Switch to DuckDuckGo Search
   October 24, 2013  
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | >

Toggle Join/Part | bottom
[00:03:31] *** badboy_ has quit IRC
[00:04:40] *** badboy_ has joined #project-FiFo
[00:07:12] <arekinath> killfill: http://uq.edu.au/
[00:08:41] <killfill> oh your from australia
[00:09:22] <killfill> so what are you using fifo for there?
[00:09:36] <killfill> giving all students new zones?
[00:14:09] <arekinath> yeah, basically. we've been using smartos for a few years now, and Solaris zones since s10 came out
[00:14:52] <arekinath> they get used a lot for teaching, but also quite a few research and infrastructure projects
[00:15:39] *** mgls has quit IRC
[00:15:53] <arekinath> we mainly run on Solaris and openbsd, with a little bit of win server and linux on the edges
[00:17:03] <Licenser> wow smartos is that old?
[00:24:18] <arekinath> our first production install was oct 2011
[00:24:28] <arekinath> we had some toys going before that
[00:26:01] <arekinath> we still have some zones on smartos-1.3.x datasets, haha
[00:26:24] <arekinath> they've been hacked a bit to keep working with a newer gz
[00:26:41] <arekinath> joyent aren't very good at backwards compat where backwards compat is > 1 year or so
[00:33:48] <arekinath> the way they run their kit seems to be that new hosts get a new image, but no older server is ever upgraded
[00:34:19] <arekinath> until eventually the old service offering is retired and everyone has to migrate to their new shiney thing
[00:37:25] <arekinath> even if the old service offering was advertised as "for life"! :P
[00:37:29] <Licenser> arekinath sadly true, backwards compatibility isn't a strong suite
[00:37:34] <arekinath> anyone for a mixed grill?
[00:37:44] <Licenser> mixed grill sounds good
[00:37:46] <Licenser> <3 meat!
[00:37:54] <arekinath> Licenser: not a real one. :P
[00:37:58] <Licenser> darn it :(
[00:38:29] <arekinath> I was talking about the textdrive stuff
[00:39:15] <arekinath> that's the way joyent deals with legacy. shut it down and give everyone a few free months on some new shiney thing that's quite different. :P
[00:39:27] <arekinath> no matter what they said about its lifetime in the past
[00:39:42] <Licenser> yap that was a rather sad episode
[00:41:54] <arekinath> but anyway, you should never maintain expectations that a company will continue to do something unless it makes them money
[00:42:26] <arekinath> even a contract that says they have to pay money if they don't continue, only puts it off
[00:42:46] <Licenser> yea I know
[00:42:52] <Licenser> it's still sad
[00:43:21] <Licenser> also it was quite a pickle not sure how to get out of that kind of corner for a company without either loosing face or lots of money
[00:51:02] <killfill> arekinath: can you give me some details about i.e. how much servers does your fifo runs on? or how many machines are runnng or been created per course?
[00:52:31] <arekinath> killfill: we currently have 4 dell boxen for smartos, each with 32 cores and 128GB ram
[00:52:40] <arekinath> killfill: hosting just over 500 zones and a handful of kvms
[00:53:18] <arekinath> we're looking at expanding that though so that we can host more kvm on it, since kvms eat up ram like mofos
[00:53:56] <killfill> heh
[00:54:13] <arekinath> out of the 500 zones about 400 are teaching zones, some are for an individual student's project, some for groups/teams
[00:54:44] <arekinath> those 400 will all be gone in the next few weeks after the semester ends and things will be very quiet until the end of summer
[00:55:06] <killfill> greate
[00:55:16] <arekinath> then we expect next year probably 500-600 teaching zones to be ordered for sem1, judging by the trend over the last few sems
[00:55:37] <arekinath> more lecturers are hearing about it and getting on board
[00:56:00] <arekinath> we're hiring another staff member to basically come and look after smartos-related queries and problems
[01:02:07] <killfill> wonderfull
[01:04:25] <arekinath> we do pretty well out of hiring our own students. :P
[01:09:22] <killfill> 'eating your own food' or something like that i guess..:P
[01:13:17] <Licenser> I'm off to bed take care guys :)
[01:16:03] *** FoxBuru has quit IRC
[02:06:17] *** jim80net has quit IRC
[02:08:37] *** FoxBuru has joined #project-FiFo
[02:11:59] <trentster> killfill: you think its worth fixing the broken "cloudview" graph thing in latest dev's ? or you think we should just retire it?
[02:56:12] <killfill> letme see why does this happend
[02:56:25] <killfill> trentster: what do you use fifo for?
[03:00:15] *** fxhp has joined #project-FiFo
[03:00:28] <fxhp> Hi guys, adding this channel to my auto join
[03:01:23] <killfill> hi fxhp, thats greate.. :)
[03:02:50] <fxhp> so, during testing I deleted my fifo "stable" and installed fifo "dev"
[03:03:26] <fxhp> I had a vm from the existing fifo install that I wanted to keep, so I did and it was detected, however the package was not detected
[03:03:46] <fxhp> in addition, the IP address was not detected as in use
[03:03:55] <trentster> killfill: I use it for all my compute stuff both privately and in the datacentre's
[03:03:55] *** noahmehl has joined #project-FiFo
[03:04:08] <fxhp> so the next host I attempted to bring up used the same ip address as the existing vm
[03:04:42] <killfill> trentster: like for clients? for the 'office'? im trying to put some example on my presentation..
[03:05:21] <killfill> fxhp: you could define in the new fifo a new iprange, with 1 ip less, and previous.. :)
[03:05:24] <trentster> killfill in the DC it manages all windows vm's, linux, zones etc. I have permissions setup so that technicians can login and manage certain infrastructure for certain clients etc.
[03:05:45] <trentster> The only thing I dont do yet is use it for customer facing interaction yet.
[03:05:46] <fxhp> Yeah, that was the work around I was going to do
[03:05:49] <killfill> fxhp: if you remember the package spec, you can created it, and 'resize' the vm to that.
[03:06:11] <trentster> There needs to be some dumb downed customer portal for that that hides 80% of the tabs shown in jingles.
[03:06:16] <fxhp> I recreated the package spec, I assume only the uuid of the spec is different
[03:06:53] <killfill> fxhp: yeah, and that info should be lost when you wiped the old fifo.. so need to reasign it.. anyway. it its the same, it wont do anything :)
[03:07:01] <fxhp> killfill - in this case I can delete all the vms and start over, I'm only playing with test vms
[03:07:13] <fxhp> but I do think that could get wierd
[03:07:18] <fxhp> (like if I was to do a vm restore)
[03:07:26] <fxhp> I wouldn't want to alter my iprange
[03:07:29] <killfill> trentster: but the busness there is to provide virtual machines to customers?
[03:07:38] <fxhp> or a vm migration from one hypervisor to another
[03:08:08] <trentster> fxhp: you can also do a fifo migration to a new zone and then install dev ontop of existing install.
[03:10:05] <fxhp> yeah
[03:10:06] <trentster> killfill: yeah, its a bit of a hybrid role, it manages all of our internal stuff and services etc, plus it also manages individual clients vm's and infrastructure. But since we look after those customers IT we manage their vm's on their behalf.
[03:11:03] <trentster> So its not really a self service cloud yet, which it really should be, like someone could sign up automatically and provision etc, billing stuff is auto handled.
[03:11:16] <trentster> which is where I would like it to be.
[03:11:30] <killfill> oh yeah, me too.. :)
[03:12:22] <fxhp> seems like this could be a priv roll
[03:17:12] <trentster> fxhp: ?
[03:20:05] <fxhp> hiding views in jinjles seems like it could be solved with user privleges
[03:20:19] <fxhp> killfill - resize to package fixed that bit
[03:20:37] <fxhp> going to see if I can "resize" the network too
[03:22:25] <fxhp> I think SmartOS + KVM is faster then Ubuntu _+ KVM
[03:22:44] <fxhp> reboots happen extreamly fast
[03:22:50] <fxhp> (for the guests)
[03:24:52] <trentster> fxhp: yeah the joyent linux datasets have smartos optimised kernels etc, they are pretty darn fast. Altho stay away from kvm if you can use zones. zones rock!
[03:26:47] <killfill> fxhp: the other day, i have to use xen on production.. i felt the other side.. its awful slow.. :P
[03:32:43] <FoxBuru> Is there a reason why i don't see datasets when doing fifoadm datasets list if i already have one dataset on the gz (the dataset that i used to create the fifo zone, of course ^^)?
[03:33:34] <fxhp> I used xen for a long while before kvm on RHEL/Ubuntu, KVM on smartos is magical
[03:34:03] <fxhp> FoxBuru - did you restart chunter ?
[03:34:04] <killfill> FoxBuru: fifo uses a different way to store datasets than smartos.
[03:34:17] <fxhp> ignore me
[03:35:29] <FoxBuru> @fxhp: Yep, i restarted it some minutes ago.
[03:35:59] <FoxBuru> @killfill: So, datasets downloaded previously are not detected by fifo?
[03:36:13] <fxhp> killfill - FoxBuru - fifo stores datasets in its own zone, and spreads the datasets over 3 sibling fifo hypervisors, else it store 3 copies on itself if it has no siblings ?
[03:36:14] <killfill> nope. you can still downlod them again tho...
[03:36:16] <trentster> FoxBuru: you picked up a slight bug there, I see "fifoadm datasets list" gives an error, yet "fifoadm datasets list -j " works
[03:36:56] <trentster> FoxBuru: you can log a ticket for it to be fixed, or I can do it if you prefer?
[03:37:14] <fxhp> ^ is my comment/question wrong
[03:37:18] <killfill> FoxBuru: was your dataset download via imgadm?
[03:37:33] <FoxBuru> killfill: yep, it was downloaded using imgadm on gz
[03:37:39] <killfill> fxhp: trying to understand it.. :P
[03:37:54] <FoxBuru> trentster: i don't see anything using "-j" on fifoadm =(
[03:37:55] <killfill> trentster: i donnt think hes dataset will be shown, right?
[03:38:45] <trentster> FoxBuru: you on dev?
[03:39:10] <FoxBuru> trentster: Yep, on yesterday dev
[03:39:11] <killfill> FoxBuru: think if you had 20 servers. you dont want to download 20 times via imgadm 1 dataset (or setup a proxy setup for that)... fifo just downloads them, and when a server needs it, it goes and places it where it need it. no imgadm involved. :)
[03:39:20] <trentster> killfill: yup it wont be shown.
[03:39:39] <trentster> FoxBuru: you will only see datasets within fifos internal dataset server with that command.
[03:40:10] <trentster> fifo handles its own datasets and provisions them to nodes when they are needed, think of it as a central imgadm repo for your entire cloud.
[03:40:11] <FoxBuru> killfill - trentster: oh right. ^^ I'll download a dataset from FiFO UI then
[03:41:51] <killfill> fxhp: it saves the dataset in 3 'partition' db thing. maybe you ment that.. :)
[03:42:01] <fxhp> yeah
[03:42:14] <fxhp> I am a complete newb when it comes to solaris, and zones
[03:43:57] <fxhp> From listening to people talk in here, I learned that fifo attempts to spread the datasets to availble fifos in the "cloud"
[03:44:11] <fxhp> But it does that separately then the gz
[03:44:33] <trentster> fxhp: yeah, its a little learning curve when you start, but its worthwhile. The benefits of zones are definately worth the innitial learning.
[03:44:42] <trentster> There are soem resources available.
[03:45:03] <trentster> Have you seen the linux to solaris cheatsheet thing yet?
[03:45:08] <fxhp> Nope
[03:46:15] <trentster> fxhp: http://wiki.joyent.com/wiki/display/www/Linux+to+SmartOs+cheatsheet
[03:46:39] <fxhp> thanks!
[03:47:15] <fxhp> trentster - a couple weeks ago I was attempting to create a dataset from a KVM snapshot
[03:47:22] <fxhp> and filed a bug
[03:47:31] <killfill> i remember i was so fool when i tried opensolaris when it came it.. i thoght 'this has no 'top' command.. i dont feel lerning every unix command again..' <-- /me ashamed.. :P
[03:47:49] <fxhp> prstat?
[03:47:51] <killfill> still lerning how to interpret prstat.. :P
[03:47:54] <killfill> :)
[03:47:56] <fxhp> (I learned that much)
[03:47:57] <fxhp> haha
[03:50:24] <fxhp> So the dataset generated from the snapshot happens REALLY fast (instantly?) could that be an issue?
[03:50:25] *** FoxBuru has quit IRC
[03:56:33] <trentster> fxhp: you talking about creating a dataset from an existing kvm vm? you still having issue with that?
[03:58:52] <fxhp> yeah, it doesn't work.
[03:59:22] <trentster> fxhp: let me check quick.. gonna create a ubuntu vm and see if it works, when I try make a dataset
[04:01:05] <fxhp> But I don't know if it is an issue with dataset creation or with the new vm generation
[04:01:31] <fxhp> I know the snapshot itself works, I tested that
[04:22:48] <trentster> fxhp: I think you are right dataset creation from kvm is broken!!
[04:23:20] <trentster> I think the kvm stuff gets a bit neglected. Will have to log a ticket and wait for Licenser to have a look at it and fix it.
[04:23:45] <trentster> fxhp: have you tried the manual method?
[04:24:00] <trentster> fifoadm datasets import ?
[04:24:32] <fxhp> manual method?
[04:25:16] <trentster> here is the reason it fails
[04:25:18] <trentster> [debug] <0.28128.8>@chunter_vmadm:wait_for_text:210 [vmadm] missing .model option for disk: {"boot":true,"image_size":0,"image_uuid":"2cf37571-6e1e-412e-b306-df41dd5bcaca","media":"disk"}
[04:25:35] <fxhp> Looks like the datasets created from snapshot are missing:
[04:25:36] <fxhp> "nic_driver": "virtio",
[04:25:52] <fxhp> "disk_driver": "virtio"
[04:26:34] <fxhp> (also ran into that bug with fifoadm dataset list , needed -j )
[04:27:04] <trentster> manual method = " fifoadm datasets import <manifest> <data file> "
[04:27:57] <fxhp> "fifoadm networks" gives me wierd usage message ... fifoadm networks
[04:27:58] <fxhp> Usage: sniffle-admin { join | leave | reip | ringready | remove |
[04:28:46] <fxhp> I'm able to use kvm from a dataset downloaded from joyant, so I think the manual method works
[04:28:55] <trentster> fxhp: to be honest the fifo dataset stuff is a little bit wonky and erratic. What I do, which may or may not work for you, is setup my own local dataset server using the "dsapi" dataset which is what datasets.at uses. I then push all my datasets into there, and sync any of the joyent datasets to it.
[04:29:12] <trentster> I then change my fifo to sue my local dsapi server rather than datasets.at
[04:29:49] <trentster> the cool thing about this is that you can configure local dsapi server to automatically sync and pull all images from datasets.at and you always have all datasets available locally to use.
[04:30:26] <trentster> This way if you ever blow away fifo, you can easily install datasets again from own dsapi server when you setup new fifo.
[04:30:44] <fxhp> I see.
[04:32:07] <trentster> Also MerlinDMC will be releasing new dsapi soon which allows easier uploads and pulling out individual datasets as well. Plus dsapi stores all datasets locally in a directory so they are not roped up in a internal database that you dont have control over.
[04:32:19] <trentster> if all that makes sense ? :-P
[04:33:15] <trentster> http://datasets.at/ui/#!/configure/e96e0516-76c1-5ffb-53ba-655a729206e4
[04:35:27] <fxhp> Make sense, I'm not anywhere near ready to put this into production, so I have fun tinkering with it on my free time
[04:35:38] <trentster> ;-)
[04:49:45] <fxhp> trenster - do you want to update jira with your findings
[04:49:47] <fxhp> ?
[04:50:09] <fxhp> http://jira.project-fifo.net/browse/FIFO-136
[04:50:58] <fxhp> oh you already had a ticket
[04:51:13] <fxhp> (just saw yours)
[04:51:32] <fxhp> Mine was from a few weeks back, I'll link to your from my older one
[04:55:32] <trentster> fxhp: cool, I also added a ticket for "fifoadm hypervisors list" not working
[04:57:57] <fxhp> "fifoadm networks" was the one I saw
[04:58:41] <trentster> http://jira.project-fifo.net/browse/FIFO-140
[04:59:17] <fxhp> yeah that one too
[05:10:57] *** noahmehl has quit IRC
[05:11:19] *** noahmehl has joined #project-FiFo
[07:43:08] <MerlinDMC> morning
[07:44:19] <MerlinDMC> trentster, I did put the "export" link onto datasets.at UI ... I thought if it kills the server it does ... if not I'm fine
[07:47:21] <MerlinDMC> trentster, there: http://glui.me/?i=y950ve0750hgqa4/2013-10-24_at_07.46.png/
[07:53:55] *** jim80net has joined #project-FiFo
[08:37:02] <trentster> MerlinDMC: yeah, I saw it earlier today, when I was on the site, pretty cool!
[08:59:11] *** noahmehl has quit IRC
[10:08:36] *** mgls has joined #project-FiFo
[11:07:45] <Licenser> Morning everyoen
[11:13:29] <trentster> howdy Licenser :^)
[11:21:30] <Licenser> how are things with you? :)
[11:23:53] <trentster> good mate, and with you?
[11:24:33] <trentster> we need to have a hard chat about the dataset server one of these days, I am not exactly its biggest fan right now :-P
[11:27:01] <MerlinDMC> does someone have a M$ subscription and can check their server speed for me? ... i'm fetching isos with 30K/s (which is if you need about 15G really slow)
[11:30:37] <Licenser> trentster you're not the biggest fan of anything ever :P
[11:31:54] *** jim80net has quit IRC
[11:32:15] *** jim80net has joined #project-FiFo
[11:34:57] <Licenser> trentster side what better sulution would you suggest?
[11:34:58] <trentster> Licenser: only you I am a biggest fan of, all else I judge with a critical eye ;-)
[12:02:41] <MerlinDMC> got that.
[12:03:04] <MerlinDMC> but I can handle judgement.
[12:07:26] <trentster> MerlinDMC: ;-)
[12:07:49] <trentster> Licenser: lets replace dataset server with dsapi server ;-)
[12:08:30] <Licenser> And make people install datasets manually on each node?
[12:08:55] <Licenser> And upload datasets manually?
[12:09:20] <Licenser> That sounds like a step backwards not forward
[12:10:00] <trentster> nope, nah, dsapi will pull datasets from datasets.at and store them locally, then when provisioning a dataset fifo will pull it from dsapi and push to node if if does not have it.
[12:10:39] <trentster> only thing that will be missing is " redundecy " and 3x duplication thing.
[12:10:43] <Licenser> but that how it's working right now?
[12:10:57] <trentster> Licenser: btw I am half kidding :-P
[12:11:14] <trentster> not sure what the solution is.
[12:11:40] <trentster> but I would like to put a bullet to the brain of the current dataset server, its really not nice…:-P
[12:12:14] <Licenser> trentster oddly enough I can't replicate the problem with timeouts I tried it today and it's just working w/o any magic for me
[12:12:59] <trentster> its like the black hole of calcutta, anything you put in there is there foreever, and consumes space till eternity even if you delete it. from the UI
[12:13:12] <trentster> you have created a singularity :-P
[12:18:07] <trentster> ymmv
[12:25:04] <Licenser> the kvm dataset ticket is updated with a suggested solution just waiting for someone to try it
[12:25:28] <Licenser> and no fifoadm will not allow to import export datasets, it will be a API call nice and clean
[13:15:05] <trentster> Licenser: thanks will test it out now and revert.
[13:18:48] *** arekinath_ has joined #project-FiFo
[13:19:11] <trentster> Licenser: but I dont think thats the only problem, from what I can tell when you try and turn the kvm vm into a dataset in fifo, it does not get created correctly and added into the dataset server. On my systems it gets instantly added, and if you click on the dataset its shows "zero" image size.
[13:19:28] *** arekinath has quit IRC
[13:24:58] <Licenser> trentster it does not copy the right image to start with which propably is the big issue
[13:27:22] <trentster> I just tried it, did not make a difference dataset created immediately, and still shows this.
[13:27:49] <trentster> Licenser: https://www.monosnap.com/image/nUv73VDzHHPztCVol1qsmyzi8
[13:28:26] <trentster> and should look like this:
[13:28:27] <trentster> https://www.monosnap.com/image/JvMfQ4yJXvd4tkLdEyGdI00L2
[13:36:05] *** arekinath_ is now known as arekinath
[13:36:10] *** arekinath has quit IRC
[13:36:10] *** arekinath has joined #project-FiFo
[13:51:21] <killfill> hi
[13:53:20] <trentster> hi killfill
[13:53:24] <trentster> how ya doing?
[14:01:11] <killfill> dont feell like crap today (jet) to greate!.. :)
[14:01:14] <killfill> so
[14:45:19] <killfill> leofs?
[14:45:40] <killfill> we will integrate with them?.. :-D~
[14:45:54] *** jim80net has quit IRC
[14:46:01] <Licenser> killfill that's the plan
[14:46:13] <killfill> what about riak-cs?
[14:48:25] <Licenser> leofs looks nicer especially form a performance perspective
[14:48:43] <killfill> cool.. (i havent test neithr both.. :P)
[14:49:02] <killfill> just used s3 for the first time last week for backups.. :P
[14:50:11] <killfill> oh and we could save accounting in there.. :) and allow some kind of billing
[14:51:55] <Licenser> that not make too much sense it's a object store not a databse
[14:52:46] * killfill still dont looses faith on accounting. :P
[14:53:31] <Licenser> heh
[15:10:54] *** jim80net has joined #project-FiFo
[15:39:50] <Licenser> killfill I think you can give up hope on accounting
[15:55:10] <killfill> greate!
[15:56:36] <killfill> how do you record videos on osx? from a chrome window
[16:10:19] *** ira has joined #project-FiFo
[16:24:11] <killfill> ok im almost done.. :)
[16:25:44] *** noahmehl has joined #project-FiFo
[16:50:36] *** noahmehl has quit IRC
[17:20:43] *** noahmehl has joined #project-FiFo
[18:05:57] *** FoxBuru has joined #project-FiFo
[18:30:12] *** noahmehl has quit IRC
[18:36:05] *** noahmehl has joined #project-FiFo
[19:23:23] <FoxBuru> hmmm, it seems like FiFo UI can't see Etherstubs =(
[19:31:13] <killfill> hm.. vm resizing doesnt work now.. :S
[19:31:30] <killfill> damn.. that was my killer feature! :P
[20:01:57] <MerlinDMC> heyho
[20:09:24] <FoxBuru> Hey ^^
[20:24:21] *** noahmehl has quit IRC
[20:28:18] *** noahmehl has joined #project-FiFo
[20:51:00] *** noahmehl has quit IRC
[21:47:17] *** pringlescan has joined #project-FiFo
[21:47:30] <pringlescan> has anyone enabled jumbo frames yet?
[22:13:24] *** mgls_ has joined #project-FiFo
[22:16:54] *** mgls has quit IRC
[22:16:55] *** mgls_ is now known as mgls
[22:30:08] <fxhp> What may I stick into User script ?
[22:30:27] <fxhp> A bash command, or a path to a script?
[22:34:03] <MerlinDMC> a bash script
[22:35:20] <fxhp> So a path, is the path relative to the GZ or Fifo zone?
[22:36:21] <MerlinDMC> what "a path"? ... you put the complete script in there and it gets executed inside the new zone
[22:37:54] <MerlinDMC> fxhp, https://gist.github.com/MerlinDMC/8fd375ad454c303b4380
[22:39:07] *** FoxBuru has quit IRC
[22:42:12] <fxhp> Oh so its not a input to a path to a "user script" its a text area for placing bash
[22:43:52] <fxhp> I would have expected to see a multi-line textarea input instead of a single-line input, that is what sent me off in the wrong direction.
[22:45:50] *** pringlescan has quit IRC
[22:47:26] <MerlinDMC> fxhp, yeah ... should probably be a textarea
[22:48:19] <MerlinDMC> don't know how that is handled internally ... maybe it's just a simple html fix in the jingles UI
[22:49:02] <fxhp> So now I'm afraid to use multilines, so I used semi-colons
[22:49:04] <fxhp> haha
[22:49:26] <fxhp> About to test bootstrapping salt on my ubuntu guest
[22:51:43] <MerlinDMC> if you do smth like that and need a config file for the bootstrap (like the json file for chef-solo) maybe user-data is also something to take a look at
[22:51:58] <MerlinDMC> I just can't recall where the data file is placed inside the zone
[22:53:09] <MerlinDMC> should be /var/db/mdata-user-data ... you you have some place for config data as well
[22:55:00] <MerlinDMC> almost 23:00 ... need to get up at 05:30 ... so i'll better get some sleep now ;)
[23:10:01] *** mgls has quit IRC
[23:14:29] <fxhp> night
[23:53:21] *** pringlescan has joined #project-FiFo

   October 24, 2013  
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | >