[00:23] <DinVitamin> I installed Ubuntu Server, which has no sound or video drivers. I installed the video drivers and x-server, etc. so that I could see everything, but how do I install the sound drivers?
[00:25] <cloakable> They're built into the kernel
[02:22] <Delerium_> Hi guys, I need your input / recommandation .. my dedicated server is up!  but they put qmail to manage emails ... I think qmail is not even supported anymore and the developpement is stop by now ... I thought about installing Zimbra instead.  Any other solutions you could think of?
[02:34] <AtomicSpark> So, the "updates avaiable" part of my MOTD hasn't been correct for a few months. How do I fix it?
[02:34] <AtomicSpark> Its stuck on saying 11 packages. :\
[02:38] <julian_c> Using 10.04 LTS?
[02:38] <AtomicSpark> Yes.
[02:38] <AtomicSpark> Is it a bug?
[02:38] <AtomicSpark> It happened after there were updates and I updated them, but the counter never reset.
[02:38] <julian_c> Kind of. When you log in, it's actually showing two MOTDs.
[02:39] <AtomicSpark> Haha! What the crap. So it is.
[02:39] <julian_c> If you delete the file </etc/motd.tail>, then it will only show the current MOTD. (The stale MOTD is in that file.)
[02:40] <AtomicSpark> The default terminal size wasn't showing the other one. I maxed the window after you said that. Sillyness.
[02:40] <AtomicSpark> I shall do that then.
[02:41] <AtomicSpark> Seems to have fixed it. Thank you.
[02:44] <julian_c> You're most welcome.
[02:46] <AtomicSpark> :)
[03:28] <Snugger_> Having trouble installing ubuntu server could someone help me?
[03:29] <Snugger_> Hello?
[03:30] <Delerium_> What's the issue first?
[03:30] <Snugger_> i get a corrupt file error while it is installing the base system
[03:31] <Delerium_> maybe a corrupted download / or wrong burning ... did you check the md5su?
[03:31] <Delerium_> mdsum
[03:31] <Snugger_> no i havn't, but i've tried burning and burning numerous times, still the same error
[03:32] <Delerium_> then might be your download ISO
[03:32] <Snugger_> should i try a mirror link?
[03:33] <Delerium_> check the md5sum, that might save you downloading everything again (and save bandwith)
[03:35] <Snugger_> I'm going to use my labtop to burn the cd this time, could you inform me how to check the md5sum though incase i have the same problem on my labtop?
[03:38] <Delerium_> do you have a Linux installation available?
[03:38] <Delerium_> command would be: md5sum FILENAME
[03:39] <Delerium_> the right md5sum is on Ubuntu download site if I'm not mistaken
[03:39] <Snugger_> I'm burning the cd via Windows Vista & Windows 7
[03:40] <Delerium_> Snugger_, not sure if this kind of tools is available on Windows...
[03:41] <Snugger_> I apoligize, i'm new to ubuntu entirely, does ubuntu server require ubuntu OS installed beforehand?
[03:42] <Delerium_> nope
[03:44] <Snugger> @Delerium_ i got disconnected from the chat, what did you write?
[03:45] <Delerium_> Snugger, Nope. you don't need a Ubuntu installation beforehand
[03:45] <Delerium_> You can always try to install Ubuntu in the VM to avoid burning CDs over and over
[03:46] <Snugger> I'm not sure how to utilize the VM without an active OS on a computer
[03:47] <Delerium_> Ho okay... you only one computer running Windows and you wish to install ubuntu as a dual boot?
[03:48] <jmarsden> Snugger: https://help.ubuntu.com/community/HowToMD5SUM  has info for checking md5sums from different OSes, including windows.
[03:48] <Delerium_> nice to know jmarsden , thanks
[03:49] <jmarsden> Delerium_: google for ubuntu howto md5sum    and you find it :) :)
[03:49] <Snugger> thanks jmarsden, delerium, i am transforming an old pc into a game server trying to use ubuntu server
[03:49] <Delerium_> jmarsden, well.. I was just trying to help Snugger ;)  I'm working at the same time so no time to google over to find answer
[03:51] <Delerium_> check the link jmarsden sent ...
[03:51] <Delerium_> you will be able to see if the .iso file is ok or not
[03:52] <Snugger> i will, thanks for being patient with me
[03:52] <Delerium_> no worries, we've all been there ... and I must say I'm not a Linux expert at all... this learning, when I have time
[03:55] <Snugger> i can tell english is not you're main language, may i ask which is?
[03:56] <Delerium_> haha! You're right... French ... I'm from Montreal
[03:57] <TVision> Hi. I'm using ircd-hybrid and irssi. /list isn't showing me +s secret channels even though I /oper. What might I be doing wrong?
[03:59] <Snugger> Ah tu parles francais?!
[04:00] <Snugger> tu aurais du dire quelque chose, je parle le francais. s'il est plus facile pour toi a epliquer l'instructions en francais, vas y =ED
[04:01] <Delerium_> Snugger: :) Tu es du Quebec aussi?
[04:01] <Snugger> non, de chicago, mais j'etudie le francais
[04:01] <Delerium_> Cool!
[04:02] <Snugger> j'ai decider a toi de parler desormais en francais?
[04:04] <Delerium_> Snugger, but I do really love to improve my english .. since I'm working 90% with english people... but you can talk me in french, that's fine ;)
[04:04] <Delerium_> I'll give you a note after ;)
[04:04] <jmarsden> On doit parler en Anglais ici, je pense.  Peut-etre vous pouvais parler en Francais en #ubuntu-fr ? :)
[04:05] <Delerium_> yup
[04:05] <Snugger> woops okay english it is =D
[04:05] <Delerium_> That was kind of not predictable ;)
[04:06] <Delerium_> Gosshh... I should really update my upload speed...
[04:06] <Delerium_> and say bye-bye to my unlimited plan
[04:07] <Snugger> it will save money, what are you uploading?
[04:10] <Delerium_> Snugger,  Zimbra ...600 megs.. I should have download it directly from the server... instead of uploading from my laptop.. oh well
[04:11] <jmarsden> Snugger: Incidentally, try /msg nickserv info Delerium_    to see a bit of info about his login, that might have given you a clue where he is from :)
[04:12] <Delerium_> ;)
[04:13] <bencc> what is the /etc/nginx/conf.d folder for?
[04:13] <jmarsden> bencc: Local customization of nginx configuration by dropping files in there...
[04:14] <bencc> jmarsden: but it only allow configuring the http context
[04:15] <jmarsden> Don't the files there get included into the nginx configuration?  I'm no nginx expert, but that is usually how confid/ directories work...
[04:15] <jmarsden> * conf.d/ directories work...
[04:16] <bencc> jmarsden: yes, it's included but it only let you configure part of the possible settings
[04:17] <jmarsden> bencc: Well, per site stuff goes in sites-enabled/  ... what else do you need to configure in nginx.conf ?
[04:18] <bencc> jmarsden: worker_processes, worker_connections, gzip, logs...
[04:18] <bencc> jmarsden: gzip and logs can be configured in conf.d but worker_processes and worker_connections can't
[04:18] <Snugger> woohoo, i burned the cd from my labtop and it is finally installing the base system
[04:18] <Delerium_> Snugger, cool!
[04:18] <Delerium_> Snugger, defective CD / Burner ?
[04:19] <jmarsden> bencc: OK, so edit them in /etc/nginx.conf itself, they are already in there... it doesn't make sense to let each app configure nginx-wide options, the apps would all fight each other... right?
[04:19] <bencc> jmarsden: yes. ok
[04:19] <Snugger> Delerium_ perhaps, i should look into it
[04:19] <bencc> jmarsden: I'll just use etckeeper to keep track of changes
[04:20] <Delerium_> Snugger, well.. at least, you will be up and running soon
[04:21] <Snugger> this is the first server i've ever made, are they hard to maintain and setup?
[04:22] <Delerium_> Snugger, Depends on what you want to install... but upgrading / installing security fixes is pretty easy
[04:24] <Snugger> Hey delerium how do you make it so the messages are directed towards me? and i want to install a minecraft server
[04:26] <Delerium_> Snugger, Just type the first letter of my nick and then hit Tab.
[04:26] <Delerium_> Snugger, as for minecraft ... I really don't know
[04:26] <Snugger> dwhat do your server(s) manage
[04:27] <Snugger> Delerium_: what do you're server(S) manage?
[04:28] <Delerium_> Snugger, Just bought a dedicated server for a personnal project ;) ... but at work, I'm managing website for a bank (WebSphere, IBM https server, iPlanet. MQ, blabla)
[04:33] <Snugger> Delerium_: Cool, i'm a bit baffled as to which server software i should install
[04:33] <Delerium_> Snugger, depends on requirement! ;)
[04:34] <Delerium_> Snugger, need a HTTP Server?  Apache.  Need a Database?  MySql will do the job. ;)
[04:34] <Snugger> Delerium_: No website or database, what is a Tomcat Java server?
[04:35] <Delerium_> Snugger, Tomcat is an applicatiion server for Java.  It will let you run java Servlet/JSP.  But no EJB.
[04:36] <Snugger> Delerium_: Could you run java applet on a samba file serveR?
[04:37] <Delerium_> Snugger, yup, since java applet are run on the browser side, not on the server side
[04:38] <Snugger> Delerium_: aAlso i will be able to install additional server software on my server if needed?
[04:38] <Delerium_> Snugger, yep, the ubuntu repository contains alot of software to be installed (open source, of course)
[04:39] <Delerium_> Snugger, Like I said, it really depends on what you need and what you want to do
[04:40] <Snugger> Delerium_: i know what i want to do, i just don't know which softwares i need to do it. I should study software and hardware, but french is my passion also
[04:40] <Delerium_> Snugger, I think you wanted to install a minecraft server at first?
[04:41] <Snugger> Delerium_:  yes that's correct
[04:42] <Delerium_> lol... just check the minecraft website... seems weird ;)
[04:44] <Snugger> Delerium_: mdr i'll google it, what are you studying or what did you use to study?
[04:47] <Delerium_> Snugger, a DEC (Collegial) in computer science... But I'm 34, so it was a bit easier to get a good job in the past
[04:48] <Snugger> Delerium_: ahh i take computer science at my high school, i'm not that good at it to be honest. you're fluent in javascript?
[04:49] <Delerium_> Snugger, Nope ... I'm no developper at all!  But I do have some developmenet basic which I need to do my job
[04:50] <Delerium_> Snugger, it's weird since, for example. WebSphere is a big container, and you then install the code on it.  So I support WebSphere, but not the code itself.  But I do have good java basis...
[04:51] <jmarsden> Snugger: Why did you install Ubuntu server before knowing what you want to run on it?  That seems odd... it is easiest to pick a server OS that has existing packages of the software application(s) that you are wanting to run :)
[04:52] <Delerium_> I guess he's playing around and learning ;)
[04:52] <Snugger> exactly, and i had no idea where to find those server OS's. I guess i like learnin the hard way
[04:52] <Delerium_> I think we all did
[04:52] <Delerium_> Well.. for me ;)
[04:53]  * jmarsden got himself a B.Sc in Computing and Information Systems in 1983... :)
[04:53] <Snugger> You did now? alot of things have changed since then
[04:53] <Delerium_> jmarsden, errr... I was 14 then!!! I thought I was old ;)
[04:54] <Delerium_> err.. no ... I was 8!! Goossshh
[04:54] <Delerium_> Born in 76
[04:54] <jmarsden> Things change, but computing itself does not, and the basic principles do not, so it has been (relatively) easy to learn new things with that bacis of knowledge to "hang things on"
[04:55] <Delerium_> jmarsden, yep, one of my teacher always told us :" I'm learning you how to learn " ... I find this pretty right
[04:56] <jmarsden> Right.  I just mentioned it to suggest that the idea that "we all" learn by playing with modern PCs and free OSes does not fit my approach very well, Linux did not exist in 1983, I don't think even Minix existed then :)
[04:57] <Delerium_> Indeed!  I guess you were mostly working on DEC / Mainframe/ Cobol /Fortran and things like that
[04:57] <Snugger> i really enjoy learning about computers and technology and i am somewhat computer savy, i'm patient with computers. I also like business and french though
[04:57] <fluvvell> jmarsden, xenix existed, I connected terminals for it.
[04:57] <jmarsden> Right, but it wasn't free :)
[04:57] <fluvvell> jmarsden, quite right.
[04:58] <Proz01d> i'm searching for domain registrar any recommendations on cheap ones? (also do most provide unlimited subdomains)?
[04:58] <Delerium_> Good old days ... Sometime. I'm missing my little Tandy TRS-80 with it's tape drive (?!?!) go figure
[04:59] <Delerium_> Proz01d, Home server without a static IP ?
[04:59] <jmarsden> Proz01d: Registrars are not concerned with subdomains at all; how you configure your DNS servers determines those.
[05:00] <Proz01d> i'm just using google dns servers.
[05:00] <jmarsden> Proz01d: Not if you register a domain, you aren't :)
[05:00] <Proz01d> Delerium: I do have a server with a static ip
[05:00] <jmarsden> Unless you can persuade Google to provide authoritative DNS for your new domain... which I somewhat doubt?
[05:01] <Delerium_> Proz01d, maybe your server provider can act as a registrar
[05:01] <Proz01d> i'd rather keep my server and registrar seperate in case i end up finding something better
[05:02] <Delerium_> not sure... I always use the same server provider / registrar in the past
[05:02] <Proz01d> whos is your provider?
[05:02] <Delerium_> iWeb.com, located in Montreal
[05:03] <Delerium_> I also use aplus.net in the past (I think they were bought by another cie meanwhile)
[05:06] <jmarsden> Proz01d: You can look at https://www.nearlyfreespeech.net for a fiercely independent place that will do domain registrations for you.  Don't cosider this a recomemndation, check them out for yourself and decide if you like them.
[05:08] <jmarsden> Proz01d: They will also do DNS for you if you need it, but if you have a server on a static IP and it is reliable, you can run primary DNS there, and find a free secondary somewhere to save $$ if you know what you are doing.
[05:17] <Snugger> quick question, so i am trying to install capability for my netgear wireless usb adapter onto ubuntu, it recognizes it, but it is not connecting to the internet
[05:30] <Proz01d> jmarsden: the server is pretty small so i want ot limit the number of things i run there.... I don't really understand dns at all so i will need to do some researc there. Regardless it still needs to be registered no?
[05:30] <Delerium_> Snugger, Sorry, But I can't help much on this ... I never use linux with a wireless usb
[05:31] <jmarsden> Correct.  Easy way is to use somewhere like nearlyfreespeech to both register it and then run DNS for it there, they have a web interface so you can add/edit DNS stuff on their servers for your domain... nice and easy.  But you will pay some small amount for that service.
[05:32] <jmarsden> Snugger: wireless USB on the Ubuntu Server?  That's not brilliantly documented, but I have something twb gave out here last time this was asked about, let me find it...
[05:33] <jmarsden> http://pastebin.com/Uu0HMbcP
[05:34] <Snugger> appreciate it
[05:34] <jmarsden> Snugger: You need to edit the two files /etc/network/interfaces and the wpa_supplicant one...
[05:35] <jmarsden> So that plus reading    man interfaces   and   man wpa_supplicant.conf     should be a start.
[05:37] <Snugger> okay i'm completely clueless as to what you just said.. should i type in what is on this page and that should fix the problem?
[05:38] <Delerium_> Gotta go guys ... Need some sleep... be back tomorrow, take care
[05:42] <Snugger_> jmarsden: so on the pastbin link you sent me, type that into the server console?
[05:43] <jmarsden> Snugger_: No, edit those config files... I can't teach you the basics of managing a server right now, I need to do other things, sorry.
[05:44] <Snugger> jmarsden: no problem, i teach myself =D
[05:45] <jmarsden> The commands to type would be    man interfaces     and    man wpa_supplicant.conf    to learn about those files, and then you can use sudoedit to edit them to look like the ones it my pastebin.
[05:49] <Snugger> i typed in sudo vi /etc/network/interfaces would that work also?
[06:23] <jmarsden> Snugger: Probably.
[08:03] <Syria> Hi, can I install ubuntu server 10.4.2 using a usb flash disc?
[08:05] <jmarsden> Syria: Probably.  One set of instructions is at http://www.pendrivelinux.com/run-ubuntu-server-edition-installer-from-usb/
[08:05] <Syria> jmarsden:  Thanks, i will read that.
[08:05] <jmarsden> You're welcome.
[09:00] <speakman> lambda_x: Hi! Just wanted to tell you that when the new disk in my raid1 finally was synchronized, it automatically switched from spare into an "ordinary" disk in the mirror. It's now sda1[0] sdb1[1] instead of sda1[0] sdb1[2].
[09:01] <speakman> lambda_x: even mdadm -E /dev/sda1 tells only two disks now.
[10:01] <Syria> Hello, I want to install Zimbra mail server on my ubuntu server 10.4.2 I have two hard discs each one is 1tb please tell me what is the best way for partitioning.
[10:06] <thefish> hi, i have a lucid server kvm image which I just found in "loading, please wait" - it had been like this for at least a few days. I've tried to use a rescue image, and this is complaining "no partitions found", though fdisk -l shows 3 partitions (/dev/vda1 as root) - I can chroot to this and all the files seem sane - any ideas what to look at?
[10:06] <thefish> ^ the fdisk stuff in the busybox shell that the installer offers
[10:07] <thefish> :)
[10:10] <Syria> Are there any useful docs for partitioning?
[10:18] <RoyK> Syria: the default will work well in most cases
[10:19] <Syria> RoyK:  please tell me what do you mean by default?
[10:19]  * RoyK is off for an MPI/parallel processing workshop
[10:19] <Syria> RoyK:  This will give the home folder a huge space right? which is useless.
[10:22] <lambda_x> speakman: it was synchronizing? well... in that case - everything is now solved :)
[10:23] <lambda_x> speakman: you just had a bit nervous weekend ;)
[10:46] <Syria> Hello, I want to install Zimbra mail server on my ubuntu server 10.4.2 I have two hard discs each one is 1tb please tell me what is the best way for partitioning.	
[11:17] <fastveg> kiall: you around?
[11:25] <Daviey> Syria: To start, i would use mdad softriad across the two.
[11:41] <Syria> Daviey:  I don't know what is that :( mdad
[11:41] <Daviey> Syria, Software Raid.
[11:42] <Daviey> Syria, https://help.ubuntu.com/community/Installation/SoftwareRAID
[12:31] <speakman> lambda_x: Solved indeed. I guess I wasn't even very nervous... :D
[14:02] <andygraybeal> so this is completely scary.  i have had logrotate cough for a while now.  i see the problem exists in a specific file with -d ... i go to list the folder with the issue.. and it won't come up :(
[14:02] <andygraybeal> i do ls -al on it.. and it just hangs
[14:02] <andygraybeal> what should i do?
[14:03] <andygraybeal> i hit ctrl-c to get out of it.. but the prompt never comes back, and the (scsi) HD's are blinking lots.
[14:03] <andygraybeal> it looks like i might have some bad data
[14:04] <andygraybeal> what should i do next?>
[14:23] <Guest21700> Hello, I have a server that restarted last night and I can't find why. I searched through the various logs and found nothing. I have eliminated electrical failure as the source of the problem. I found nothing in syslog, dmesg or kern.log to justify the reboot. Where else might I look to find the reason for this reboot ?
[14:42]  * SpamapS stretches
[14:43] <RoAkSoAx> SpamapS: morning... up early again huh?
[14:43] <SpamapS> RoAkSoAx: yeah been up for > 1 hour already
[14:44] <andygraybeal> can anyone help, i can't do a ls -al on a folder, or a rm -rf ... i am sudo -i'd in ... i think there might be some data problem.  i don't know what to do.
[14:44] <RoAkSoAx> SpamapS: heh... I;ve been up for a couple hours already.. dunno why but I'm waking up earlier than always lol :)
[14:44] <SpamapS> andygraybeal: can you expand on "can't" .. like.. what is the error message?
[14:44] <SpamapS> RoAkSoAx: you're *SUPPOSED* to be up for a couple hours now though.
[14:45] <andygraybeal> SpamapS, it just kinda hangs there, and i can't ctrl-c out of it - i kill it after awhile with htop.
[14:45] <andygraybeal> it doens't give me a listing or anything
[14:45] <SpamapS> andygraybeal: try 'dmesg' .. any errors?
[14:45] <RoAkSoAx> SpamapS: I officially start at 9 so I usually wake up around 8.15/8.30. Been waking up at 7
[14:45] <SpamapS> Oh to be young
[14:46] <andygraybeal> SpamapS, i don't think there is anything in dmesg
[14:47] <andygraybeal> yea, nothing in dmesg - i'm doing a tail -f /var/log/dmesg
[14:48] <andygraybeal> and then i did a ls -al while tailing.. and nothing showed up
[14:48] <SpamapS> no
[14:48] <SpamapS> dmesg
[14:48] <SpamapS> not /var/log/dmesg
[14:48] <SpamapS> /var/log/dmesg is a snapshot from when the system booted
[14:49] <andygraybeal> aah okay
[14:49] <SpamapS> the command, dmesg, shows you the latest contents of the kernel's output buffer
[14:49] <andygraybeal> yea, i don't see anything in dmesg either... i mean that makes sense to me
[14:49] <andygraybeal> tehre's some mtrr: type mismatch whichs sounds very scary
[14:49] <SpamapS> ok, and anything in /var/log/syslog ? no errors reported?
[14:49] <andygraybeal> nothing in syslog
[14:50] <SpamapS> Bleh.. ok.. were it my system, I'd scheduled downtime, put it in single user mode, and run forced fsck's and maybe even check for bad blocks.
[14:50] <andygraybeal> okay, thank you
[14:51] <SpamapS> andygraybeal: if your drives / controllers have support for SMART, might be good to do that as well.
[14:51] <andygraybeal> mdadm says all is good as far as the raid goes
[14:51] <SpamapS> hanging indefinitely is generally not a good sign
[14:51] <andygraybeal> i dont' know if that matters
[14:51] <SpamapS> mdadm can do very basic checks..
[14:51] <SpamapS> but it doesn't go through and make sure all the blocks are returning what they should
[14:52] <andygraybeal> what can i expect from fsck .. ultimately?
[14:52] <SpamapS> well fsck is just going to make sure its not a filesystem issue
[14:52] <SpamapS> btw you'll want to pass -f to fsck so it forces a full check.
[14:53] <andygraybeal> okay thank you for that
[14:53] <SpamapS> andygraybeal: I'm curious, what kind of drives, and what kind of RAID?
[14:53] <andygraybeal> raid 10... softraid... and i honestly don't remember the drive types.. it's been maybe 2 years.. but they are seagate
[14:54] <andygraybeal> is there a way to get the type of drives out of the system?
[15:00] <RoAkSoAx> zul: ping
[15:02] <SpamapS> andygraybeal: yeah, hdparm /dev/XX should give some info..
[15:10] <andygraybeal> cool thank you SpamapS
[15:11] <andygraybeal> i scheduled downtime tomorrow morning at 6am
[15:11] <SpamapS> andygraybeal: good luck.
[15:11] <andygraybeal> yea, if things go bad..... how bad could it get?
[15:11] <SpamapS> andygraybeal: got backups? ;)
[15:12] <andygraybeal> it's a raid 10 ... yea, i have backups
[15:12] <SpamapS> just to make sure I understand you right.. raid 10 is NOT your backup solution, right?
[15:12] <andygraybeal> aah .. it came back with 'bad system call' !!
[15:13] <SpamapS> I'd check dmesg again.. might be some kernel errors now
[15:13] <andygraybeal> ok
[15:13] <SpamapS> but more to the point..
[15:13] <SpamapS> when you say you have backups.. you have snapshots of the data stored somewhere other than on the RAID, right?
[15:13] <andygraybeal> dmesg doesnt' have anything that i can see
[15:13] <andygraybeal> yea, it's all on tape
[15:13] <SpamapS> Ok.. some people don't understand the difference. Just making sure. :)
[15:14] <andygraybeal> yea, no problem, thank you
[15:14] <andygraybeal> for being persistent and thorough :)
[15:15] <andygraybeal> actually my backups have had a hard time.... every day this month since the 9th.. i think it is because of this problem
[15:15] <andygraybeal> so i have data, but from the 9th.. which sort of sucks but.. oh well
[15:19] <SpamapS> andygraybeal: if its just one directory, I'd recommend backing up everything else manually before doing your fsck
[15:20] <andygraybeal> cool, i'm configuring bacula to not touch that directory right now
[15:20] <andygraybeal> ;)
[15:20] <andygraybeal> just as you said that i saved :)
[15:20] <andygraybeal> now to restart the service
[15:20] <SpamapS> andygraybeal: I once had an FTP server that took 3 hours to fsck, resulting in a) a clean filesystem, and b) about 20% of the files disappearing due to corruption.
[15:20] <andygraybeal> SpamapS, very scary
[15:21] <andygraybeal> i don't know how to do a diff between my backups and my filesystem :(
[15:21] <andygraybeal> i guess i'll learn :)
[15:21] <SpamapS> andygraybeal: It had been running for 4 years w/o a reboot
[15:21] <andygraybeal> nice
[15:21] <SpamapS> and getting something like 50,000 uploads a day
[15:21] <andygraybeal> what o/s?
[15:21] <SpamapS> RedHat 8
[15:22] <SpamapS> on a Compaq Proliant DL380 (1st Generation)
[15:22] <andygraybeal> nice
[15:27] <evilsushi> how can I make server-wide bash prompt changes regardless of users .bashrc?
[15:29] <andygraybeal> i like your nick evilsushi :)
[15:29] <ph8> I realise it's a silly question, but i've come across a legacy intrepid box on my network, i don't particularly want to risk upgrading it but I want to install a few things - is there anywhere i can still get intrepid packages?
[15:30] <evilsushi> thanks =)
[15:30] <andygraybeal> sorry i can't help yuo :)
[15:32] <ruben23> hi guys i have uuntu server 10.04 LTS the problme is i can install lame, in anyway- i can really install it-any suggestion guys
[15:32] <evilsushi> You can install lame? what does that mean?
[15:32] <kpettit> Samba question.  ON my linux server I'm trying to navigate to a Windows share.  I can mount the server and main share.  But inside that share there are hidden directories.  I can't naviagte to the child of a hidden directory.
[15:33] <kpettit> For example //server/share/HidenParrent/ChildINeed  I need to get to "ChildINeed" but sense I cant' see HiddenParent I can't get there.  Any ideas?
[15:34] <kpettit> In Windows I can just paste in the whole URL and get there.  But in Linux sense one of the parrent directories I hidden or rather I don't have permisisons to see it, I can't navigate to the child I do have permissions to
[15:37] <SpamapS> ph8: http://old-releases.ubuntu.com/
[15:39] <ruben23> evilsushi: sorry i mean i cant install lame apps
[15:39] <ruben23> any idea guys how do i worked around on it
[15:39] <andygraybeal> LAME as in the mp3 encoder?
[15:39] <andygraybeal> or lame as in sucky
[15:40] <andygraybeal> or am i not up on my acronyms any longer
[15:41] <ruben23> lame sa mp3 encoder
[15:41] <ruben23> mp3 encoder- there is not pacakge available for ubuntu server 10.04 LTS
[15:42] <ph8> thx SpamapS
[15:44] <BBBThunda> anyone know a good resource for setting up kvm on ubuntu server for someone who's never used kvm?
[15:44] <ph8> You should probably look up libvirt-bin BBBThunda
[15:44] <BBBThunda> I've only used VirtualBox and VMWare Server mostly
[15:44] <evilsushi> ruben23: what error are you getting? why cant you install? we need more information.. 10.04 LTS?
[15:44] <ph8> it's an admin package for KVM virtual machines, lets you have XML config files and a nice interface
[15:45] <ph8> that and some knowledge about using qemu-img create to make qcow2 disk images is probably all you need
[15:45] <ph8> (read the kvm project page)
[15:45] <BBBThunda> nice... thanks ph8
[15:45] <evilsushi> ruben23: apt-cache search lame
[15:46] <ruben23> it just say no pacakge available
[15:48] <SpamapS> !serverguide
[15:48] <SpamapS> BBBThunda: ^^ that has some stuff about libvirt too
[15:49] <BBBThunda> Thanks SpamapS, I will check that out as well.
[15:52] <ruben23> evilsushi: you there..?
[15:52] <andygraybeal> SpamapS, running backup right now.. stuck a new tape in too
[15:52] <andygraybeal> evilsushi,  you don't happen to be a cook do you?
[16:38] <robbiew> hallyn: FedEx man arrived...ARM netbooks ;)
[16:39] <hallyn> robbiew: cool, are they shiny?
[16:39] <robbiew> heh...not sure, will open the box and let you know :P
[16:40] <robos> ubuntu 9.10 -- is there an easy way to disable ipv6? Is it as easy as echo 1 > /proc/sys/net/ipv6/conf/all/disable_ipv6 ?
[17:10] <ScottK> robos: First point is you should upgrade to 10.04 since there's no more security support for 9.10.
[18:02] <Jasonn> Is there any problem with running 64bit OS on a vps with 256mb ram??
[18:03] <cloakable> Besides pointlessness?
[18:04] <Jasonn> yes
[18:04] <Jasonn> cloakable: I need to run it because there is a problem with rtorrent on 32bit
[18:04] <Jasonn> Its an addon, it needs to run on 64bit
[18:05] <cloakable> Aha
[18:05] <Jasonn> Ok, wait how about on a server with 512 mB ram ?
[18:06] <Jasonn> is there any problem with that??
[18:06] <cloakable> Pointlessness? :P
[18:06] <Jasonn> lol
[18:06] <Jasonn> other than that ?
[18:06] <cloakable> Also, depends on if the vps supports 64bit
[18:06] <Jasonn> yeah, it does
[18:07] <cloakable> Then 256mb will (pointlessly) support it.
[18:07] <cloakable> You won't see the benefit of 64 though.
[18:08] <Jasonn> but will it effect the performance?
[18:10] <cloakable> Apparently not :)
[18:10] <bencc> I've installed ubuntu-minimal basic on vbox. after it asked me what to install (openssh, postgresql...) I just see a purple screen with a gray line at the bottom
[18:10] <bencc> is it ok, or does it still installing?
[18:11] <cloakable> Yea
[18:11] <bencc> ok, now it says installing grub... wonder why there was no indication before. thanks
[18:15] <robos> ScottK, do you have a second point?
[18:15] <robos> 9.10 is what we run for now and I immediately have to disable ipv6
[18:15] <ScottK> I was going to say that I think what you suggested is correct.
[18:15] <robos> ah
[18:17] <robos> what's weird is i used ifconfig eth0 inet6 del and also disabled it using /proc but if I run a nestat -na |grep tcp I still see tcp6 connections
[18:18] <RoyK> robos: connections or packets?
[18:18] <robos> I assume connections. An example line is this:
[18:19] <robos> tcp6       0      0 10.251.1.56:44632       10.251.1.142:11213      ESTABLISHED
[18:20] <robos> that's from netstat
[18:21] <maswan> that's a v4 connection, not v6
[18:21] <robos> hmm. what's tcp6 then?
[18:21] <RoyK> fe80::x:x:x something
[18:22] <RoyK> or x:x:x::x..
[18:23] <hggdh> smoser: can you set up an ec2 kernel for karmic-2.6.31-308.29?
[18:23] <maxagaz> I'm connected to my server with "ssh -D 1234 myuser@myserver", I've set Firefox to use proxy on localhost:1234, but it doesn't work (connection refused by proxy server), where could it come from ?
[18:23] <maswan> tcp6       0      0 2001:6b0:e:2018:::51494 2001:6b0:e:4040::123:22 ESTABLISHED
[18:23] <robos> i'm having problems with this box -- intermittent connections. Should I be concerned that i see tcp6 using netstat or shrug it off as normal?
[18:23] <maswan> that's an example of a v6 connection
[18:24] <robos> Yeah, i see it's certainly a ip4 IP. But not sure why netstat thinks it's 6. Even if I use netstat -na -6 it shows the tcp6 connections -- which my understanding of netstat is that should only show ipv6 connections
[18:25] <maswan> tcp6 can also do v4 connections, I forget how you use that, but it might be up to the application doing the connection
[18:25] <robos> yeah.. i have coldfusion listening on ipv6 addresses
[18:26] <robos> which i need to figure out how to turn that into ipv4 only, but that's another story :-)
[18:26] <RoyK> s/coldfusion/confusion/
[18:27] <robos> exactly -- haha
[18:27]  * RoyK isn't very fond of coldfusion
[18:28] <robos> i don't think anyone really is except those programmers who refuse to switch
[18:28] <RoyK> and when they do, they end up polluting your network with .net :P
[18:29] <robos> haha
[18:30] <maswan> deploy ipv6 instead. :)
[18:34] <SpamapS> ColdFusion is to the internet as COBOL was to corporate computing.
[18:37] <RoyK> SpamapS: heh - I was at a workshop today about using mpi/openmp for distributed stuff - FORTRAN is still holding the fort there, somehow :)
[18:37] <maswan> RoyK: Not surprising, it's better than C for writing fast code after all
[18:37] <RoyK> not really
[18:37] <maswan> (under some circumstances)
[18:38] <RoyK> it may be easier with newer fortran versions to write understandable code, but C is still fast
[18:38] <maswan> Much less aliasing problems to start with, then arrays "laid our correctly", etc also help
[18:38] <RoyK> we still have f77 in production with some models, and that code is rather ugly, beleive me :P
[18:39] <maswan> Yes, I know. I work in academic HPC.
[18:39] <RoyK> :)
[18:39]  * RoyK works for nilu.no
[18:39] <ScottK> ColdFusion was really cool in 1998.
[18:39] <ScottK> I think it was 1998.
[18:39] <ScottK> It was a long time ago in any case.
[18:39] <maswan> ah, hpc2n.umu.se here - but currently "hired out" to ndgf.org for the nordic LHC tier1
[18:40] <RoyK> last year, Eyjafjallajökull made us invest a bunch of money in storage, now, Grimsvötn may open for more compute power, which is rather nice
[18:40] <maswan> heh
[18:41] <RoyK> just got this quad machine, that is, four mobos in 2U, each with two 12-core opterons and 64 gigs of ram
[18:41] <zul> SpamapS:  why are you still using ftp ;)
[18:41] <RoyK> rather nice piece of hardware
[18:41] <maswan> Well, we're rather close to the part of cs.umu.se that does research into making matrix math go faster.
[18:41] <maswan> And they are writing new and fresh fortran.
[18:41] <cloakable> RoyK: very much so
[18:41] <cloakable> lmao
[18:42] <RoyK> I wonder how much they can stuff into 1U in a couple of years
[18:42] <maswan> Well, you could also get a 1U quad-socket machine for the same core density, but bigger nodes
[18:42] <RoyK> maswan: using openmp or mpi or something?
[18:43] <monaDeveloper> Hi there
[18:43] <maswan> RoyK: yes
[18:43] <RoyK> either of them or both?
[18:43] <monaDeveloper> I'm trying vgcreate volume /dev/loop0 but it always says device /dev/loop0 not found
[18:43] <maswan> RoyK: mpi definately, don't know if/how they do threading
[18:44] <cloakable> monaDeveloper: first, does /dev/loop0 exist?
[18:44] <maswan> RoyK: quite a lot of it is even in the core code for the chunks that then are parallelised
[18:44] <monaDeveloper> @cloakable yes
[18:44]  * RoyK just came from a conference of parallel processing http://www.notur.no/notur2011/programme.html
[18:44] <cloakable> Hmmmm.
[18:45] <RoyK> maswan: I don't think they have room for so much memory per node, though
[18:46] <RoyK> or perhaps they do...
[18:46] <maswan> RoyK: ehm, 512G is doable on the quad socket ones
[18:46] <RoyK> if they can rewrite the model to use GPUs, it may be a bit better than just using old CPUs
[18:47] <maswan> that's some of the research programs they have, yes
[18:47] <maswan> RoyK: heh, I was kind of hinted at that I should go for some parts of that, but I decided not to
[18:48] <RoyK> maswan: I guess it was about price as well - we got this box for NOK 130k with 4 x 2 x 12cores + 4 x 64GB RAM, which isn't that bad
[18:48] <maswan> RoyK: too much travel to geneva to be bothered
[18:50] <maswan> RoyK: As HEP folks have found recently, for the AMD case 1 x 4 socket is quite competitive with 2 x 2 socket price-wise.
[18:50] <maswan> but anyway, are you going to the ARC sessions tomorrow? :)
[18:50] <RoyK> yep
[18:51] <maswan> excellent. :)
[18:51] <RoyK> looking forward to it - it's quite new to me
[18:52] <maswan> that's part of what I'm doing for the last few years
[18:52] <RoyK> :)
[18:52] <RoyK> tell me, what exactly is ARC?
[18:53] <RoyK> just to get some background before tomorrow :P
[18:55] <maswan> ARC is a middleware that gives a common interface to many compute clusters, so that you can by one client remotely submit jobs to lots of clusters at once. They will then stage data from storage elements, run the calculations, and upload output to defined URLs.
[18:55] <RoyK> ok
[18:55] <maswan> Well, in the ARC project you also have clients, glue to put everything together (information system etc), but that's the basic part
[18:57] <monaDeveloper> I'm trying vgcreate volume /dev/loop0 but it always says device /dev/loop0 not found any idea?
[18:57] <RoyK> at this openmp session, a lady asked if it was possible to autoadjust the number of jobs according to load - so I wrote a little daemon to update the load average to allow for adjustments :P
[18:57] <monaDeveloper> device not found (or ignored by filtering)
[18:58] <RoyK> maswan: in your systems, do you overcommit or undercommit by terms of cpu-intense jobs?
[19:00] <maswan> RoyK: Neither, we currently schedule one job slot per core and only one user per node. in the future we'll do slurm with cgroups to schedule jobs with strict memory and cpu limits
[19:10] <RoAkSoAx> zul: ping?
[19:11] <zul> RoAkSoAx: on holiday whats up?
[19:11] <RoAkSoAx> zul: quick thingy then, Do you have any idea of why a package that has build-dep libldap2-dev won't build in oneiric as it will say openldap is too old, but it will build in natty?
[19:12] <zul> RoAkSoAx: no idea
[19:13] <RoAkSoAx> zul: k
[19:31] <cr3> hi folks, might anyone have a cloud init script for running ubuntu-desktop in ec2?
[19:52] <cloakable> why would you want to?
[19:53] <koolhead17> hi RoAkSoAx
[19:57] <zul> cr3: smoser might have one
[19:59] <ruben23> guys any idea how to install lame mp3 encoder on ubuntu-server 10.04 LTS
[20:00] <RoyK> ruben23: apt-get install ubuntu-restricted-extras
[20:00] <cr3> zul: I'm trying it by hand now, fun!
[20:02] <zul> cr3: cool but there should be desktop images already
[20:02] <ruben23> RoyK: then whats next install the lame..?
[20:03] <RoyK> libmp3lame0
[20:03] <RoyK> there may be others
[20:04] <cr3> zul: I couldn't find them here: http://uec-images.ubuntu.com/releases/*/release/
[20:05] <zul> cr3: meh sudo apt-get install ubuntu-desktop maybe ;)
[20:07] <cr3> zul: done, now evaluating vnc vs nx vs x2go
[20:09] <butter_fs> How does ubiquity behave when installing to an existing partition?
[20:10] <butter_fs> (specifically, I want to install to an already-made partition, and I really don't want it to blow away /home)
[20:20]  * butter_fs stamps his feet, and demands free support from skilled technical workers on a statutory holiday
[20:25] <Nelis> I'm trying to install 11.04 x64 server to a software raid 1 array but at first boot i get a black screen right after grub. Any ideas on how to solve this (tried reinstall)? The two drives both have a swap and ext4 partition mounted on / in raid 1. Grub is installed on the mbr of the first drive.
[20:27] <butter_fs> Nelis: remote quiet splash from the grub line, and try again;  that'll give you some more debugging info at least
[20:27] <Nelis> I pretty much followed this guide: https://help.ubuntu.com/10.04/serverguide/C/advanced-installation.html
[20:27] <Nelis> : ok
[20:28] <Nelis> butter_fs: No info, just a black screen right after. It is completely unresponsive.
[20:31] <butter_fs> Nelis: do you get the grub menu?  i.e., is it freezing at grub, or during boot after grub?
[20:35] <Nelis> butter_fs: Ok, I managed to get something on the screen now. I commented out the graphics line in the grub config. It now drops me into a busybox shell because it cannot find the root file system... I'll try editing grub some more, brb.
[20:36] <butter_fs> ah, k
[20:36] <butter_fs> Nelis: it's useful to see if you can mount the filesystem from busybox, so you know exactly what piece isn't working
[20:42] <Nelis> butter_fs: Before dropping me to busybox it sais: "alert! /dev/disk/by-uuid/[uuid] does not exist.". This is the device set as root in grub. It looks like i can see the whole filesystem from busybox. Any ideas?
[20:43] <butter_fs> for now, just use the device path as root, rather than root=uuid=
[20:43] <butter_fs> what filesystem is it?
[20:43] <Nelis> ext4
[20:43] <Nelis> nvm
[20:43] <Nelis> it was ofc the ramdisk..
[20:45] <butter_fs> heh
[20:47] <Nelis> could it be that the raid array is not loaded fast enough and the search for the root device times out?
[20:47] <Nelis> there don't appear to be any md* devices present yet
[20:48] <SpamapS> hallyn: hey, bug #350936 .. I think we should put our heads together at the rally and solve that.
[20:48] <SpamapS> hallyn: seems like it sits nicely at the intersection of our spheres of concern.
[20:52] <rcsheets> i've noticed that a key difference between https://help.ubuntu.com/10.04/serverguide/C/windows-networking.html and https://help.ubuntu.com/11.04/serverguide/C/windows-networking.html is that the 11.04 version does not mention Likewise Open. is it no longer recommended?
[20:53] <hallyn> SpamapS: excellent, let's do that
[20:53] <butter_fs> rcsheets: look at the relevant section in 11.04 :p
[20:53] <butter_fs> https://help.ubuntu.com/11.04/serverguide/C/samba-ad-integration.html
[20:53] <SpamapS> rcsheets: Its still in main.. 11.04 shipped with v6.0 .. not sure why it would have been removed from the server guide.
[20:54] <butter_fs> it wasn't
[20:54] <butter_fs> it's just not referenced by name in the table of contents
[20:54] <butter_fs> (because a name of a product is useless there)
[20:54] <rcsheets> oh i see
[20:54] <rcsheets> it's just been integrated into the rest of the docs better
[20:54] <rcsheets> silly me *sigh*
[20:54] <butter_fs> "it says likewise open, why would I care about that?  ooooo! it's how I join domain!  Why didn't they just say that?!"
[20:54] <SpamapS> RTFM - Read the FULL manual ;)
[20:54] <rcsheets> butter_fs: thanks
[20:55] <Nelis> butter_fs: What do you think, could it just be a timeout issue?
[20:55] <butter_fs> Nelis: could be;  can you paste your grub boot line here?
[20:55] <butter_fs> and/or check if you have rootwait set?
[20:55] <SpamapS> Nelis: RAID1 has some.. issues with 11.04 ... I'm working on a nasty bug right now
[20:55] <butter_fs> (if you don't have rootwait, try adding it)
[20:55] <SpamapS> Nelis: if you edit the grub line and remove 'quiet' .. that may help
[20:56]  * butter_fs just uses btrfs, hence the nick :p
[20:56] <SpamapS> butter_fs: god bless you for that. :)
[20:56] <rcsheets> ha, oh dear. the linked Likewise Open Installation and Administration Guide does not seem to have been updated to reflect version 6.
[20:56] <SpamapS> I wish they'd take the damn EXPERIMENTAL tag off it so we could all feel good about using it.
[20:56] <butter_fs> SpamapS: NO.
[20:57] <SpamapS> butter_fs: its got an fsck now doesn't it?
[20:57] <butter_fs> fedora _might_ be using by default this september, in which case it _might_ be sane to take the experimental tag off a year _after_ they don't have any fatalities to their name :p
[20:57] <SpamapS> I hear the main problem now is free space management.
[20:57] <butter_fs> SpamapS: no.
[20:57] <Nelis> SpamapS: I did, it's complaining it cannot find the boot device. When i look at /dev in the busybox shell it doesn't list any md devices, so maybe it hasn't loaded the raid array yet?
[20:57] <butter_fs> btrfsck is a readonly checker, which has no place being called fsck.btrfs
[20:57]  * butter_fs strangles a certain ubuntu dev
[20:58] <butter_fs> SpamapS: there might be a proper one by the end of this month though
[20:58] <njin> hello, can someone look at bug 787055 thanks
[20:58] <Nelis> butter_fs: rootwait is not set.
[20:58] <butter_fs> SpamapS: we still get messages like "this is meant for integration testing only, and should not be run by anyone who doesn't love crashes" from the lead dev in the mailing list :)
[20:58] <butter_fs> butter_fs: worth a shot
[20:59] <butter_fs> (get rid of any rootdelay lines)
[20:59] <SpamapS> butter_fs: :(
[20:59] <butter_fs> SpamapS: feel free to use it, I do :)
[20:59] <butter_fs> just don't come whining to #btrfs if it breaks with your only copy of your financial data!
[20:59] <SpamapS> Nelis: note that you may be dealing with bug #778520
[20:59] <butter_fs> (true story, some idiot)
[21:00] <SpamapS> butter_fs: nobody trusts ext4 with their only copy of anything either. :)
[21:00] <butter_fs> SpamapS: amazingly enough, yes they do :p
[21:00] <SpamapS> nobody sane. ;)
[21:00] <butter_fs> sec, you'll like this link
[21:00] <SpamapS> or informed
[21:00] <butter_fs> oh, he was informed
[21:00] <butter_fs> he was informed last november after the dmraid people went to heroic lengths to save his company, that switching to btrfs wasn't a reasonable solution
[21:01] <butter_fs> 5 months later, he's in btrfs's mailing list complaining about how we're neglecting the community, and how he has a business to run, and doesn't have _time_ to deal with this
[21:01]  * butter_fs was unimpressed :p
[21:02] <SpamapS> doesn't have time for backups either eh?
[21:02] <SpamapS> or transaction logs
[21:02] <butter_fs> apparently
[21:02] <butter_fs> "cacook", that's his name
[21:03] <butter_fs> (sorry, I had it muted in gmail, so it's a little tricky to find :p)
[21:03] <SpamapS> Well I feel that btrfs is stuck in a rut .. FS's are hard.. yes.. but at some point one has to wonder if it will ever be "stable"
[21:04] <SpamapS> How long was XFS in testing before SGI let us use it?
[21:04] <butter_fs> """What do you mean 'does Debian support BTRFS'?  The kernel supports it.  And why would they know more about BTRFS than you?
[21:04] <butter_fs> My whole system is installed over BTRFS.  If this is non-functional in any OS there should be a warning indicating it is non-functional."""
[21:04] <SpamapS> I'd say the same for JFS .. but JFS once ate my machine for breakfast.. ;)
[21:04] <butter_fs> (update-grub was bailing because it didn't understand the partitions, which broke his package manager)
[21:05] <SpamapS> I hate blamers
[21:05] <butter_fs> SpamapS: stable is determined by use by those able to deal with the risks and who need the features, or the foolhardy :)
[21:05] <SpamapS> If you don't know what it does, thats a risk. Accept the risk, and plan accordingly.
[21:05] <Nelis> SpamapS, butter_fs: looks like it is indeed that bug. Exactly the same problem.
[21:05] <butter_fs> \o/
[21:05] <SpamapS> Nelis: did removing 'quiet' at least get you past the initramfs ?
[21:06] <SpamapS> Its a nasty bug.. I think it may be udev/kernel timing problems.
[21:06] <SpamapS> race between when udev exposes the device and it can actually be used
[21:06] <butter_fs> incidently, it's fairly straightforward to modify an initramfs to force things to work
[21:06] <Nelis> SpamapS: i'm not booting with quiet and i get into initramfs every time.
[21:07] <butter_fs> basically, it's just adding a script to /etc/initram.d/hooks/local-top/
[21:07] <butter_fs> quiet is unrelated, it just shows less stuff before it dumps you
[21:07] <SpamapS> Nelis: quiet is on by default... you have to remove it. If that doesn't work, the rootwait may work.. but I wasn't able to get it to.
[21:08] <SpamapS> butter_fs: it has a nice net effect of making the race win in your favor though
[21:08] <SpamapS> by slowing the kernel timing down just a tiny bit
[21:08] <butter_fs> what?  really!?
[21:08] <butter_fs> eww?
[21:08] <SpamapS> terrible work around
[21:08] <butter_fs> Nelis: just modify initramfs, it's easy
[21:08] <butter_fs> (I had to do similar for btrfs before ubuntu added support)
[21:08] <SpamapS> But allows me to get in and install new packages / initrd without re-attaching a disk.
[21:08] <Nelis> butter_fs: ok, if you talk me through it, i'm not that savvy with linux -_-
[21:08] <butter_fs> /etc/initramfs-tools/hooks
[21:08] <butter_fs> is the relevant folder
[21:09] <SpamapS> yeah you actually just have to sleep 1 in there and it should work
[21:09] <butter_fs> er, I might have lied, sec, I need to log into work
[21:09] <butter_fs> SpamapS: you could do things right too :p
[21:09] <butter_fs> (give me a sec, just logging into work)
[21:10] <SpamapS> butter_fs: "right?"
[21:10] <butter_fs> SpamapS: "non racy"
[21:10] <SpamapS> butter_fs: by all means if you have a better grasp on the bug, I'd love to hear the right way.
[21:10] <lifeless> SpamapS: o/
[21:11] <butter_fs> Nelis: http://pastebin.com/dWzHNYV8
[21:11] <butter_fs> is the sort of file that goes in /etc/initramfs-tools/scripts/local-top/whatever_name
[21:11] <SpamapS> butter_fs: the race is happening somewhere between when the kernel says "ok you can use the device with UUID X" and the MD driver actually being able to see the mdX device and tell you if its degraded or not
[21:11] <butter_fs> most of it is boilerplate
[21:11] <SpamapS> lifeless: howdy
[21:12] <butter_fs> SpamapS: how would you politely ask mdadm to tell you?
[21:12] <SpamapS> lifeless: you're not going to ask me for cassandra packages again are you? ;)
[21:12] <rcsheets> do bugs against the official documentation on help.ubuntu.com go in launchpad somewhere?
[21:12] <butter_fs> rcsheets: yes
[21:13] <butter_fs> Nelis: if you're lazy, you can just change the btrfsctl to a sleep
[21:13] <SpamapS> butter_fs: the idea is to wait for the root device event to arrive on udev. Once it arrives, continue and try to use the device normally via mount. If there's a failure, you run mdadm to see if its a degraded array and if it is, act accordingly....
[21:14] <SpamapS> butter_fs: the problem is somewhere between that udev event and mdadm checking the device, there's a race.
[21:14] <lifeless> SpamapS: :P no, but I did just mail evand a link to oops-repository
[21:14] <Nelis> butter_fs: i'll need some more details on how to implement your solution :)
[21:14] <butter_fs> SpamapS: so putting some blocking call to mdadm (or a loop if it comes to that) in a local-top script should make that all go away :p
[21:14] <SpamapS> lifeless: I think at this point, they've gone so far off the reservation with java packages, we're better off using upstream's fully embedded jars. :(
[21:15] <SpamapS> butter_fs: precisely what I've been working on. But a way to get back up and running until we fix that.. is sleep 1. :-/
[21:15] <butter_fs> SpamapS: damnit, fix it proper :D
[21:15] <nimrod10> how can I kill a process that it is marked in ps as    DL     ?
[21:15] <rcsheets> butter_fs: oh sorry, just saw it at the bottom of the page. *facepalm*
[21:15] <butter_fs> nimrod10: you basically can't :p
[21:15]  * SpamapS gets out his Proper Brand Duct Tape
[21:16] <nimrod10> damn it ! I don't like it when I'm right !
[21:16] <butter_fs> nimrod10: if it's blocking inside a kernel function, then killing the task would kill a random kernel function without cleanup
[21:16] <SpamapS> lifeless: the last I checked, 0.7.4 and later failed with on disk format problems because of one of the archive java jars we're using instead of theirs.
[21:17] <butter_fs> if you can figure out why it's blocking, and fix the underlying condition, you can sometimes get things back quicker than otherwise
[21:17] <butter_fs> but if it's, for instance, because of a buggy driver, you're probably out of luck
[21:17] <SpamapS> like the old "bring up another box with the same IP as the downed NFS server" trick
[21:17] <butter_fs> exactly :D
[21:17] <nimrod10> thanks butter_fs
[21:17] <butter_fs> Nelis: the file I pasted goes in the folder I pasted, and you edit the last line according to your needs
[21:18] <butter_fs> (probably a sleep 1 for now)
[21:18] <butter_fs> and then you remake your initramfs's, via "update-initramfs -k all -c" (or just update-initramfs -k <kernel version> -c" if you know it and don't want to make _every_ installed kernel have the change)
[21:19] <butter_fs> that way, every future kernel will also get a matching initramfs with those changes automagically
[21:19] <butter_fs> SpamapS: I'm assuming that the tricky part is making it not act silly for everyone who doesn't have a degraded raid1 array
[21:20] <SpamapS> butter_fs: no that works perfectly no matter what
[21:20] <SpamapS> butter_fs: the tricky part is making *damn* sure you detect the degraded array
[21:20] <SpamapS> So you don't accidentally sync a RAID1 in the wrong direction
[21:20]  * butter_fs huggles his btrfs checksums
[21:21] <SpamapS> Yeah since reading about ZFS and BTRFS .. I can't believe we're still doing this RAID thing. ;)
[21:21] <butter_fs> nay, /me makes sweet love to his btrfs checksums
[21:21] <SpamapS> !ohmy
[21:21] <nimrod10> something is definitely wrong with my setup. butter_fs, I hate that I have to reboot every time keepassx locks like this  with DL
[21:21] <SpamapS> butter_fs: just a friendly reminder. :)
[21:21] <Nelis> butter_fs: i do these changes from the busybox shell?
[21:21] <butter_fs> SpamapS: I had my own ubottu factoid once :p
[21:21] <SpamapS> btrfs checksums have feelings too.. make sure you call them. ;)
[21:21] <butter_fs> Nelis: no
[21:22] <butter_fs> Nelis: you do them from a booted system, however you manage that (chroot or whatever)
[21:28] <Nelis> butter_fs: thanks for the help, but that's too much tinkering for me. For now anyways. Any idea in what timespan I can expect a fix for this issue to be available?
[21:28]  * butter_fs points at SpamapS
[21:28] <Nelis> :D
[21:33] <SpamapS> Nelis: it will be in natty-proposed some time this week. It needs to sit there for 7 days and have some independent verification (you would be a great candidate for verification!) and then it will go into natty-updates
[21:33] <SpamapS> Would have been fixed sooner but UDS sort of blew it up. :-P
[21:34] <Nelis> SpamapS: I'd be happy to do some testing, but I'd need someone to talk me through it, as I'm not that good with linux.
[21:35] <SpamapS> Nelis: understood. For now, do you have two disks?
[21:36] <SpamapS> Nelis: if so, when you get to the initramfs> prompt, try 'cat /proc/mdstat'  .. does it show your RAID arrays?
[21:36] <Nelis> SpamapS: 2x 640gb WD drives
[21:37] <robbiew> hallyn: got an ARM netbook for you :)
[21:38] <hallyn> robbiew: I'm ready :)
[21:38] <Nelis> SpamapS: no devices shown.
[21:39] <SpamapS> robbiew: I'll give you a 12oz bar of gold for one of those and the lamp from Christmas story...
[21:40] <SpamapS> Cause everyone knows a 12oz bar of gold costs an ARM and a Leg
[21:40] <SpamapS> Nelis: Interesting!
[21:40] <SpamapS> Nelis: did you create the RAID during installation?
[21:40] <Nelis> SpamapS: yes
[21:40] <robbiew> SpamapS: heh
[21:41] <Nelis> SpamapS: i used this guide: https://help.ubuntu.com/10.04/serverguide/C/advanced-installation.html
[21:43] <SpamapS> Nelis: fdisk -l /dev/sda .. what do you get?
[21:43] <Nelis> SpamapS: sda should be a raid device right? I have 6 2 TB WD drives at sda-sdf, sdg and sdh are the 640GB raid drives
[21:45] <Nelis> SpamapS: looks like there's no fdisk available -_-
[21:45] <SpamapS> no, sda should be a normal device with auto-detect RAID partitions
[21:46] <SpamapS> ah.. hrm
[21:46] <SpamapS> ahh sdg and sdh should be the regular drives then
[21:47] <SpamapS> Nelis: try 'mdadm --misc --query --scan'
[21:48] <Nelis> SpamapS: you think if i installed the raid drives as sda and sdb they might be mounted faster to get around this bug? -_-
[21:48] <SpamapS> slightly possible
[21:48] <SpamapS> shouldn't really matter tho
[21:49] <Nelis> SpamapS: "create user root not found - create group disk not found - no devices given."
[21:49] <SpamapS> I'm more surprised that there's nothing in mdstat
[21:59] <Nelis> SpamapS: is there anything else you want me to do?
[22:11] <SpamapS> Nelis: I'm really curious why you're not seeing the array. Do you have a file, /etc/mdadm.conf ?
[22:14] <Nelis> in the initramfs?
[22:14] <Nelis> no
[22:21] <SpamapS> Nelis: interesting. It sonds like your RAID wasn't actually created properly.
[22:21] <SpamapS> sounds rather
[22:21] <SpamapS> Nelis: the system did boot at one time right?
[22:22] <Nelis> SpamapS: no, i just installed as per the guide.
[22:24] <SpamapS> Nelis: what does mdadm -E /dev/sdg1 show ?
[22:25] <Nelis> SpamapS: "create user root not found - create group disk not found - no devices given."
[22:25] <Nelis> ehm
[22:25] <Nelis> SpamapS: cannot open /dev/sdg1: no such file or directory
[22:26] <Nelis> SpamapS: are no partitions shown in /dev for sdg and sdh
[22:32] <SpamapS> Nelis: yeah it seems to me that your partitions were not created properly
[22:33] <Nelis> SpamapS: the installer actually shows the partitions and raid though
[22:33] <SpamapS> Nelis: entirely possible its a bug.. those install directions look pretty straight forward
[22:34] <SpamapS> Nelis: can you boot with 'debug' and then examine the file it writes? Its something like /run/initramfs/initramfs.debug
[22:35] <Nelis> SpamapS: when i do a reboot it sais "stopping all md devices", kinda weird since there aren't any? :D
[22:36] <RoAkSoAx> adam_g: ping?
[22:38] <Nelis> SpamapS: I added debug to the boot options but I'm not seeing the file you're talking about
[23:02] <SpamapS> Nelis: hrm
[23:14] <bencc> I'm installing ubuntu server 1..04 on vbox. do I need LVM?
[23:18] <Nelis> SpamapS: I'll probably be back tomorrow after some testing.
[23:36] <kirkland> hallyn: ping
[23:36] <kirkland> hallyn: do you have an documentation/instructions/blog post on how exactly to use the multiple encrypted private directories feature we added to Natty?
[23:37] <kirkland> hallyn: looks like we have a teaser in the manpage
[23:43] <hggdh> smoser: how to debug an instance that does not boot on AWS (in this case, Maverick 2.6.35-29-virtual)?
[23:44] <hggdh> smoser: i386, m1.small
[23:54] <hallyn> kirkland: one sec
[23:55] <hallyn> kirkland: i never did a new one for the merged version.  my original post was http://s3hh.wordpress.com/2010/10/27/emount-ecryptfs-mount/
[23:56] <hallyn> kirkland: but really it's still all the same, especially since i alias 'mounte' to mount.ecryptfs_private, and umounte to umount.ecryptfs_private
[23:56] <kirkland> ah
[23:56] <kirkland> hallyn: k
[23:56] <kirkland> hallyn: thakns
[23:56] <hallyn> so i'm not sure anything changes :)
[23:56] <hallyn> np.  had fun out east?
[23:56] <hallyn> i just saw your copious blog posting :)
[23:57] <kirkland> hallyn: yeah, was good
[23:57] <kirkland> hallyn: i need to put up 2 more for the last 2 days ;-)
[23:57] <kirkland> hallyn: plitvice was amazing
[23:57] <hallyn> but, i didn't see a pic of you with ivoks' cats :)
[23:58] <hallyn> jinkeys, linus is talking about '3.0'