02 <skvidal> these error reports will come out half-hourly
19:34:17 <skvidal> I have not found a way to make them driven by when puppet runs on the host
19:34:38 <skvidal> only by the change of state in the reports
19:35:20 <skvidal> smooge: oh!
19:35:25 <smooge> oh?
19:35:26 <skvidal> one thing I completely blanked on is bxen01
19:35:31 <skvidal> the box is reinstalled
19:35:38 <skvidal> and needs to be moved to the community network/rack
19:35:48 <skvidal> then I'll install the autoqa02/bastion## box over there
19:35:50 <smooge> yes.. the box for the hardware move that we will someday do when I GET A FING NETWORK
19:35:56 <skvidal> ah
19:35:59 <skvidal> so I'm not holding anything up
19:36:01 <skvidal> fabulous
19:36:28 <skvidal> there are some other tasks to clean up our hosts which I wouldn't mind some other eyeballs on
19:36:32 <skvidal> if anyone is interested in them
19:36:35 <skvidal> some of them are DULL
19:36:44 <skvidal> but shouldn't take very long
19:36:57 <skvidal> I added them to our FI clean up 2011 list
19:37:01 <skvidal> https://fedoraproject.org/wiki/Infrastructure_Cleanup_Tasks_2011#Fix_all_the_things_that_we_have
19:37:03 <skvidal> at the bottom
19:37:35 <smooge> no no.. just keeping me from reaching for the battle axe and woad
19:37:47 <skvidal> the cron entries in /var/spool/cron and /var/spool/mail
19:37:57 <skvidal> I know its not interesting but imo we shouldn't have ANYTHING in either of those directories
19:38:03 <skvidal> and if we do it damn well better be known about in puppet
19:38:09 <skvidal> if it is not then SOMETHING IS WRONG
19:38:13 <smooge> I am going to start on those on Monday. I hope I get a weekend off without systems having drama issues
19:38:40 <skvidal> no kidding
19:38:54 <smooge> speaking of which... log02
19:39:15 <smooge> if we can move to the next item?
19:39:31 <skvidal> yah
19:39:37 <smooge> #topic logging
19:40:04 <smooge> we had a bit of a scare on Saturday with log02's xen box dropping one of its disks and everything going into degraded mode.
19:40:09 <smooge> s/log02/log01/
19:40:27 <smooge> so I started working on a log02 and CodeBlock picked up hat I broke and started cleaning things
19:40:56 <smooge> I am hoping that after this meeting I will start an rsync of the data from one to the other
19:41:06 <smooge> and we should be able to 'move' over to log02 next week
19:41:42 <skvidal> do we need to run over to log02?
19:41:56 <skvidal> is the box that log01 is on in that bad of shape?
19:42:06 <smooge> skvidal, well xen10 goes out of service in early June.
19:42:21 <smooge> so I want to get ahead of the curve
19:42:26 <skvidal> okay - I'm just wondering - can we take some time to make log02 'better'
19:42:30 <skvidal> and get it configured
19:42:39 <skvidal> before we start inundating it with data
19:42:41 <smooge> sure..
19:43:00 <smooge> I am not sure what better is so I guess we can talk about that and then do
19:43:12 <skvidal> 1. I'd like to pursue epylog reports
19:43:33 <skvidal> 2. I'd like to break our hosts out into report groups - so we can send log reports from the proxy servers to one group of people
19:43:36 <skvidal> (for example)
19:44:06 <skvidal> 3. it might be nice to consider a real log-rotation/expiration policy that isn't 'keep everything forever'
19:44:29 <skvidal> both to protect ourselves from issues and to just limit the sheer size of things.
19:44:35 <skvidal> does 3 seem ridiculous to anyone?
19:44:51 <jsmith> skvidal: Not to me
19:44:54 <smooge> hmmm pretty much every place I have ever worked has always found that in the end it comes down to "keep everything forever". So I just figured to assume that first :)
19:45:05 <skvidal> smooge: everywhere I have worked has had specific policies
19:45:09 <skvidal> with expiration timeouts
19:45:15 <skvidal> for CYA against lawsuits/subpoenas
19:45:32 <smooge> yeah. we had policies and every time they came up to be deleted some law came in and we ahd to revise and keep things longer.
19:45:34 <skvidal> if you have a policy of 'we keep logs for 6 months and kill them' then a subpoena gets nothing
19:45:35 <fenrus02> compress and archive to off-site storage only. that way, legal can get the logs if required.
19:45:50 <skvidal> fenrus02: and if you have an official policy you can hand that back to legal
19:45:56 <skvidal> PROVIDED YOU FOLLOW THE POLICY
19:46:06 <fenrus02> point
19:46:18 <fenrus02> "keep everything forever" is simply not sustainable.
19:46:42 <skvidal> smooge: we don't work for the gov't of any kind
19:46:48 <skvidal> we have no obligation to keep everything forever
19:47:06 <smooge> skvidal, heck I was thinking of startups and .edu
19:47:28 <skvidal> so was I - and all the lawyers and infosec people said "have a policy and keep to it"
19:47:32 <skvidal> duke was militant about it
19:47:35 <smooge> but we don't service anything medical
19:47:55 <skvidal> you know what
19:48:00 <skvidal> I don't want an argument about it
19:48:02 <skvidal> I'll work on epylog
19:48:07 <skvidal> if y'all decide if you want a policy
19:48:08 <smooge> I am not arguing geez
19:48:08 <skvidal> great
19:48:09 <marchant> the "have a policy" bit is the trick
19:48:39 <skvidal> if not, then provided we can continue to poop infinite disk space, fantastic
19:49:28 <skvidal> but before we deploy log02 it might be a good idea to know what reports we want and what reports we actually are GETTING and READING
19:49:38 <skvidal> do we know this? is it written up anywhere?
19:49:44 <marchant> poop disk space?
19:50:06 * marchant that's amazing
19:50:37 <skvidal> marchant: we have a goose which instead of laying golden eggs
19:50:40 <skvidal> it poops diskspace
19:50:54 <skvidal> it's fabulous, actually, but sheesh cleaning the disk space before you use it is usually advised
19:51:14 <skvidal> smooge: what's next?
19:52:02 <smooge> https://fedorahosted.org/fedora-infrastructure/query?status=new&status=assigned&status=reopened&group=milestone&keywords=~Meeting&order=priority
19:53:22 <skvidal> so we've covered a number of these in the earlier conversations
19:53:37 <skvidal> which is good
19:53:47 <skvidal> 2591 is curious - what's our status on blogs?
19:55:31 <sijis> i put a list together of what i found on the usage of blogs
19:55:42 <sijis> i attached it to the ticket.
19:56:11 <skvidal> nod - so it's a larger number than I expected
19:56:20 <smooge> a lot larger number
19:56:21 <skvidal> do we have any concern/reticence with moving them?
19:56:31 <skvidal> nod
19:56:35 <skvidal> it is a lot more
19:57:40 <smooge> ugh I am not sure.
19:58:31 <smooge> a lot of them don't seem to be updated.. which is normal for blogs.. but its a lot to archive/move OR deal with otherwise
19:58:56 <jsmith> Has the blog infrastructure been difficult to manage in general?
19:59:09 <skvidal> jsmith: yes
19:59:17 <skvidal> wordpress requires pretty much constant attention
19:59:28 <smooge> it is about the same case as our transifex issue
19:59:43 * jsmith doesn't know of a web application that doesn't require a lot of attention, unfortunately
20:00:00 <smooge> it is a definate behind the curve.
20:00:26 <jsmith> Is it just a matter of updating to the latest wordpress package, or are there other items that keep you up at night?
20:01:06 <smooge> well updating it is part of the problem
20:01:18 <smooge> we are i believe running on stuff that is dead software now
20:01:36 <skvidal> jsmith: the problem is that wordpress is frequently vulnerable
20:01:41 <smooge> and updating it is going to be an act of faith and a large focus
20:01:45 <skvidal> jsmith: and the -mu we're using is a dead bracnh
20:02:02 <jsmith> Right... the -mu functionality got rolled into the main branch
20:02:30 <skvidal> jsmith: the issue in my mind is that the time we spend baby sitting this is expensive
20:02:30 <jsmith> I'm pretty good friends with some of the automattic folks -- is it worth asking if they'd be interested in helping with a migration?
20:02:37 <jsmith> Or is it just not worth the effort?
20:02:43 <skvidal> I dunno
20:02:46 <skvidal> the questions I'd ask are
20:02:59 <jsmith> For me, I see it as a valuable resource for the community, but don't have any way to gauge the sysadmin drain it causes
20:02:59 <skvidal> 1. why do we have 100 blogs hosted?
20:03:10 <skvidal> 2. what do we get out of having it run locally?
20:03:37 <skvidal> 3. how many hours have been spent on it since it was installed
20:03:44 <skvidal> #3 might be a question for ricky
20:03:49 <skvidal> since I know he's done a lot of work on it
20:04:18 <smooge> 4. who do we go to fix it when its not working.
20:04:35 <smooge> which usually ends up being ricky or nb (I think)
20:05:43 <skvidal> nod
20:06:48 <lmacken> i'd love to see us switch to a flat-file git based blog system, like blog-o-file or pyblosxom :)
20:07:17 <skvidal> lmacken: pyblosxom --- haven't heard that in years
20:07:29 * nirik returns, reads up
20:07:31 <skvidal> lmacken: is it still being maintained?
20:07:33 <lmacken> skvidal: I had lunch with the creator of it at PyCon  :)  then worked with him to upgrade lewk.org to the latest code
20:07:37 <lmacken> skvidal: yup
20:07:42 <skvidal> lmacken: wow, that's cool
20:07:48 <skvidal> lmacken: I used it a looooooong time ago
20:07:52 <lmacken> yeah, he's based out of Boston and hacks on Miro
20:07:57 <skvidal> lmacken: I liked editing  in my editor and rsyncing the files up :)
20:08:07 <skvidal> lmacken: (this was all pre-git)
20:08:25 <lmacken> yeah, I do like the concept of blog-o-file too... git push hooks that compile your entries down with Mako templates, so your blog is purely static files
20:08:44 <skvidal> lmacken: makes comments systems harder
20:08:50 <skvidal> sorry, this is off-subject a bit
20:08:52 <lmacken> yeah, most people just throw disqus in there
20:09:07 <lmacken> sorry for de-railing :)  I hate wordpress.
20:09:26 <nirik> ok, we are over an hour too. ;) should we open floor/close out?
20:09:43 * sijis sorry. had to step away
20:10:54 <nirik> #topic Open Floor
20:11:03 <nirik> anything for open floor?
20:11:28 <maxamillion> uhmmm... not in official space, just wanted to say hi to everyone and apologize for my tardiness :)
20:11:34 <smooge> hi maxamillion
20:11:36 <maxamillion> $dayjob got in the way of the meeting
20:12:00 <nirik> no problem. welcome maxamillion
20:12:19 <maxamillion> I'm currently trying to learn my way around the infrastructure layout a bit and then I'll be taking on a couple of the line items from https://fedoraproject.org/wiki/Infrastructure_Cleanup_Tasks_2011 ... time and perms permitting anyways
20:12:31 <nirik> excellent.
20:12:56 <maxamillion> I also have no idea what all I have permissions to ssh into (if anything at all)
20:13:08 <nirik> we can sort that out...
20:13:10 <maxamillion> I assume that's handled by FAS groups
20:13:13 <maxamillion> nirik: ok, sounds good
20:13:15 <nirik> yep.
20:13:39 <smooge> you have permission to log into any host you are allowed to. all systems you aren't will initiate a dd if=/dev/zero of=/dev/sda
20:13:46 <smooge> of your system
20:13:54 <maxamillion> smooge: lol
20:14:14 <maxamillion> sounds like fun
20:14:17 <nirik> ok, thanks for coming everyone...
20:14:52 <nirik> #endmeeting