[00:48] <wiredfool> hey, I've got a 10.04 server running 2 bridges ( and quite a few :0,:1 extra interfaces) -- I get about a 2 minute hang on reboot at deconfiguring network interfaces. any ideas?  is there a way to just timeout dropping the interfaces in a couple seconds and just reboot the thing already?
[01:27] <econnell> hello all.  We have started working our migration from Solaris to Ubuntu Server because of the recent licensing changes with Oracle, and I thought the Ubuntu server team may be interested in some observations.
[01:29] <econnell> First, let me explain our infrastructure a bit, our data centers are unmanned US east and west coast facilities, and we access the console of these servers via SSH to an ILOM.  I imagine this is a fairly common use case.
[01:30] <econnell> As such, netboot via PXE is a requirement, as well as serial port installations.  This is 100% doable via the Ubuntu server install, but it is not documented at all.  The fixes to do this are extremely simple (once you figure out what needs to be done).  Basically, syslinux.cfg in the netboot img needs to be modified to have "serial 0" in the config and all of the kernel parameter option lines need to be modified to specify the console as
[01:30] <econnell>  ttyS0
[01:31] <econnell> i was quite surprised this wasn't documented anywhere actually... or at least, not that I could find
[01:33] <econnell> perhaps there should be an option in the installer for "install using serial console" along with the cli and normal install methods.
[01:44] <JanC> econnell: it's probably documented _somewhere_
[01:44] <yann2> econnell, I'd warmly welcome a blog post about your migration once it's done
[01:45] <yann2> it's a bit weird for me, am waiting for oracle to finish btrfs so I can move away from my only solaris box (ZFS)...
[01:45] <JanC> lol
[01:45] <lostogre> I assume that UEC can run windows 2003 hosts, but I haven't seen anything about this. Can it?
[01:45] <lostogre> guests that is....
[01:46] <yann2> lostogre, I understannd UEC is based on KVM - I suggest you to have a look there: http://www.linux-kvm.org/page/Guest_Support_Status
[01:46] <lostogre> yann2, that's what I thought, but I haven't heard of anyone trying it.
[01:47] <yann2> I've quite a few windows XP, 2000 running on KVM, but not on UEC, and not windows 2003 I'm afraid
[01:47] <lostogre> how stable is it? I am looking for a replacement for VirtualBox.
[01:48] <yann2> mmmh.
[01:48] <econnell> yann2: i'm already missing zfs :(
[01:48] <yann2> using KVM with libvirt, not UEC - rather stable, wouldnt recomment previous versions
[01:48] <JanC> our locoteam VPS is a kvm VM and seems to run fine
[01:48] <yann2> and still a few issues with 10.4 but managed to work around so far
[01:49] <econnell> and zones... man am i missing zones
[01:49] <lostogre> issues with 10.04?
[01:49] <yann2> lostogre, bah nothing too bad, a PXE boot that wasnt happy this morning
[01:50] <lostogre> don't use pxe so it wouldn't matter.
[01:50] <yann2> should be alright
[01:50] <yann2> disks a bit slow, but havent tried the virtio drivers for that yet
[01:51] <yann2> (under windows mostly)
[01:51] <yann2> I think the virtio drives would be of good help, for network and disks
[01:51] <lostogre> right now we have a virtualbox server that is giving us fits. First problems with asynchronous IO problems with xfs (not VB's fault, I know), but when I upgraded the kernels to fix the problems the guest drivers wouldn't recompile causing other filesystem problems.
[01:52] <yann2> VB is oracle's anyway, probably the next on the list imo :)
[01:52] <lostogre> I have to admit, that is a concern.
[01:52] <lostogre> how bad are the virtio drivers to work with anyway?
[01:53] <yann2> msi files, click on them and it's done?
[01:53] <yann2> (as far as I remember)
[01:53] <lostogre> Oh. huh.
[01:53] <lostogre> so why haven't you tried the virtio drivers for disks?
[01:53] <yann2> maybe you have to do the add hardware > find drivers stuff, cant remember
[01:54] <yann2> fairly new I believe. Not that concerned about performance on windows in my case
[01:54] <yann2> I definitely should give them a try
[01:54] <lostogre> I see.
[01:54] <lostogre> heh.
[01:57] <yann2> off to bed, good night everyone, good luck lostogre
[01:57] <lostogre> I'm looking at the guest support status you gave me; I see that win2k3r2 runs under qemu-kvm, but Ubuntu 10.04 runs under kvm-84.
[01:57] <lostogre> Oh, ok.
[01:57] <lostogre> thanks.
[01:57] <yann2> it's fine, it'll work ;)
[01:58] <lostogre> what gives then? what's the difference?
[01:59] <lostogre> can anyone tell me the difference between qemu-kvm and kvm-84?
[02:00] <JanC> kvm-* are old versions before their fork got merged into upstream qemu
[02:01] <lostogre> I see. So qemu is it now?
[02:02] <JanC> well, there is the kvm kernel module, which is part of the upstream kernel, and there is qemu-kvm which is part of upstream qemu
[02:03] <lostogre> I see. Ok, so it is about like I remember.
[02:03] <lostogre> thanks.
[02:04] <lostogre> Different topic: what kernel do I need in order to successfully patch the asynchronous IO bug in XFS?
[02:04] <lostogre> This is a lucid server.
[02:11] <lostogre> MikeGuo, you work for canonical?
[02:11] <MikeGuo> lostogre, urmm. yeah.
[02:11] <lostogre> mind if I ask a question?
[02:12] <MikeGuo> lostogre, sure. if I can help. ;),
[02:12] <MikeGuo> lostogre, how do you know I work for canonical?  :)
[02:13] <lostogre> MikeGuo, your login says that your host is emperor.canonical.com.
[02:13] <todd> hmm
[02:13] <lostogre> On a lucid server, which kernel in the official repos has the XFS problems fixed.
[02:14] <lostogre> XFS/ext4 problems.
[02:15] <MikeGuo> lostogre, sorry. man,  I really don't know about your question.  but  may be you can find the answer on LP
[02:15] <lostogre> LP?
[02:15] <JanC> wasn't that before Lucid?
[02:15] <todd> launchpad
[02:15] <lostogre> ah.
[02:15] <MikeGuo> yeah.
[02:15] <lostogre> JanC, no, I have a server that is suffering from those problems.
[02:16] <lostogre> or at least something like it.
[02:17] <lostogre> I have several VM running in VirtualBox that are halting with XFS errors.
[02:18] <detrix> Hi everyone.  problem: wife has a netbook (acer) been running Ubuntu Standard Gnome (not the UNE). She found a program that was very similar to a PIM program she used on windows, but it is a bunch of KDE programs.  Would this cause her computer to run really slow?
[02:19] <detrix> Sorry wrong channel
[02:19] <lostogre> the only slowdown that I can think of is that she will have gnome libraries and qt libraries in ram at the same time.
[02:20] <detrix> hmmmm
[02:51] <ASULutzy> Hey all, I wrote a little shell script here that is pretty simple. Wanted to remux a bunch of .avi's I had into .mp4's. http://pastebin.com/icdbxxN1 For some reason, the script fails, and I don't get why, output here: http://pastebin.com/W0PGAKu9 Finally, here's the output of echo'ing $file instead of ffmpeg'ing it http://pastebin.com/01KerYRM Thanks for your help!
[02:53] <twb> find -type f -name \*.avi -exec sh -c "ffmpeg -i "$0" -vcodec copy -acodec copy "${0%.avi}.mp3"' {} \;
[02:54] <ASULutzy> Thanks, I'll try that
[02:54] <ASULutzy> Any idea why mine fails?
[02:54] <twb> That's just better style, I don't know what's failing
[02:54] <patdk-lap> dunno
[02:54] <patdk-lap> mine works
[02:54] <ASULutzy> In the second pastebin, it gives "uered the World.avi: no such file or directory"
[02:55] <ASULutzy> It's chopping up $file for some reason it seems like?
[02:55] <patdk-lap> fine . -type f -name '*.avi' | while read i; do echo "Test " $i; done
[02:55] <patdk-lap> oh, spaces
[02:55] <ASULutzy> It works fine when I echo it
[02:55] <patdk-lap> you have to escape them :)
[02:55] <patdk-lap> or put it in "
[02:55] <patdk-lap> fine . -type f -name '*.avi' | while read i; do echo ffmpeg -i "$i" ....; done
[02:55] <ASULutzy> It's quoted?
[02:55] <patdk-lap> some for twb's
[02:55] <patdk-lap> "{}"
[02:55] <twb> Whitespace issues wouldn't account for the problem you're seeing.
[02:55] <ASULutzy> http://pastebin.com/icdbxxN1 source
[02:56] <ASULutzy> Yea, it's kind of baffling to me
[02:56] <ASULutzy> I'll try your different style and see if it works
[02:56] <twb> patdk-lap: you only need to quote {} in csh and tcsh
[02:56] <patdk-lap> I have had issues in bash
[02:56] <twb> If you say so.
[02:58] <ASULutzy> twb: Did you mean to have that last single quote in what you typed?
[02:58] <twb> Oh, sorry
[02:58] <twb> find -type f -name \*.avi -exec sh -c 'ffmpeg -i "$0" -vcodec copy -acodec copy "${0%.avi}.mp3"' {} \;
[02:59] <ASULutzy> So far yours seems to be working better already
[02:59] <twb> Well, I *am* a genius.
[02:59] <ASULutzy> I'm still really confused as to why mine didn't work. I thought maybe one of the files had a ' in it
[02:59] <ASULutzy> so I trimmed ' from all the filenames
[02:59] <ASULutzy> Still didn't work
[02:59] <twb> Ask #bash sometime
[02:59] <ASULutzy> Good call
[03:00] <ASULutzy> Another quick question, you think it would be worth shoving gnu parallel in there sometime over an nfs share on Gigabit Ethernet, or you think the network is probably the bottleneck there?
[03:00] <ASULutzy> Probably a dumb question, I guess I could just ffmpeg -threads 4
[03:01] <patdk-lap> "${file/%.avi/.mp4}"?
[03:01] <patdk-lap> shouldn't it be like "${file%.*}.mp4"
[03:01] <patdk-lap> or atleast that is how I do it
[03:01] <patdk-lap> heh
[03:01] <twb> patdk-lap: no.
[03:01] <ASULutzy> I originally had base=`basename "$file" .avi` dir='dirname "$file"`, but my friend who is quite the guru told me I was a noob and suggested that, the problem existed both ways
[03:02] <twb> Because that will chop of extensions other than .avi
[03:02] <patdk-lap> we are already grepping on avi extentions, who cares :)
[03:02] <patdk-lap> but then, I normally convert from more than just avi
[03:04] <ASULutzy> Thanks for your help though, I tried asking in #ubuntu first, but since it wasn't related to the same 5 questions asked all day, didn't get much response ;)
[03:05] <ASULutzy> Your fancier one liner seems to be working though
[03:29] <overrider> Hello there, i have a vga monitor connected to my server and am using to monitor some statistics. Unfortunately, it will go blank after a few minutes not using the keyboard. How to make it so that the monitor never goes off / blank / standby? Many thanks
[03:41] <demonspork> how can I do an incremental backup of a directory on my server, the directory is large and just archiving it is going to take a lot of space because I intend to keep the backups for a long time
[03:43] <patdk-lap> probably want to look at rsyncdiff
[03:43] <patdk-lap> or backuppc
[03:43] <patdk-lap> all depends on what kind of route you want to go
[03:47] <twb> I use rsnapshot instead of rsyncdiff; anything based on hard-linking and the rsync algorithm should be about the same.
[03:49] <Dantes13> Did anyone here upgrade from 8.04 to 10.04 instead of doing a clean install?  How did the upgrade go?  Any issues?
[03:49] <twb> I remember having problems with d-r-u
[03:49] <twb> I only tried it in a VM prior to the actual release of 10.04
[03:49] <patdk-lap> I'm not sure I have upgraded any 8.04 to 10.04 yet
[03:50] <patdk-lap> the only two that I have, went through 8.10->9.04->9.10
[03:50] <ScottK> Dantes13: I've done a couple with no issues.
[03:52] <Dantes13> Awesome ScottK.  That's what I wanted to hear!  :)
[03:59] <fluvvell> Dantes13, I've done 4 upgrades from 8.04 to 10.04, not without incident but easier than earlier upgrades
[04:01] <fluvvell> but I have to say, eve  a clean install where I was migrating a system with raid1, the 10.04 setup does not make for easy raid configuration.
[04:02] <fluvvell> no, the upgrades were not true "server" installs.
[04:44] <kiall> Hiya - anyone around who's familiar with UEC? (specifically EBS volumes claiming to attach correctly, but never appearing inside the instance)
[04:45] <databits> when you are typing a password or info into a txt box they have the X and check indicators does anyone know what kinda of function that would be to add to a website
[04:45] <databits> I'm trying to google, but I'm not sure exactly what to search for
[04:50] <segv`> weirdest issue today, have a system that as soon as "hardware detection" runs, the drives decide to go kaput
[04:50] <segv`> So I can't install ubuntu server, BUT if I load them from a live cd, just fine, if I go into rescue mode on the same ubuntu live cd, i can mount them, only during install..
[04:51] <segv`> and the USB goes out as well
[04:51] <segv`> acpi issue?
[05:08] <alex88> time
[05:09] <twb> segv`: you're booting the d-i CD?
[05:09] <twb> segv`: does "kaput" mean a kernel oops?
[05:10] <twb> segv`: if not, switch to vt2, hit RET to create a shell, then examine the end of /var/log/syslog and dmesg for hints
[06:08] <kiall> Any UEC users about?
[06:16] <twb> soren: kiall (presumably) has a UEC question
[06:16] <twb> !anyone > kiall
[06:20] <kiall> twb, hiya
[06:21] <kiall> yea its a UEC question / bug .. (99.99% sure its a bug .. not sure if its QEMU/eucop tho .. half way through building hacked version that will point me in the right direction ;)
[06:24] <twb> kiall: you have no describe the issue.
[06:26] <kiall> sorry - With UEC (10.10) EBS volumes fail to attach correctly, nothing appears in the instance ... manually attaching via virtsh works - as long as you use bus=virtio ... and telling UEC to use bus=virtio doesnt seem to work .. its a tad difficult to explain ;)
[06:26] <twb> No worries.
[06:26] <twb> That message should've been your first message
[06:26] <kiall> been at this about 8 hours - brain melted at this stage ;)  soryr!
[06:26] <kiall> sorry*
[07:32]  * pennyless is away: Gone away for now
[08:54] <sander^work> mount.nfs: rpc.statd is not running but is required for remote locking. mount.nfs: Either use '-o nolock' to keep locks local, or start statd.<-- How can I get rpc.statd to run on bootup?
[09:02] <twb> sander^work: on 10.04?
[09:04] <twb> This is how I do it: http://paste.debian.net/99437/
[09:04] <twb> Yet, it's disgusting.
[09:04] <twb> Yes, it's disgusting. *
[09:05] <sander^work> twb, yes... on 10.04
[09:08] <maedox> Does anyone know how I can add L-IPA, L-NAM and L-DOM in the PowerAdmin interface?  It's there in an older version on another server, but not in the latest one I set up.
[09:14] <twb> maedox: I'm not familiar with any of those terms.
[09:15] <twb> Ah, apparently its a web UI for the pdns DNS server.
[09:15] <maedox> twb: Me neither really, but they are very handy to check that the DNS zones are set up correctly. You can see IP's with PTR and A etc. and e.g. IP's with non-matching A and PTR are marked in red.
[09:15] <maedox> twb: correct.
[09:15] <twb> I usually just use dig
[09:16] <sander^work> twb, what does that script actually do?
[09:16] <twb> sander^work: edits files
[09:16] <sander^work> twb, where should I put it?
[09:16] <sander^work> just run it?
[09:16] <twb> You need to understand it and then make the changes by hand, since you don't have the associated interpreter.
[09:16] <maedox> twb: yeah that works fine, but just for finding e.g. typos/issues quickly the web UI is better.
[09:16] <twb> Shrug.
[09:17] <twb> I guess if it was me, I'd set up something to send me an email when it noticed a problem, rather than having to look at a web page for colours.
[09:19] <maedox> That's probably a good idea, but AI is not perfect still. :) It's hard to make a script figure out which domain name/A/PTR is the correct one. It's basically for finding issues with older records that may have been changed without care. There was another guy running the DNS servers a while back, and I have been finding strange issues ever since.
[09:20] <maedox> If I add new records I always check that it works as expected, but it would be nice to have those features in PowerAdmin anyway.
[09:21] <maedox> They might have just deprecated them...
[09:32] <sander^work> twb, Isn't it just to start rpc.statd on bootup?
[09:33] <twb> How do you know *when* to start rpc.statd?
[09:34] <twb> There are already upstart jobs there, but they have a cyclic dependency.
[09:34] <twb> What my code does is replace one of the nodes in the cycle with a simple loop that says "try again" every tenth of a second
[09:35] <twb> About half the problem is that upstart took over the job of mounting filesystems before it was actually capable of doing so, which is why there's this intermediary kludge mountall(8) daemon.
[09:36] <sander^work> twb, I get an error at bootup sometimes.. saying it can't mount an nfs partition.. and wondring if I should do it manually.. or just continue.
[09:36] <sander^work> ..Is that part of the problem?
[09:36] <twb> I think it's the same issue
[09:37] <twb> I'm *booting* from NFS, so I probably just notice it more.
[09:37] <sander^work> twb, ok.. but that error message I pasted to you.. I got long time after I booted.. just because I didnt have rpc.statd started.
[09:38] <twb> rpc.statd is required for rw nfs mounts.
[09:38] <sander^work> Yeah.. I guess it wasnt started because I didnt add the nfs lines untill after I booted
[09:38] <sander^work> ..in fstab.
[09:38] <twb> Possibly
[09:39] <twb> Although the portmap-based services OUGHT to start regardless of whether they think they're needed
[09:39] <sander^work> twb, so how do I use your script, without understanding it all?
[09:39] <twb> That's not possible.
[09:40] <sander^work> twb, I understand it.. But don't understand under what context I should use it.
[09:47] <alvin> twb: interesting script :)
[10:26] <Dravekx> anyone know how to get rid if the update error on Maverick? " A error occurred during the signature verification. The repository is not updated and the previous index files will be used. GPG error: http://extras.ubuntu.com maverick Release: The following signatures couldn't be verified"
[10:27] <twb> Dravekx: are you going via a proxy?
[10:28] <Dravekx> twb. no. Im typing sudo apt-get update via ssh on my server.
[10:28] <Dravekx> no proxy
[10:28] <twb> The the most likely scenario is that repository is currently being updated
[10:29] <kiall> the extras repo doesnt actually have anything in it .. no idea why its in your sources list
[10:29] <Dravekx> idk
[10:29] <twb> Oh, or it might not be signed by a key in your current keyring
[10:30] <kiall> best bet is to just remove (or comment out) the repo from /etc/apt/sources.list - it most likely shouldnt be there anyway
[10:30] <twb> kiall: maybe he ran one of those stupid scripts that claim to DTRT
[10:30] <twb> You know, "one click to get mp3 and divx" or whatever.
[10:33] <Dravekx> which part to comment out? Canonical 3rd party?
[10:34] <Dravekx> its a simple media home server ftw.
[10:34] <_ruben> the referring to extras.ubuntu.com obviously ;)
[10:34] <_ruben> the one even
[10:34] <Dravekx> got it :)
[10:35] <Dravekx> that did it
[10:35] <Dravekx> thanks
[10:36] <niitty> i have problem with install -> itry to install from usb. i dont have cdrom. and installation "The Failing step is: Detect and mount CD-ROM. have anyone this problem too?
[10:37] <serard> hello
[10:37] <niitty> so how i can tell that it search all files from usb?
[10:41] <twb> niitty: presumably you're booting an install image that you downloaded.  What URL did you download it from?
[10:43] <niitty> i downloaded it from ubuntu.com/server (64-bit) and put that image to usb with those instructions.
[10:45] <twb> The full URL, please
[10:48] <niitty> oh sorry: image is from: http://www.nic.funet.fi/pub/mirrors/releases.ubuntu.com//maverick/ubuntu-10.10-server-amd64.iso and I mount that image to usb with http://www.pendrivelinux.com/universal-usb-installer-easy-as-1-2-3/
[10:49] <niitty> http://www.ubuntu.com/server/get-ubuntu/download this page include all those infos
[10:50] <twb> Humph, OK.
[10:51] <twb> Try this one: http://archive.ubuntu.com/ubuntu/dists/lucid/main/installer-amd64/current/images/netboot/boot.img.gz
[10:52] <twb> I can't provide support for that pendrive converter thing.
[10:53] <niitty> ok. ill check that. thanks!
[10:58] <eriksson25> Anyone that could help, I am about to create my secund raid on a server. Have md0 up and runing and wants to create md1. I have started md1 with mdadm --create but is there anything I shuld think on before i create a new mdadm.conf file? Since I alredy have one md0 runing.
[10:58] <niitty> uhm, but what i should do with that img?, sorry to bother but im a bit new here.
[11:00] <twb> niitty: gunzip it and dd it onto a USB key.
[11:02] <eriksson25> Anyone know If I need to create a new mdadm.conf since I created me secund raid
[11:03] <twb> eriksson25: no, you just ensure that the existing one mentions both arrays
[11:03] <eriksson25> Oki, will have a look into that.
[11:03] <twb> eriksson25: e.g. http://paste.debian.net/99442/
[11:04] <eriksson25> My newly created array was created with one spare, but I would like to have it used. I will add 3 more hdd in just a day and will use one of them as spare.
[11:04] <eriksson25> How can i activate the spare disk?
[11:05] <eriksson25> and thanks for the link.
[11:05] <twb> I don't remember
[11:06] <eriksson25> If I dont have a spare, and one disk fails is the raid fucked? Or do I just need to add a other disk and recover the array and then remove the faulty disk?
[11:06] <twb> RAID1 can lose up to one node without data loss.
[11:06] <eriksson25> raid 5
[11:07] <twb> Er, sorry
[11:07] <twb> RAID5 can lose up to one node without data loss.
[11:07] <twb> RAID1 can lose n-1 disks.  RAID6 can lose 2 disks.
[11:07] <eriksson25> So why have disks as spare?
[11:07] <twb> So that the disk isn't spinning
[11:07] <twb> i.e. isn't subject to mechanical wear
[11:07] <eriksson25> I see.
[11:08] <twb> ...I think.  I've never actually created an array with a spare disk
[11:08] <eriksson25> I just folowed a guide and it got created with a spare by defoult.
[11:09] <twb> I wish people would not follow "a" guide
[11:09] <jpds> eriksson25: Hot-spares are useful for when a drive fails and you want something to kick in immediately.
[11:09] <twb> The number of times I've caught a certain coworker doing something on (say) 10.04, using a guide written for (say) 6.06...
[11:10] <eriksson25> mm, understand =)
[11:10] <twb> jpds: but how are they better than just adding another disk (e.g. three-way RAID5 becomes four-way RAID6)
[11:10] <_ruben> that'd be a "cold spare", a "hot spare" does spin .. and gets added to a raid volume automatically whenever it becomes degraded
[11:10] <twb> _ruben: I thought a cold spare didn't have power going into it
[11:11] <eriksson25> Well do I need to let it rebuild 100% bofore I change it to include the spare.
[11:11] <twb> _ruben: a hot spare was free to spin down because there are no writes to it
[11:11] <_ruben> twb: fair enough, tho that's more a powermanagement feature than a raid feature
[11:11] <twb> Granted
[11:12] <twb> http://en.wikipedia.org/wiki/RAID#Hot_spares
[11:13] <twb> _ruben: I'm just saying that if it weren't for the property that it will spin down (if power management is doing its job), then I don't see the point of a spare drive
[11:14] <twb> I suppose with md raid, a 4-way raid6 might be more computationally intensive than a 3-way + 1-spare raid5
[11:15] <twb> I let our DBA talk me into doing 3-way RAID1 instead of 3-way RAID5 for our new box :-)
[11:15] <kiall> your DBA was probably right for a DB server .. depending on load that is ;)
[11:16] <twb> It's not a DB server
[11:16] <twb> Well, mostly not
[11:16] <kiall> ;)
[11:17] <twb> But the argument is that it's a lot easier to misalign app/filesystem/raid stripes on RAID5 than on RAID1
[11:17] <twb> That, and it'll be easier to upgrade online because I can unplug all but one drive without taking the system offline
[11:17] <eriksson25> Anyone have a good comand to include the spare disk in the array.
[11:18] <_ruben> twb: spinning isn't the only difference, movement of the head(s) could be important as well
[11:18] <_ruben> never really gave it that much thought thou
[11:18] <_ruben> gh
[11:19] <twb> 4TiB SSDs can't come soon enough.
[11:19] <_ruben> + affordable
[11:19]  * kiall chews up another 40 mins of PPA build server time :/ If only I could get this to build locally .. each iteration would take 2 mins.. grr
[11:19] <twb> Right
[11:20] <_ruben> if only RAMSAN would become "cheap" :)
[11:20] <kiall> and persistent ;)
[11:21] <twb> kiall: what are you building, openoffice?
[11:21] <kiall> eucalyptus ... (UEC)
[11:21] <twb> The userland, or a disk image?
[11:22] <twb> _ruben: I don't like them calling it "RAMSAN" if it isn't a SAN
[11:22] <kiall> eucalyptus itself  .. not an image ...
[11:22] <twb> kiall: bummer.
[11:23] <h4p0z3u> INSTANCE i-42320867 emi-E0831073 <public ip> 0.0.0.4 running 0 m1.small 2010-11-10T10:57:49.442Z mycloud eki-F7121103 eri-0B951170, 0.0.0.4 -> <local ip> like 192.168.1.4, how?
[11:23] <kiall> It would literally take 2 mins to rebuild locally .. but it only compiles against a release 10.10 .. do a dist-upgrade and it fails ..
[11:23] <_ruben> twb: agreed
[11:23] <kiall> h4p0z3u, was there a question in that?
[11:23] <h4p0z3u> my VM's start without IP
[11:24] <h4p0z3u> how can I solve it?
[11:24] <kiall> That could be lots of things .. what networking mode are you using? and is this a 1 server setup, or more than 1 server setup?
[11:25] <twb> kiall: why can't you just use pbuilder or sbuild, like a normal dd?
[11:25] <kiall> because im not building an image - im building the stuff that runs on bare metal ;)
[11:25] <h4p0z3u> internet --- server1(public ip) --- LAN --- server2(private ip)
[11:25] <h4p0z3u> server1 is router too
[11:26] <twb> Um, the job of pbuilder/sbuild is to create a chroot that contains (say) a 10.10 compilation environment for dpkg-buildpackage
[11:26] <h4p0z3u> http://imagebin.org/122632
[11:26] <twb> So if you are on, say, 10.04 and need to create a .deb for 8.04, you use pbuilder instead of creating an 8.04 VM
[11:27] <kiall> twb :| now you tell me? Will have to check that out ..
[11:28] <twb> Broadly: sudo pbuilder create --distribution hardy (once), then "pdebuild" instead of "debuild" in the source package
[11:28] <kiall> h4p0z3u, we get that part .. pastebin your eucalyptus.conf from server1 and ifconfig from server2
[11:30] <twb> kiall: the binary and source packages end up in /var/cache/pbuilder/result/
[11:31] <twb> I think sbuild is actually better, but I haven't gotten around to looking at it
[11:31] <kiall> installing it at the mo .. 1 thing i cant find yet ... can you prevent it from doing a clean build each time ;)
[11:31] <twb> I'm waiting for someone to implement btrfs-snapshot-based COWs for it instead of LVM snapshots
[11:31] <twb> kiall: no, the whole point is that it does a predictable, clean build each time.
[11:32] <twb> The manpage seems to indicate there's a way to prevent that, but I don't know about it.
[11:32] <kiall> thats a good feature for doing a release build ... but for dev iterations can be a pain :)
[11:32] <h4p0z3u> kiall, http://pastebin.com/ZU5hdLzf
[11:32] <twb> Nod
[11:32] <twb> I tend to use a mix of "debian/rules build" and "pdebuild"
[11:33] <kiall> twb, yup - but *something* un maverick-updates is preventing the backing from building .. and im just trying to trace a bug!
[11:33] <kiall> in maverick-updates*
[11:34] <kiall> h4p0z3u, on server1 .. what interface is connected to the internet? eth0?
[11:34] <h4p0z3u> yes
[11:34] <kiall> and whats the ip of that 192.168.100.x?
[11:34] <h4p0z3u> node controller
[11:36] <kiall> okay .. 1 issue anyway is that VNET_SUBNET should be something not used anywhere else
[11:36] <kiall> the default is .. VNET_SUBNET="172.19.0.0" (in /etc/euca../euca....local.conf)
[11:37] <eagles051387_> hey guys is there a way to recieve emails if a service goes down?
[11:37] <eagles051387_> or possibly an sms
[11:37] <kiall> your looking for "monitoring" ;)
[11:39] <h4p0z3u> VM's IP and NC IP have to be same subnet?
[11:39] <kiall> h4p0z3u, I cant see any other issues there.. the bridge looks right, the conf looks right (other than the VNET_SUBNET and associated settings .. mine looks like http://pastebin.com/z8yzxuMm where 10.2 is the public range, and 10.8 is the private (CC<->NC) range)
[11:39] <eagles051387_> kiall: ya what can i setup on lucid to do that?
[11:40] <kiall> nagios / cacti / munin / loads more can be setup .. but .. you generally need to set this stuff up somewhere else .. if it goes down .. how is the email supposed to get out?
[11:41] <kiall> there are also companies who will do it for free (if your only monitoring 1 or 2 services..)
[11:41] <eagles051387_> true
[11:41] <eagles051387_> humm
[11:42] <eagles051387_> man i really need to setup a cloud
[11:42] <kiall> h4p0z3u, no - your VNET_SUBNET is wrong there .. it cant be in the same range as server1's eth0 *or* eth1
[11:42] <eagles051387_> has anyone here used the cloud computing feature of ubuntu server in a private environment?
[11:44] <kiall> yes - but ... honestly .. if it was something you actually wanted to use .. you probably wouldnt be asking about monitoring ;)
 ok now i restarted it with boot.img and it works. but problem is that it says: No network interfaces detected. Do you know some solution for this?
[11:44] <eagles051387_> kiall: im considering revamping my entire setup
[11:45] <eagles051387_> kiall: atm just have a single server but considering turning my current desktop into a server and building a new desktop for me
[11:45] <eagles051387_> kiall: question regarding the cloud. does it require a cpu that supports virtualization or can it be run on one that doesnt
[11:45] <twb> niitty: buy better hardware
[11:45] <eagles051387_> niitty: is the network card built onto the motherboard
[11:46] <niitty> yes
[11:46] <eagles051387_> i had that happen to me where the onboard network card was a bit flakey
[11:46] <twb> niitty: what that error means is that your shitty Nvidia/AMD motherboard's NIC is not supported by open-source drivers.
[11:46] <kiall> eagles051387_, the bare min setup is 2 servers, 1 "managing" the cloud, and 1 running the actual VMs .. that second one needs VT
[11:46] <eagles051387_> twb: thats an easy fix
[11:46] <eagles051387_> ok well my quad has that kiall :) ok sweet
[11:46] <eagles051387_> twb: all niitty would need to do is buy a pci nic and throw it in. thats what i did and it worked fine
[11:47] <niitty> i use Asus cheap motherboard.
[11:47] <twb> True
[11:47] <eagles051387_> niitty: you mean asrock <-- low end asus
[11:47] <twb> niitty: asus is the manufacturer.  The chipsets are usually from nvidia or intel
[11:47] <eagles051387_> niitty: i would try with a pci nic card they arent that expensive
[11:48] <h4p0z3u> kiall,I fix but not work
[11:48] <niitty> aah, i have naybe some old network card in garage... so maybe ill try em
[11:48] <eagles051387_> niitty: gigabit nics arent expensive
[11:48] <h4p0z3u> INSTANCE        i-427707BC      emi-E0831073    200.137.66.120  0.0.0.2 running         0               m1.small        2010-11-10T11:46:55.359Z        cloudinf        eki-F7121103    eri-0B951170
[11:49] <twb> I've had consistently good success with e100 and e1000 chipsets.
[11:49] <eagles051387_> twb: who makes those
[11:49] <twb> While they're reputedly very shitty hardware, I've also had mostly good success with rtlNNNN chipsets.
[11:49] <twb> eagles051387_: they're intel chipsets
[11:49] <eagles051387_> and i have had great luck with nvidia's nforce
[11:50] <eagles051387_> twb: maybe for server but desktop users are getting slammed hard again with graphics bugs
[11:50] <twb> eagles051387_: this is #ubuntu-server.
[11:50] <eagles051387_> i know but im saying im apprehensive about intel chipsets
[11:51] <h4p0z3u> kiall, http://pastebin.com/iuu875qt
[11:51] <eagles051387_> desktop/possible server is an nforce 750i chipset that i have on
[11:51] <eagles051387_> kiall: could the server managing the cloud also host some things as well?
[11:56] <kiall> h4p0z3u, that should be fine ... clean restart the lot .. and spark up a new instance .. if it still doesnt work, you still have another issue
[11:59] <twb> kiall: btw, recent pbuilder magically supports ccache.  Dunno if that would help with re-builds of a java thingamy
[12:00] <h4p0z3u> kiall, INSTANCE i-46F00928 emi-E0831073 x.x.x.x 0.0.0.2 running... fail! =(
[12:00] <h4p0z3u> ops
[12:00] <h4p0z3u> now works
[12:00] <h4p0z3u> 0.0.0.2 -> 192.168.122.205
[12:00] <kiall> twb, humm .. interesting :) .. esp since i'm now confident the bug is in libvirt (or kvm .. or qemu)
[12:01] <kiall> 0.0.0.2? that cant be right ;)
[12:01] <kiall> that should be 172.19.1.2
[12:07] <h4p0z3u> kiall, ping <public ip> crash and ping 192.168.122.205 crash
[12:07] <kiall> crash as in it crashes? or just doesnt work?
[12:08] <h4p0z3u> From x.x.x.x icmp_seq=1 Destination Host Unreachable
[12:08] <kiall> if it just doesnt work .. pastebin "euca-describe-groups" from your workstation (aka not server1 or server2)
[12:09] <h4p0z3u> GROUP	h4p0z3u	default	default group
[12:09] <h4p0z3u> PERMISSION	h4p0z3u	default	ALLOWS	tcp	22	22	FROM	CIDR	0.0.0.0/0
[12:09] <kiall> you haven't allowed ICMP pings to it :)
[12:09] <kiall> hence its rejecting them
[12:09] <kiall> try ssh'ing to it
[12:09] <kiall> (since you have allowed ssh)
[12:10] <h4p0z3u> ssh h4p0z3u@x.x.x.x
[12:10] <h4p0z3u> ssh: connect to host x.x.x.x port 22: No route to host
[12:11] <kiall> odd .. this wont fix that .. but you prob want it anyway ... euca-authorize -P icmp -s 0.0.0.0/0 default
[12:11] <kiall> has the VM booted up correctly? euca-get-console-output i-XXXXXXX will show you whats up
[12:12] <h4p0z3u> iptables rules on server1
[12:12] <h4p0z3u> http://pastebin.com/m8VScaBL
[12:13] <jpds> h4p0z3u: Line 13 seems pretty conclusive for your ping 'crash'.
[12:13] <kiall> that looks a but wrong ;)
[12:14] <kiall> you should have some stuff like .. "-A PREROUTING -d 10.2.6.10/32 -j DNAT --to-destination 172.19.1.2"
[12:14] <h4p0z3u> for all 172.19.1.x?
[12:15] <kiall> just for running instances
[12:15] <h4p0z3u> euca-get-console-.... http://pastebin.com/8ktYxE54
[12:15] <kiall> aha .. I love that one ..
[12:16] <kiall> The only fix i've found is to wipe out the emi / eki /eri and publish fresh ones
[12:16] <kiall> the 10.10 ones on uec-images.ubuntu.com work well ..
[12:18] <kiall> most images should have "-----BEGIN SSH HOST KEY FINGERPRINTS----- ........" near the end of that output if it booted successfully
[12:22] <niitty> nice! i find my old nic and it work like a charm. thanks for the tips :)
[12:32] <h4p0z3u> kiall, I have to leave out now, thx for all....
[12:33] <kiall> no problem
[12:49] <eagles0513875|2> hey guys how can i fix this error drwx------ 16 jonathan jonathan     4096 2010-11-10 12:51 Maildir
[12:49] <eagles0513875|2> what group by default does dovecot use?
[12:50] <eagles0513875|2> reason im asking is i get this error jonathan@eagleeyet.net>: maildir delivery failed: create maildir file /home/jonathan/Maildir/tmp/1289392498.P4804.eagle: Permission denied
[12:50] <eagles0513875|2> should the user and the group be dovecot in this case?
[13:06] <serard> I have a problem installing squirrelmail, on the web interface I get this : Error connecting to IMAP server: localhost
[13:06] <serard> 111 : Connection refused
[13:15] <serard> Ok I finally came to make it work. But I can't send an email, it tells me 501 5.1.7 Bad sender address syntax
[13:27] <[diablo]> afternoon all... any Canonical reps around please?
[13:39] <pmatulis> [diablo]: this is a technical channel.  what is your main question?
[13:51] <hallyn> jdstrand: hey, my irc connection has dropped a few times, so just to make sure - did you have any outstanding questions on the libvirt-0.8.5v2 merge req?
[14:35] <[diablo]> pmatulis, sorry was afk... just wanted to know if any canonical reps hang around here cos I have some Landscape questions
[14:36] <pmatulis> [diablo]: i believe there is a #landscape Freenode channel
[14:36] <ScottK> [diablo]: There are people that work for Canonical on this channel, but it's really not on topic.
[14:36] <[diablo]> np
[14:38] <jdstrand> hallyn: I do not. I mentioned I've yet to look at it again, but hope to today
[14:38] <jdstrand> hallyn: the things I had were pretty small, so assuming you addressed them, I'll just upload
[14:51] <hallyn> jdstrand: ok, thanks
[15:35] <tgwoollard> Good afternoon. I have a 32 Core (4x Eight-Core AMD Opteron) HP BL685 G7 Server that i am installing with Ubuntu 9.10. For some reason, the Operating System (and the installer) will only load when i specifiy a "maxcpus=16" or lower, boot parameter. How can i get Ubuntu to see the other 16 cores? Many thanks in advance
[15:39] <manjo>  tgwoollard, any reason you are using 9.10 instead of 10.04 (LTS) or 10.10 ?
[15:41] <manjo> tgwoollard, ProLiant BL685c G7 was ubuntu ready under 10.10
[15:41] <tgwoollard> manjo: 10.04 is of course my preferred OS, and i will 100% be using that release if i can't resolve this issue. I am purely using 9.10 as it is a request from our head of application development as his preffered environment. Whether or not i use 10.04 or not, it's going to really bug me if i don't figure this out now! I guess it's kernel related, but i need a few pointers i guess. Any suggestions?
[15:41] <claude2> tgwoollard: phew, thats quite a problem to have
[15:42] <tgwoollard> FYI 10.04 works perfectly on the BL685 G7 with all Cores. 9.10 is the preffered option for us though so it would be nice to resolve if poss.
[15:43] <pmatulis> tgwoollard: let me guess, 9.10 preferred because of PHP stuff?
[15:44] <tgwoollard> More than likely, but i also think there is simply a reluctance to move away from the preffered OS, in this case 9.10
[15:44] <manjo> tgwoollard, you could use LTS backports and see if that works
[15:45] <manjo> tgwoollard, the backports kernel have patches backported
[15:46] <tgwoollard> manjo, would you care to elaborate? i've never used backports kernel before.
[15:46] <pmatulis> manjo: LTS backports kernel on 9.10???
[15:47] <manjo> pmatulis, ah he said 9.10... I kept thinking 10.04...
[15:47] <manjo> tgwoollard, scratch LTS backports
[15:47] <manjo> idea
[15:47] <tgwoollard> manjo, just to confirm. 10.04 works perfectly. This issue is purely around prior releases, specifically 9.10 in this case
[15:54] <ScottK> tgwoollard: 9.10 goes out of support in ~5 months.  It's probably not too soon to think about migrating forward anyway.
[15:54] <tgwoollard> ScottK; that i did not know. That's certainly a rather large bullet to load my gun with when i push for 10.04. Many Thanks
[15:55] <tgwoollard> This is still going to bug the hell out of me though, so if anyone has any ideas they would be warmly welcomed
[15:55] <ScottK> tgwoollard: You're welcome.  In comparison, 10.04 will be supported for another ~53 months.
[17:05] <tonyyarusso> I know there was some chatter a while back about switching the Ubuntu mysql packages to use the MariaDB source tree.  What is the current status / plan on that front?
[17:18] <ScottK> tonyyarusso: No action AFAIK.
[17:18] <ScottK> IIRC the mariadb people didn't even show up at UDS this time.
[17:18] <tonyyarusso> o rly?  Hrm.
[17:18] <tonyyarusso> No mneptok?
[17:19] <ScottK> If he was there, I managed to avoid him.
[17:24] <zul> tonyyarusso: there is a blueprint that deals with this i forgot which one now
[17:26] <tonyyarusso> zul: looks like https://blueprints.launchpad.net/ubuntu/+spec/server-lucid-mariadb-inclusion
[17:26] <tonyyarusso> Priority undefined, Series none, Assignee none, Definition new, Implementation deferred, Milestone none - not exactly active :P
[17:28] <tonyyarusso> ScottK: "This issue was discussed on the second day of UDS.", so I'd be surprised if mneptok wasn't there for that.
[17:28] <ScottK> tonyyarusso: Look what release that spec is for.
[17:28] <Pici> tonyyarusso: he wasn't physically at UDS at all.
[17:29] <tonyyarusso> oh, right, that would have been a DIFFERENT UDS.
[17:29] <tonyyarusso> Some days I have a brain, and some of those it even works, I swear.
[17:59] <antoniocs> hello!
[17:59] <antoniocs> is it possible to change to another user while I am logged in?
[18:00] <alex88> antoniocs, just execute command as another user?
[18:01] <antoniocs> well I want use the sudo command but my user isn't in the sudo list but I do know the password for a user that is in the sudo list
[18:01] <antoniocs> so I just want to know if it's possible to just switch to that user and perform the sudo
[18:02] <alex88> if you're connected via ssh open another session, if you're in front of the pc press ctrl-alt f2 (or f3-f4-f5..)
[18:02] <alex88> well you have to know the another user password
[18:02] <antoniocs> ok I just thought I could switch to the user I would like
[18:02] <antoniocs> I do I do
[18:02] <alex88> ok
[18:03] <uth> hello
[18:03] <alex88> hi
[18:03] <uth> I would like to know how to use ssh
[18:03] <patdk-wk_> very interesting :)
[18:04] <patdk-wk_> just upgraded a few webservers from rhel 5.5 to ubuntu 10.04
[18:04] <EvilPhoenix> uth:  define  "use ssh"
[18:04] <patdk-wk_> nfs loads on them, for peek weekday traffic, is down to weekend levels :)
[18:04] <alex88> uth, you have to install openssh-server on the server, and on the client run ssh user@Ip_of_the_server
[18:04] <alex88> and you're on the remote console
[18:05] <Seidr> Heya - does anyone have experience with OpenVZ under Ubuntu? I'm having a strange problem with FTP conncetions while using an OpenVZ kernel. This issue was solved a while ago (with a vanilla kernel) by using the FTP helper with IPTABLEs. The very same entries are active at the moment and I've ensure all kernel modules that WERE loaded with the previous kernel are loaded with the current
[18:05] <Seidr> (well, all that should make a difference) however I'm still getting a 'Connection reset' on only certain directories through FTP.
[18:05] <uth> ok alex88
[18:05] <Seidr> Sorry for the wall of text! Does this problem ring any bells with anyone? ;)
[18:05] <uth> but I have to configure files ? for the server ?
[18:06] <alex88> uth, nope.. but it's useful to deny root login
[18:06] <alex88> and for perfect security use key-based auth instead of password..
[18:07] <uth> so every user can connect by ssh ?
[18:07] <alex88> uth, every user of the server machine, yes..
[18:08] <uth> yes
[18:08] <uth> how do I deny root login ?
[18:08] <alex88> there is a line like rootlogin
[18:08] <alex88> in /etc/ssh/sshd_config
[18:09] <uth> ok
[18:09] <alex88> if you want you can create a group and let only users in that group to ssh login
[18:09] <uth> Is there problems with Internet box ?
[18:09] <Seidr> So - no-one else ever run into problems with FTP access while using an OpenVZ kernel? Even the log files are not giving any hints (verbose mode is on)
[18:10] <alex88> Seidr, i haven't that problem
[18:10] <alex88> uth, like?
[18:11] <uth> Should I do a port redirection ?
[18:11] <EvilPhoenix> uth:  only if you want to
[18:11] <EvilPhoenix> uth:  i personally run ssh on a port that is not the standard port
[18:12] <EvilPhoenix> but its up to you :P
[18:12] <uth> ok thanks alex88 and EvilPhoenix
[18:12] <alex88> uth, for restricting users you can use AllowGroups AllowUsers DenyUsers DenyGroups directives
[18:12] <alex88> np
[18:13] <alex88> look at man /etc/ssh/sshd_config
[18:13] <EvilPhoenix> or ask here if you get confused with the manpage :P
[18:14] <uth> ok thanks
[18:26] <mathiaz> hallyn: hi
[18:27] <mathiaz> hallyn: it seems that I'm running into small issues with ksm on maverick
[18:27] <mathiaz> hallyn: my laptop is running maverick
[18:27] <mathiaz> hallyn: and I start vm guest via libvirt
[18:27] <hallyn> mathiaz: if you don't mind,
[18:28] <hallyn> plz ping me again in 30 mins.  i need to get on a call, and suddenly my (maverick) laptop is failing getpt - can't even pull up a terminal!
[18:28] <hallyn> weirdness
[18:28] <hallyn> biab
[18:30] <pmatulis> why is the lucid netboot kernel [1] dated april 2010?  this is old.  am i missing something?
[18:30] <pmatulis> [1] http://archive.ubuntu.com/ubuntu/dists/lucid/main/installer-amd64/current/images/netboot/ubuntu-installer/amd64/
[18:38] <hallyn> mathiaz: ok, i'm back.  what sort of ksm errors are you having?
[18:44] <ScottK> pmatulis: Because that's when Lucid was released.
[18:45] <pmatulis> ScottK: i saw that connection but i figured it would be updated
[18:45]  * jpds hopes everything under dists/lucid/* is frozen forever more.
[18:46] <ScottK> pmatulis: http://archive.ubuntu.com/ubuntu/dists/lucid-updates/main/installer-amd64/current/images/netboot/ubuntu-installer/amd64/
[18:46] <Seidr> Lucid itself isn't - the packages used are
[18:46] <Seidr> provided* rather
[18:46] <pmatulis> ScottK: ah
[19:32] <mathiaz> hallyn: hi - so I run multiple guests on my laptop with libvirt+kvm
[19:32] <mathiaz> hallyn: and when one of the guest is powered off, the other running guests froze for 5-10 seconds
[19:32] <mathiaz> hallyn: on the host the ksm process is taking a lot of cpu
[19:33] <mathiaz> hallyn: once ksm has done its job running guests are responsive again
[19:34] <mathiaz> hallyn: so ksm may require some tweak - so that one guest going down doesn't freeze other guests
[19:35] <hallyn> mathiaz: hm.  that happens every time you shut one down?
[19:35] <mathiaz> hallyn: yes
[19:35] <mathiaz> hallyn: AFAICT
[19:35] <RoyK> mathiaz: how is the I/O load on the host when this happens?
[19:36] <mathiaz> RoyK: I haven't measure that
[19:36] <RoyK> mathiaz: try
[19:36] <hallyn> mathiaz: can you file a bug (against both kvm and kernel), and i'll try to reproduce
[19:36] <mathiaz> hallyn: great - thanks
[19:36] <hallyn> ksm is just black magic to me :)
[19:36] <RoyK> mathiaz: run sar 1 1000 or something and stop a guest - that'll show the load per second
[19:36] <hallyn> biam
[19:56] <burhan> what would I install if I wanted to run a newsgroup *server* - that caches locally newsgroups which I can access using common news clients.
[19:56] <RoyK> ye gods - we have three 10kVA UPSes - ok - only two of them are clustered as of now, for reasons old, so we called the company supporting them, and asked them to help us clustering the three of them, since we have to power down the whole datacentre now, and they told us they could help us some time next year and then they needed 5+ hours during day-time, between 8 and 16, on a workday, to fix it...
[19:57] <RoyK> burhan: some nntp server
[19:57] <burhan> RoyK: got a name I can google? Everything I try leads me to news readers.
[20:01] <RoyK> burhan: sorry - not sure
[20:05] <RoyK> burhan: apt-get install inn2
[20:05] <RoyK> http://www.eyrie.org/~eagle/faqs/inn.html
[20:05] <burhan> isn't that a reader?
[20:06] <RoyK> inn2 - 'InterNetNews' news server
[20:06] <RoyK> from apt-cache
[20:08] <mathiaz> hallyn: bug 673705
[20:10] <RoyK> mathiaz: did you check the i/o load?
[20:10] <mathiaz> RoyK: I ran sar - the output of the file is attached to the bug
[20:15] <RoyK> mathiaz: the sar.output file is a binary fila
[20:15] <RoyK> file
[20:16] <mathiaz> RoyK: yes - and you can read it with "sar -f filename"
[20:17] <RoyK> mathiaz: that's not good - as I said, start "sar 1 100" - shutdown the guest and pastebin the sar output
[20:18] <RoyK> the sar file normally only holds 10 minute averages
[20:18] <mathiaz> RoyK: not the one I've posted - it has a 1 second average
[20:19] <RoyK> mathiaz: I tried sar -f sar.output - it gave me nothing
[20:20] <RoyK> mathiaz: just post text - that works for all
[20:39] <hallyn> mathiaz: you're sure that the whole guests freeze, and not just networking to them, right?
[21:01] <KB1JWQ> Hmm.  Not sure what the other admin did to this box, but nothing starts by default now.  What's the command within Ubuntu to start services on boot?
[21:02] <KB1JWQ> update-rc complains about upstart.
[21:02] <tonyyarusso> maybe just making the symlinks directly?  hrm
[21:02] <guntbert> KB1JWQ: it used to be insserv but I always forget what works when :-)
[21:03] <KB1JWQ> tonyyarusso: That's what's odd, the symlinks should exist in /etc/rc3.d, correct?
[21:03] <tonyyarusso> insserv appears to still exist
[21:03] <tonyyarusso> KB1JWQ: Ubuntu's default runlevel is 2
[21:03] <KB1JWQ> Hmm, "runlevel" returns "unknown."
[21:03] <KB1JWQ> That'd explain it.
[21:03] <guntbert> !runlevel | KB1JWQ
[21:04] <KB1JWQ> Hey, after reboot the ONLY processes running are init, my tty, and bash.
[21:04] <KB1JWQ> Hothing else.
[21:08] <guntbert> KB1JWQ: try insserv -d (for enabling default services)
[21:08] <jeiworth> say, what command can i use to check what tme and date a certain server has?
[21:09] <KB1JWQ> guntbert: Hmm, http://pastebin.com/xjCPsH0V
[21:09] <KB1JWQ> jeiworth: date
[21:09] <jeiworth> KB1JWQ, sorry, not specific enough, i mean a _remote_ server, don't have login to that
[21:10] <KB1JWQ> Is it running an ntp server?
[21:10] <guntbert> KB1JWQ: strange indeed - might be the other removed "everything"?
[21:11] <KB1JWQ> Yeah, I'm not thrilled. :-)
[21:11] <jeiworth> no, it's a vm i have here for a cert we are doing, of course i could just boot it up with another live system and change root pw but surely there must be a way to check what is the current time and date of a server from remote?
[21:12] <guntbert> KB1JWQ: maybe try with aptitude update, aptitude safe-upgrade ??  (I'm only guessing)
[21:12] <KB1JWQ> guntbert: If I could get networking up that'd be swell. :-)
[21:13] <guntbert> KB1JWQ: ouch, I overlooked that -- is the server important?
[21:13] <KB1JWQ> guntbert: Not particularly.
[21:14] <guntbert> jeiworth: there is another time protocol as well, point nmap to the server to see if it listens on that port (don't remmeber
[21:14] <guntbert> I don't remember its name
[21:15] <ejv> can anyone assist me in backing up a GPT partiton table? :)
[21:15] <guntbert> KB1JWQ: I'm not sure if it is worth the effort to rebuild all - what about you ditch it?
[21:16] <bluefrog> guntbert, jeiworth nmap -sU -p123
[21:16] <jeiworth> guntbert, kk, well
[21:16] <KB1JWQ> guntbert: Not a bad idea.
[21:16] <guntbert> bluefrog: thx
[21:17] <jeiworth> well, thx, but that machine is not running an ntp instance
[21:17] <bluefrog> yaeh sorry just read that
[21:18] <bluefrog> jeiworth, well no ntp and no remote login guess u're stuck
[21:18] <bluefrog> webserver on it?
[21:18] <bluefrog> or email?
[21:20] <guntbert> jeiworth: bluefrog: I was thinking of DAYTIME (port 13) and time protocol(port 37)
[21:20] <guntbert> they don't need a ntp server
[21:20] <bluefrog> yes maybe sry out oy my league at this point
[21:21] <jeiworth> hmmm i could try to have it send me a test email, good thinking bluefrog
[21:22] <bluefrog> that's more of what I was thinking yes
[21:24] <jeiworth> Date: Tue, 9 Nov 2010 15:51:17 -0600 (CST) <-- lol, well that explains a lot
[21:24] <bluefrog> email then?
[21:25] <jeiworth> bluefrog, jupp
[21:26] <jeiworth> the system has a webfrontend where you can confiugre outgoing mail for notification and a button to test the config
[21:26] <jeiworth> 1d and half an hour off
[21:26] <jeiworth> tztztz
[21:27] <jeiworth> strange though, i am pretty sure there is a command that queries the current time and date of a remote machine... hmmm
[21:29] <apw> hallyn, which chroot are you using for your natty build ?
[21:30] <bluefrog> jeiworth, this? http://www.logix.cz/michal/devel/nagios/
[21:30] <bluefrog> but you need  to have tweaked the server before hand
[21:31] <jeiworth> bluefrog, well, that obviously, but kind of overkill for my use case
[21:32] <jeiworth> well, passed the exam anyway, now for some beers... hmmm
[21:40] <hallyn> apw: sbuild...
[21:40] <hallyn> apw: is that what you are asking?
[21:42] <apw> hallyn, hmmm, not quite sure, i presume that makes its own chroot on the fly as there arn't any for natty yet :)
[21:45] <hallyn> apw: no, i think i created a maverick one, then substituted in /etc/apt/sources.list
[21:47] <apw> hallyn, ahh, the shared ones should be back in a few hours, they are somewhat slow to buidl
[21:47] <hallyn> there are shared ones?
[21:47] <apw> /usr3/chroots yep, all exported under schroot
[21:48] <hallyn> on tangerine?
[21:48] <apw> but perhaps they arn't in the right format or something
[21:48] <apw> yep, on there, though they are building right now cause we wiped the machine today
[21:48] <apw> maverick is not as yet there
[21:48] <hallyn> ah, i see them.  cool
[21:50] <hdon> OT question: does anyone know a pastebin that can strip trailing whitespace? SCREEN always causes my copy/paste to include about 300 trailing spaces
[21:50] <hdon> and the pastebins i've tried do not reflect the trailing spaces well
[21:50] <hallyn> apw: sad to say, i hadn't even considered that maybe i need to build my natty kernel on tangerine in a natty chroot :)
[21:50] <hallyn> maybe that's why it's failing
[21:54] <apw> hallyn, heh, not sure i know of any natty specific requirements, think i am still building some of my test builds in maverick chroots myself
[22:02] <kuick> Hi guys, anyone available to give me a quick run through on my boot setup. I had all working great then I did an update and could not boot again. Have spent many hours on this but now I need some input. TX
[22:04] <kuick> Using grub as boot loader
[22:05] <kuick> On Lucid
[22:08] <ehcah> Can anyone point me to a good book on administering Ubuntu Server for file storage?
[22:25] <hallyn> apw: heh, well after the rebuild, my kernel binary-debs build just fine.  huh  :)
[22:29] <mathiaz> hallyn: hm - seems to be network related
[22:29] <mathiaz> hallyn: I've updated the bug 673705
[22:36] <hallyn> mathiaz: then it sounds like a dup
[22:36] <hallyn> which is great, bc i've never been able to reproduce, so maybe you can help finally nail htat one :)
[22:36] <mathiaz> hallyn: great - I can reliably reproduce it
[22:37] <hallyn> mathiaz: ok, i'm going to review the old bugs in a bit, and try and come up with a plan for bugging you tomorrow :)
[22:37] <hallyn> thanks
[22:37] <hallyn> bbl
[22:37] <mathiaz> hallyn: o^5
[22:37] <hallyn> ?
[22:37] <hallyn> is that one of your handshakes? :)
[22:45] <wideyes> hello all! Anyone interested in giving a little dhcp setup help or directing me to a good channel?
[22:47] <RoyK> wideyes: pastebin your dhcp conf
[22:48] <RoyK> wideyes: dhcp setup is quite easy....
[22:48] <mathiaz> hallyn: it's a high-5 - o^5! ;)
[22:52] <wideyes> I'm looking for some ltsp setup help, specifically
[22:53] <wideyes> I'd be happy to pastebin the config, but maybe we could talk about the ltsp setup first
[22:53] <wideyes> the issue is, I'm using one server for dhcp, and a separate one for ltsp
[22:54] <wideyes> I can't get the dhcp server to correctly direct my thin clients to the ltsp server
[22:54] <wideyes> I think I'm misusing the next-server directive
[22:55] <wideyes> the ltsp server is ubuntu 10.04
[22:55] <wideyes> but the dhcp server is debian sarge 3.1
[22:57] <hggdh> soren: I have some changes to the autobuilder for regression tests, will propose a merge in a few
[23:09] <soren> hggdh: I won't be upset if you just apply them, but if you want a review, that's cool too.
[23:09] <hggdh> soren: I do not think I have write access to the bzr :-)
[23:09] <soren> hggdh: You do.
[23:10] <soren> hggdh: Make the changes on the server, do bzr commit, bzr push, win.
[23:10] <hggdh> oh
[23:10] <hggdh> OK
[23:10] <soren> I'm happy to review it, if you feel it's necessary, but really, just go ahead if you want.
[23:10]  * hggdh was considering a local (own laptop) branch ;-)
[23:11] <hggdh> soren: OK. What I intend to do: make it run against both Maverick and Natty
[23:11] <soren> hggdh: Hm.. Ok.
[23:12] <hggdh> and, eventually, also against Lucid (to be considered, yet)
[23:12] <soren> hggdh: Ok, I can see that would be valuable.