[00:48] <jfig1234> hi all, having a problem here at home, I have a server running for months without problem, now without any change (except for upgrades to ubuntu) the server only works for a few minutes connected to a specific network switch and then stalls. connector to another switch seems to work to. other equipment (freeNAS, router, etc) works fine
[01:02] <twb> jfig1234: do you have physical access to the host?
[01:04] <jfig1234> twb: yes
[01:04] <twb> jfig1234: is the problem currently occurring?
[01:06] <jfig1234> twb: no, but i can easly re-plug the server on the switch
[01:08] <twb> jfig1234: you want to collect information like "dpkg -l network-manager", "ethtool", "ip addr" and "ip route" both before and after the problem occurs.
[01:08] <twb> jfig1234: you may also want to use tshark or tcpdump to perform packet sniffing
[01:09] <twb> It may just be a switch or cable is dead, and needs to be replaced.
[01:10] <jfig1234> thanks, i'll give those a try, has for ports, cables, et all i've tried several... and oother equipment on that switch works fine
[01:11] <jfig1234> since its a test version i'll also try and install a older version of ubuntu and see if it works again
[01:13] <twb> I do not think a different release will change anything.
[01:18] <cef> anyone used the 'rescue mode' on the server disk? why is it asking me about a root fs? the root fs on this machine is trashed and the disks are going to fail. All I want to do is boot up a minimal install and copy the files off the machine, preferably across the network.
[01:19] <jfig1234> twb: i dont see how either but this has gotten me very bugged... same server works on one switch and not on another, switch thats causing the problems works fine with other equipment... that server was working for several months - on that switch - without and problem.....
[01:19] <twb> I assume you have rebooted the switch
[01:19] <jfig1234> :) several time
[01:19] <twb> cef: the rescue target will not help you there
[01:20] <cef> twb: yeah.. and there isn't even an ftp client on the thing.. grr
[01:20] <twb> cef: at least, it'll be far less hassle to find generic live media that includes lvm2/mdadm/whatever you need.
[01:20] <cef> or scp
[01:20] <twb> cef: you need to request sshd specifically
[01:21] <twb> cef: it will appear as a checkbox if you boot with expert
[01:21] <cef> twb: in the rescue? ahh ok.. I didn't boot  with expert on the rescue mode
[01:21] <`jpg> jfig1234: Clear the cache on the switch if it doesn't on reboot?
[01:22] <twb> cef: try "rescue priority=low"
[01:22] <twb> cef: but it'd be easier to just use a knoppix or centos live cd
[01:22] <cef> just frustrating that the rescue mode is next to useless if the root fs is trashed, which, imho is one of the reasons you need a rescue disk
[01:22] <twb> The ubuntu ones won't do because they don't include mdadm and lvm2 out of the box, and it's a PITA to install them
[01:22] <cef> don't need either.. hardware raid
[01:23] <twb> cef: I don't think rescue mode is intended to be a full-featured rescue system
[01:23] <cef> but it's an older machine so any livecd that has a gui is going to run like a dog
[01:24] <twb> cef: both centos and knoppix live CDs do not force you to run a GUI
[01:24] <twb> This is another reason they are better than ubuntu's live CD for rescue purposes.
[01:30] <`jpg> cef: Also take a look at hirens boot cd etc, there are lots of console only rescue cds out there that do a great job.
[01:30] <twb> Yu
[01:30] <twb> *Yup
[01:31] <twb> I sometimes roll my own using Debian's live-helper
[01:31] <cef> `jpg: ta
[01:31] <`jpg> I am wanting to build a high availablity cloud using Ubuntu UEC. Has anyone deployed a cloud that can advice in making the cluster controller more redundant.
[01:31] <`jpg> *advise on
[01:38] <cef> well can't find anything at short notice that has the mylex DAC960 driver in it, so trying expert mode just in case
[01:41] <twb> LSI make real hardware raid?
[01:41] <cef> ahh openssh-client-udeb option :D
[01:41] <cef> Mylex, not LSI
[01:42] <cef> (at least, in this system)
[01:42] <twb> "September 2002 - LSI acquired Mylex from IBM."
[01:42] <cef> well, this IS an old machine
[01:42] <twb> Shrug
[01:42] <cef> and this is a Mylex card
[01:44] <cef> and hey, it works. :D
[01:47] <marcelcohrs> hey geys
[01:47] <cef> twb: ta for the help btw re: expert mode on the rescue disk
[01:47] <twb> cef: no problem
[01:48] <twb> cef: the reference is [...]/dists/hardy/main/installer-i386/current/images/netboot/pxelinux.cfg/default
[01:48] <twb> Or just isolinux/isolinux.cfg if you have the CD in front of your :-)
[02:03] <duvnell1> hi, I've livebooted a ubuntu 9.10 CD on a dell, when I plug in my usb external HDD, it's saying a device 2-1 was plugged in, but it is not creating a device.  Any ideas?  should I just get the newest version and hope for the best, or is there something that I can do to prod it alone to create a device?
[02:06] <twb> duvnell1: we can't help you with the live CD, since it's effectively ubuntu *desktop*.
[02:07] <`jpg> Never seems to be any Cloud experts about. :(
[02:08] <twb> `jpg: is there a cloud-specific channel?
[02:08] <twb> e.g. #eucalyptus
[02:10] <`jpg> Hmm..will check, cheers twb.
[03:31] <oh_noes> is there an inittab on hardy server?
[03:31] <cef> is there much different from the server install and the desktop install (apart from default server apps being installed and a gui on the desktop).. just the kernel, isn't it?
[03:31] <cef> or is there base differences in the core somewhere?
[03:33] <twb> oh_noes: upstart only understands a strict subset of inittab
[03:33] <twb> oh_noes: and hardy runs upstart, IIRC
[03:33] <oh_noes> oh thx, didnt realized it was upstart.
[03:33] <twb> oh_noes: most of what would be inittab is in /etc/event.d
[03:34] <ScottK> oh_noes: Dapper was the last Ubuntu release with inittab.
[03:35] <twb> With a default one, that is :-)
[03:36] <oh_noes> thanks guys, looks like upstart has an easier config than inittab too
[03:36]  * cef realises that if he wants to boot off a cdrom, it might be useful to have one in the machine.. oops!
[03:36] <oh_noes> The reason I ask is, I have an OS Appliance where I want to create a "Press F1 to do something" at the console CLI login
[03:37] <oh_noes> so I think (and please tell me if im wrong), modify /etc/event.d/tty1 and use a different getty?
[03:38] <oh_noes> I wonder if any getty alternatives have a 'create a keypess script' so when a keycombo is pressed it just runs a script
[04:05] <maxagaz> is there a tool that can compare the content of files
[04:05] <maxagaz> ?
[04:05] <maxagaz> between two servers
[04:05] <maxagaz> in order to synchronize it
[04:06] <maxagaz> like comparing configuration files
[04:09] <pmatulis> maxagaz: you want one file to become like the other?
[04:09] <pmatulis> maxagaz: or give each file what is missing from the other?
[04:12] <maxagaz> pmatulis, i want some config files to have some common content
[04:13] <maxagaz> pmatulis, but keeping also some special content too
[04:13] <maxagaz> i hear about something like ogeas...
[04:13] <pmatulis> !info ogeas
[04:14] <ScottK> augeaus I think.
[04:15] <pmatulis> !info augeaus
[04:16] <ajmitch> !info augeas
[04:16] <ajmitch> bah :)
[04:17] <ajmitch> it's augeas-tools, augeas-lenses
[04:17] <pmatulis> !info augeas-tools
[04:18] <pmatulis> maxagaz: go for it
[04:22] <maxagaz> thanks!
[04:27] <marks256> when i try to use rsync via ssh to back up to a module, rsync errors out with rsync: connection unexpectedly closed (0 bytes received so far) [sender]
[04:27] <marks256> rsync error: error in rsync protocol data stream (code 12) at io.c(600) [sender=3.0.5], but it works fine if i use a direct path. what gives?
[04:33] <tbielawa> soren: did you have any difficulty merging jaunty's libvirt from debian? I'm running into libtool errors but I see no mention of it in the changelogs.
[04:34] <tbielawa> i backported the patch from libvirt 0.6.5 to fix Bug #368084, I just can't get it to build.
[04:35] <tbielawa> perhaps I should poke around in motu while he's away, they may be more recent on libtooling issues
[04:42] <hackeron> I see error messages like: "Dec  8 12:58:12 AmurgDVR udevd-work[598]: pipe failed: Too many open files" and "unable to create db file '/dev/.udev/db/block:sda': Too many open files" - even though there are only ~3k open files according to lsof? - any ideas?
[04:43] <tbielawa> check out `cat /proc/sys/fs/file-max`
[04:44] <hackeron> tbielawa: 96570
[04:46] <tbielawa> not sure, but I'm reading this. http://www.netadmintools.com/art295.html it mentions the difference between lsof and /proc/sys/fs/file-nr
[04:46] <jmarsden> hackeron: ulimit -n   is another way max open files can be restricted
[04:47] <hackeron> jmarsden: ah, ok, that shows 1024 and udevd has 1075 open -- hmmm?
[04:48] <jmarsden> udevd may have been started from an environment with a higher limit.  Also check /etc/security/limits.conf
[04:48] <hackeron> http://dpaste.com/131078/ < what's happening there?
[04:49] <hackeron> jmarsden: that file has nothing but comments - this is a vanille ubuntu karmic
[04:49] <jmarsden> OK.  Your pastie... I'm not sure what that means, looks almost as though udev is finding the drive, losing it, finding it again, losing it... ??
[04:50] <hackeron> same thing on a few other ubuntu karmic servers I have
[04:50] <hackeron> is udevd broken in karmic? -- what does your lsof -n | grep udev | wc -l show?
[04:51] <jmarsden> 37 here on Karmic desktop amd64.
[04:51] <hackeron> mine's a karmic server amd64
[04:52] <jmarsden> But of those, most are GNOME-related
[04:52] <jmarsden> lsof -n | grep udevd | wc -l    # shows only 4
[04:52] <hackeron> # lsof -n | grep udevd | wc -l
[04:53] <jmarsden> I can boot  a Karmic server VM and try it there too...
[04:53] <hackeron> 1073
[04:53] <jmarsden> Wow... that's... a lot :)
[04:53] <jmarsden> Have you modified the udev rules in any way from the defaults?
[04:53] <hackeron> everything seems to work just fine, but I get occational messages in syslog that udevd pipe failed: too many open files
[04:53] <hackeron> nope, everything is completely vanilla
[04:54] <jmarsden> Then that looks like a bug of some sort to me.
[04:54] <hackeron> only thing I changed is the apache config and the flag not to prompt on fsck during boot in /etc/default/rcS
[04:56] <jmarsden> I get 55 in a freshly booted Ubuntu Karmic Server VM, so there is something different about udevd action on servers, it would seem
[04:57] <jmarsden> And it seems to be slowly increasing over time... 57 now...
[04:57] <jmarsden> Have you searched Launchpad for relevant bugs?
[04:58] <hackeron> hmm, this looks relevant < https://bugs.launchpad.net/ubuntu/+source/udev/+bug/463347
[04:59] <hackeron> I'll try to distupgrade - haven't done so in a couple of weeks
[05:00] <jmarsden> It's not clear that fix is in karmic-updates yet, at least not from the bug report... you might have to grab it from karmic-proposed
[05:01] <hackeron> jmarsden: how do I do that?
[05:01] <KismetGFX> does it matter how i partition my stuff with 3 hard drives cant seem to get it to boot. as in /boot /usr /var swap on one drive, root and swap on 2nd, home and swap on 3rd had it workin at one point rebooted and it was borked
[05:01] <jtaji> KismetGFX: sata drives? using /dev/names or UUIDs?
[05:02] <jtaji> in your fstab
[05:02] <jmarsden> KismetGFX: Or, if you have a machine you can experiment with, first try a single drive default install, let the installer partition it for you and see if *that* boots OK?
[05:02] <KismetGFX> jtaji:  one ide two sata i didnt manually specify a /dev
[05:03] <jtaji> KismetGFX: no I mean are your drives listed as, e.g. /dev/sda1 or UUID=xxxxxxxxxxxxxxxxx?
[05:03] <jmarsden> hackeron: Correcting myself, udev 147~-6.1 is in karmic-updates, so that should take care of your issue.
[05:04] <KismetGFX> jtaji: ya sda1 sdb and sdc each partition tried moving /boot to each drive didnt make any difference huhu
[05:04] <hackeron> jmarsden: ok, trying :) - been bugging me since beta karmic, every 4 hours I get emails from syslog with this udev error - hope this fixes it!
[05:04] <KismetGFX> er sda
[05:04] <jtaji> KismetGFX: problem is sata drives can change order, best thing is to use UUIDs
[05:04] <jtaji> !uuid | KismetGFX
[05:05] <KismetGFX> jtaji: kewl thanx i thought it was a MBR issue i was doing fdisk /mbr's and fixmbr i'll check that out
[05:19] <oh_noes> is it possible to make upstart set tty2 to automatically login instead of displaying a LoginPrompt
[05:20] <oh_noes> ie. run /sbin/sulogin instead of /bin/login
[05:21] <jmarsden> oh_noes: edit /etc/init/tty2.conf to suit your peculiar needs :)
[05:27] <oh_noes> can;t ... im on hardy which is upstart
[05:28] <oh_noes> I'm playing around with it now ...
[05:40] <holiday> Hello, is it possible to configure ssh to use password authentication for a single user, while the others use public key authentication?
[05:41] <jtaji> and the others can't use password?
[05:42] <holiday> yes, the others cant use password
[05:42] <jmarsden> Create keypairs for all the others, passwd -l all the others ... seems weird, but should work??
[05:42] <jtaji> it should
[05:47] <holiday> okay, would you recommend just forcing everyone to use a public key?
[05:48] <holiday> it worked btw thanks
[05:48] <jmarsden> No problem.  Why are you forcing things?  Who is the special person... if he determines your salary, do not force him :)
[05:58] <jmarsden> oh_noes: Did you get the sulogin on tty2 idea to work?
[06:03] <oh_noes> still looking at it now ..
[06:04] <oh_noes> not having any luck, hardy uses upstart so Im modifying /etc/event.d/tty2
[06:05] <oh_noes> exec /sbin/getty -l /sbin/sulogin 38400 tty2
[06:05] <oh_noes> doesnt seem to work which is kinda wierd.  You would expect it to.
[06:05] <jmarsden> Sounds like it should, yes.
[06:10] <jtaji> oh_noes: do you have an /etc/init/tty1.conf ?
[06:10] <jtaji> tty2.conf
[06:11] <jmarsden> jtaji: On Hardy it is all in /etc/event.d/ instead
[06:11] <jtaji> oh ok
[06:14] <oh_noes> i'll pastebin my /etc/event.d/tty2
[06:14] <oh_noes> http://pastebin.com/m6704fcbe
[06:15] <oh_noes> when I go to tty2 ... it still displays /bin/login .... which is weird.
[06:15] <jmarsden> oh_noes: This is after a reboot to make sure you picked up the changes, right?
[06:16] <oh_noes> of course
[06:19] <jmarsden> Could you just exec a program that will output something, instead of getty, there, as a test?  Like top or whatever?  Just thinking out loud, I have not messed with getty and ttys for a very very long time...
[06:22] <oh_noes> yeah same ... im trying all those things now.
[06:25] <oh_noes> naa you cant exec straight into top.   Which kinda makes sense, it's getty job to handle all character displays in/out (I think)
[06:32] <cef> oh_noes: tried passing the standard getty the -l option?
[06:35] <cef> nah ignore that
[06:59] <oh_noes> cef: even with -l it's still displaying a login prompt.
[06:59] <jmarsden> oh_noes: He already said <cef> nah ignore that
[07:00] <cef> btw: -l option of getty runs the program with the 'username' as a command line option afaik
[07:03] <jmarsden> oh_noes: Is the -n option any use to you?  Don't prompt for a login name?
[07:03] <jmarsden> In combination with -l ...
[07:05] <oh_noes> cef: i thought -l was to "use a different login program"
[07:07] <oh_noes> Trying -n -l /sbin/sulogin now
[07:08] <cef> note: if it does, it runs as rooot
[07:08] <cef> err root
[07:10] <jmarsden> oh_noes: For me on Karmic, editing /etc/init/tty2 to do  exec /sbin/getty -8 -n -l /usr/bin/top 38400 tty2     # works
[07:10] <jmarsden> Major security risk, but it works :)
[07:12] <oh_noes> OH NICE, yep -n -l works :D
[07:12] <oh_noes> well -n -l /any/bin works
[07:12] <oh_noes> thanks jmarsden, i knew it was missing a single option
[07:12] <jmarsden> You're welcome.
[08:03] <|saturn|> åñòü êòî-íèáóäü ðóññêîÿçû÷íûé?
[08:05] <jmarsden> |saturn|: Wrong language, unfortunately I can't figure out which one :)  This channel uses English.
[08:06] <jmarsden> !cn
[09:18] <maxagaz> how to reload /etc/fstab once it has been changed ?
[09:18] <Jeeves_> maxagaz: you should mount / unmount stuff
[09:18] <Jeeves_> maxagaz: What did you change?
[09:22] <maxagaz> Jeeves_, I changed the uuid of the swap
[09:33] <Jeeves_> Ah, than you need to run swapoff and swappon
[09:44] <maxagaz> Jeeves_, thanks!
[09:45] <Jeeves_> np
[09:55] <maxagaz> how is set the uuid ?
[09:56] <maxagaz> where is it taken or generated from ?
[10:18] <gamla_kossan> hi people!
[10:19] <gamla_kossan> where can I find a log containing a log for the latest packages installed?
[10:20] <arj> dpkg.log?
[11:34] <ghostlines> anyone know any howto's for installing xen on ubuntu karmic?
[11:36] <jpds> ghostlines: I don't think you can.
[11:37] <jpds> Well, you could technically, but I think support for it was dropped after Hardy for the preference of KVM.
[11:38] <Daviey> it works, but i'd rather pull my eyes out.
[11:43] <ghostlines> lol
[12:01] <gamla_kossan> arj: sweet, tahnks =)
[12:41] <tolbrino> Hi, I'm desperately trying to configure dhcpd to ignore a range of mac addresses without success so far. See http://gist.github.com/252448 .But the dhcpd still gives out IPs to machines with mac addresses starting with "d0:0d". Any hints on what to change in that file? Thanks
[12:54] <sponzor> ok i m installing ubuntu server, but i have problems with partition.. i want to do this.. http://pastebin.com/m8365263 but it is text mode install so i dont really know what to do :P is there any howto for advenced partition?
[13:00] <sponzor> anyone?
[14:00] <ttx> Team meeting on #ubuntu-meeting
[14:05] <smoser> soren, or others interested, i'm looking at bug 494185.
[14:05] <smoser> do you think it is reasonable to
[14:05] <smoser> a.) check if <region>.ec2.archive.ubuntu.com exists
[14:05] <smoser> b.) verify that http appears up on that host
[14:05] <smoser> maybe 'b' isn't necessary
[14:13] <zul> shouldnt it already be doing a?
[14:13] <smoser> it is not doing a right now. it just picks a, not checking the host.
[14:14] <smoser> and is busted (code is pasted into that bug)
[14:15] <zul> ah ok
[14:15] <smoser> ttx, just for the record, i can't do that right? i have to ask an iso team person ?
[14:16] <smoser> zul, above, i was asking about 'b' because if it tried that now and quickly timed out, we'd be ok
[14:16] <ttx> smoser: i can't do that either.
[14:16] <smoser> i know only of slangesek. are there others i can ping?
[14:16] <zul> smoser: ping ara i think
[14:17] <zul> smoser: yeah that would work also a mirror in us-west-1 would work as well
[14:17] <smoser> well right now a mirror in us-west-1 wouldn't fixt he problem.
[14:17] <smoser> because it picks 'us.ec2.archive'
[14:18] <smoser> anyway. i was mainly afraid of "checking" and having that check hang or cause other general mishap
[14:20] <zul> yeah that would have to be done very carefully if that was done
[14:29] <soren> smoser: I've talked to one of the "other cloud providers" about kernels. Everyone offers Ubuntu, but with all sorts of odd kernels. I would like for them to use our kernels so that security updates come from us in a timely manner and everything is just more straightforward. Because they offer lots of other distros and want to use the same kernels everywhere, they compile everything into the kernel to avoid having to shove modules into all their im
[14:29] <soren> smoser: If they were to use our kernels, they would have to follow the same approach: No modules, everything built in.
[14:30] <smoser> 100% "no modules" is not reasonable or possible
[14:31] <illuminai> hello there
[14:31] <smoser> and realistically, i dont think we can promise that our kernels will run everyone elses user space perfectly. i'm certain the kernel team isn't willing to sign up to support that.
[14:31] <smoser> well, i didn't ask them, but..
[14:32] <illuminai> i've a strange problem on my server (ubuntu 9.04, 2.6.28-16-server)
[14:32] <illuminai> after swapping drive /dev/sdb fdisk says: Unable to read /dev/sdb
[14:32] <illuminai> but dmesg |grep SCSI
[14:32] <illuminai> says it is there, [    3.528223] sd 3:0:0:0: [sdb] Attached SCSI disk
[14:32] <soren> smoser: I'm not suggesting we explicitly volunteer to support everyone's user space.
[14:33] <illuminai> also, i'm using software raid 1
[14:33] <smoser> right, but building everything in just isn't reasonable. there are loads of kernel modules, iirc some dont even allow a 'Y' (that could be false)
[14:33] <illuminai> any clues what could it be?
[14:35] <soren> smoser: I know there's a lot of them. Luckily, only a subset are reasonably needed in a cloud context.
[14:35] <smoser> well, some that are reasonable are not reasonable to build in
[14:35] <smoser> i'm thinking fuse
[14:36] <soren> Why not?
[14:36] <soren> (fuse, specifically)
[14:36] <smoser> i dont know... just sdoesn't seem reasonabl e to me. one thing i would think is not so nice about everything built in is that it then can't be switched out.
[14:37] <smoser> but i coudl be wrong about that too
[14:37] <smoser> ie, i would think that if fuse is built into kernel than i can't service it with a kernel module upgrade
[14:37] <smoser> that may not be that big of a deal
[14:38] <soren> Right, they'd have to upgrade the kernel images for all updates, even the ones that don't break ABI.
[14:39] <smoser> i was thinking even on a lower level, as in a user wants to build a more functional version of a kernel module in their instance
[14:39] <smoser> and use it , and maintain it.
[14:39] <smoser> lots of pepole do that for the ec2 provided kernels (xfs) for example.
[14:56] <soren> smoser: Eh? Why?
[14:57] <smoser> because the amazon kernels a.) dont get serviced frequently b.) dont have enough stuff built in or as modules
[14:58] <smoser> the idea of "build more into the kernel" generally seems perpendicular to where things are going
[14:59] <smoser> things needed for boot i woudl consider a different category than "everything else"
[15:00] <erichammond> For some reason we also needed to build the latest fuse kernel module from source when using the Amazon kernels: http://groups.google.com/group/ec2ubuntu/web/compiling-2-6-21-kernel-modules-from-source-for-amazon-ec2
[15:04] <erichammond> Well, the sun's up, so I'm heading to bed.
[15:05] <RoyK> erichammond: bah - it's 4pm and the sun set an hour ago
[15:06] <erichammond> I just realized I'm wrapping up my 3rd decade programming in January.  That's 30 years.
[15:17] <savid> I'm a bit new to administering ubuntu servers -- I have a production server for a website & database running ubuntu 8.10  Do server admins usually do dist-upgrades on servers, or just keep the current distro as long as possible?
[15:18] <savid> The reason I ask is because I have rkhunter installed, and it is complaining about GPG, OpenSSL, and OpenSSH being outdated
[15:22] <alvin> savid: dist-upgrade will not update the release, only the installed packages. It is pretty safe.
[15:22] <savid> alvin, ok, thanks :)
[15:24] <alvin> savid: $ sudo do-release-upgrade will upgrade to a newer release. In your case, that will be jaunty (9.04) After that, you can run do-release-upgrade again to upgrade to 9.10. There could be issues.
[15:24] <savid> I see.
[15:25] <savid> Well, it looks like it wouldn't matter because even 9.10 has outdated versions of those programs :-P
[15:30] <savid> Would that be a bug in ubuntu, a bug in rkhunter, or neither?
[15:31] <alvin> savid: hardy has Gnupg 1.4.6, karmic has 1.4.9. What version does rkhunter needs?
[15:31] <savid> gpg on my my server is updated to 1.4.9,   not sure what version rkhunter is expecting
[15:33] <savid> looks like I'm not the only one:  http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=560157
[15:33] <alvin> savid: Ubuntu 10.04 will also have gnupg 1.4.9, (this can still change)
[15:34] <alvin> savid: Ah, it's a bug in rkhunter then. If really needed you can ask for an upgraded package (see: https://wiki.ubuntu.com/UbuntuDevelopment/NewPackages)
[15:41] <smoser> kirkland, ttx can you verify... if=<what> for kvm booted eucalyptus instances ?
[15:41] <smoser> i think its 'scsi'
[15:43] <ttx> smoser: no setup running right now, maybe in a few
[16:41] <ttx> smoser: test(s) in progress
[16:41] <smoser> i dont know how they will run
[16:41] <ttx> + the "i386 bundled without -r, does it run on UEC/i386" test
[16:42] <smoser> jjohansen, bug 494565
[16:42] <TeTeT> what to do when eucalyptus reports 'user admin not found' after just installing it?
[16:42] <smoser> ttx what did that mean, are you asking ?
[16:43] <ttx> smoser: no, I mean: "i will be running that test as well"
[16:44] <smoser> oh. ok.
[16:44] <TeTeT> nevermind, after a five minute wait it's here
[16:44] <smoser> i was thinking that '-r' == '--ramdisk', which i'm wondering about. see bug above.... i couldn't get it to work on "plain kvm" right now.
[16:55] <ttx> smoser: right, the current uec image tarballs do not have ramdisk
[16:55] <smoser> that is true
[16:56] <smoser> and by design
[16:56] <smoser> my question is if they boot
[16:56] <smoser> :)
[16:56] <ttx> hm
[16:56] <ttx> I have to adjust my testing script
[16:56] <smoser> i swear that i tested this, but today couldn't make kvm -ddrive if=scsi work with that kernel
[16:57] <smoser> ttx, you may want to utilize http://bazaar.launchpad.net/%7Eubuntu-on-ec2/ubuntu-on-ec2/uec-tools/annotate/head%3A/register-uec-tarball
[16:57] <sponzor> i want to do this http://pastebin.com/m8365263 is it posiblle?
[17:04] <ttx> smoser: if=scsi, confirmed
[17:05] <ttx> smoser: I confirm that it does not boot
[17:05] <smoser> bugger
[17:05] <ttx> smoser: i confirm that UEC i386 doesn't seem to mind booting an image marked x86_64
[17:06] <smoser> ttx, can you put that in that bug
[17:06] <ttx> smoser: do you need the euca-get-console-output for this one ?
[17:07] <smoser> ok, first, please comment in https://bugs.launchpad.net/ubuntu/+source/eucalyptus/+bug/494084 that i386 doesn't seem to mind booting an image marked x86_64
[17:07] <ttx> smoser: sure
[17:07] <smoser> second, this really sucks.
[17:08] <smoser> because according to https://wiki.ubuntu.com/ServerLucidCloudKernelRamdisk i verified boot of 20091201 on uec
[17:08] <ttx> you want the console output ? Which bug ?
[17:08] <smoser> but 20091201 has since fallen off, so i can't easily confirm or refute my assertion
[17:09] <smoser> we need a new bug for "uec images dont boot". you can but console output there, please.
[17:09] <smoser> jjohansen, ping
[17:13] <ttx> smoser: bug against what package ? "linux" ?
[17:13] <smoser> yeah
[17:13] <smoser> linux-virtual
[17:13] <smoser> fails for both i386 and amd64 ?
[17:14] <ttx> hm, no time left to test /that/
[17:18] <ttx> smoser: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/494611
[17:19] <ttx> smoser: sorry I don't have time for more complete tests, just posted what I have
[17:19] <smoser> ttx, thanks. its good enough.
[17:39] <jongbergs> hi, which dns software is reliable and recommended for production use: bind, powerdns, mydns, djbdns, to name just a few..
[17:42]  * ScottK likes unbound
[17:51] <zul> bind
[17:51] <soren> I use bind.
[18:53] <alex88> how can i set the owner of file transferred with samba?
[18:53] <arj> force something
[18:54] <alex88> in which file? sry but i haven't used samba
[18:55] <mneptok> alex88: the Samba config file
[18:55] <alex88> mneptok: there are the samba shares there? and i force with "force create mode = 0770
[18:56] <alex88> ?
[18:56] <mneptok> alex88: unsure. i'm not a Samba user. but i know permissions are forced via the config file. i'm sure there are many examples on the web.
[18:57] <alex88> searching for them
[18:58] <arj> force create user I think
[18:58] <blackxored> hello guys, I need a compressed, incremental, and windows-aware backup solution for setting up an enterprise backup server, I'm taking a look at amanda, after giving up with bacula because of it's complexety, do you think I'm in a good path? If I am, can you point me to some updated documention about setting this up with 8.04.3??
[18:58] <RoyK> alex88: samba has unix extensions enabled by default, so it should work out of the box, given the same UIDs on both sides
[18:59] <RoyK> without identical UIDs, there's no way it'll work
[18:59] <alex88> ok, thank you all..
[19:08] <blackxored> anyone?
[19:08] <blackxored>  hello guys, I need a compressed, incremental, and windows-aware backup solution for setting up an enterprise backup server, I'm taking a look at amanda, after giving up with bacula because of it's complexety, do you think I'm in a good path? If I am, can you point me to some updated documention about setting this up with 8.04.3??
[19:09] <alex88> blackxored: i've also tried bacula and i can agree that's complex..
[19:09] <blackxored> alex88, extremely to my goals
[19:11] <alex88> i'm now using webmin ftp backup..but it's not incremental
[19:11] <blackxored> alex88, you're using something for personal back-ends I assume
[19:12] <blackxored> I need something enteprise-grade, centralized, incremental, compressed, windows smb/<custom soft> backups and the like
[19:12] <blackxored> so I'm now installing amanda
[19:13] <alex88> good luck so.. btw yes, personal, my small vps..
[19:14] <RoyK> blackxored: I'd rather take a look at opensolaris, the later -dev releases with dedup
[19:14] <RoyK> linux can't do that stuff
[19:15] <blackxored> RoyK, hheheh really, don't make me laugh
[19:15] <blackxored> amanda does all that
[19:15] <blackxored> bacula does taht
[19:15] <blackxored> <some other software I haven't even heard off> does that
[19:15] <RoyK> realtime deduplication and compression on disk?
[19:15] <blackxored> and in the end time
[19:15] <blackxored> solaris = unix
[19:15] <blackxored> :P
[19:15] <RoyK> all the while open source and very fast and very secure?
[19:15] <blackxored> is not something martian
[19:16] <arj> but block level dedup is
[19:16] <blackxored> RoyK, don't market solaris here, I was just asking a question
[19:16] <RoyK> opensolaris works with native cifs sharing and ACLs and so on
[19:16] <blackxored> RoyK, be happy to heard that I'd prefer that to windows
[19:16] <blackxored> RoyK, linux too
[19:16] <RoyK> and cifs/smb without samba, that is
[19:16] <RoyK> in-kernel cifs server
[19:16] <arj> wait what?
[19:16] <RoyK> opensolaris rocks, even if you don't like it
[19:17] <RoyK> arj: what what?
[19:17] <arj> in-kernel cifs?
[19:17] <alex88> but this isn't #opensolaris..xD
[19:17] <RoyK> arj: yeah, with full ACL support
[19:17] <RoyK> nfs4 compatible
[19:17] <blackxored> alex88, +1
[19:18] <blackxored> RoyK, see:
 RoyK, don't market solaris here, I was just asking a question
[19:18] <RoyK> alex88: I know, but blackxored asked for stuff not in linux, so I just suggested something else
[19:18] <blackxored> RoyK, I have windows *clients*
[19:18] <RoyK> blackxored: I'm not 'marketing', I'm suggesting something that works
[19:18] <blackxored> that's what I said
[19:18] <alex88> oh, so i haven't seen that request.. i'm sorry..
[19:19] <blackxored> alex88, neither did I, my fingers may have betrayed me :P
[19:19] <alex88> =)
[19:19] <blackxored> smb or custom software for windows clients
[19:19] <blackxored> RoyK, but help is always welcome, I'm not blaming you, but see ^^^^
[19:20] <RoyK> btw, anyone here ever managed to make windows ACLs work with samba? as in changing them?
[19:20] <blackxored> I'm starting a virtualization setup at work, and this is my first step, the backup server
[19:20] <RoyK> that's why I suggested what I did - disk backup with dedup is very nice
[19:21] <blackxored> but my setup = linux hosts <=> linux guests <=> windows clients
[19:22] <RoyK> then either setup an opensolaris guest or even better, install a new box with a bunch of el-cheapo SATA drives with opensolaris and raidz2 for backup
[19:22] <RoyK> no hardware raid - just raidz(2)
[19:22] <RoyK> and dedup and compression
[19:23] <RoyK> raidz is WAY safer and faster than 'hardware raid'
[19:23] <blackxored> RoyK, linux user and developer, sorry I won't consider anything besides linux, probably only netbsd for the firewall
[19:23] <RoyK> same idea as BTRFS, except BTRFS isn't stable
[19:23] <RoyK> blackxored: heh - as you like - I was a sceptic myself too before I dug into it
[19:24] <Aison> is there some documentation of to configure pppoe? in my case, I would like to setup 3 pppoe connections  on the same machine
[19:25] <blackxored> RoyK, also if we're going to market, I'm debian/ubuntu developer so I'm practically stuck to those distros for principles hehehehhehe
[19:26] <Aison> I already found this documentation, but I dont know if this pppoeconf let me setup 3 distinct connections
[19:26] <Aison> https://help.ubuntu.com/community/ADSLPPPoE
[19:27] <RoyK> blackxored: I don't think people will hate you if you use something else for the backup solution
[19:27] <RoyK> blackxored: that is, unless you want to rely on btrfs, which doesn't have dedup and whose structure is still not fixed
[19:28] <blackxored> RoyK, finished installing, tweaking it, thank you
[19:39] <jumbers> Hi, I'm having a frequent problem on my server which is becoming quite an issue. Soon after rebooting, my server will halt without warning. I've checked my logs, but I don't see anything in them that would cause this
[19:40] <jumbers> Does anybody have a suggestion as to what could be wrong here?
[19:55] <Sam-I-Am> jumbers: halt as in lock up or shutdown cleanly?
[19:57] <jumbers> As I don't have physical access to the datacenter, I do not know. I assume a lock up because the logs just show normal messages and then silence in the log until the next time it boots up
[19:58] <jumbers> Sam-I-Am: It almost feels like a RAM problem because it only happens very soon after boot, which I would think is when the most memory would be getting allocated
[19:59] <Sam-I-Am> if it just locks up its a hardware problem
[19:59] <Sam-I-Am> kernel issues will generally throw out a bunch of panic info
[20:00] <jumbers> The only kernel messages I see are say 10 minutes after the problem, at the next boot. And they're just normal kernel boot messages
[20:01] <jumbers> Would a faulty RAM stick make the most sense here? Like I said, it only happens soon after a reboot, which I try to only do when a kernel update is released
[20:01] <Sam-I-Am> who knows
[20:01] <Sam-I-Am> could be anything
[20:02] <Sam-I-Am> bad ram is usually a good first shot for testing
[20:02] <Sam-I-Am> try memtest86
[20:02] <Sam-I-Am> download it, boot from it, let it run
[20:02] <jumbers> Again, no physical access to the machine :-/
[20:04] <jumbers> I guess what I'll do is put in a support ticket and explain that this has become a recurring problem and that I think the RAM may be faulty
[20:04] <neonfreon> kernel issues can also cause lockups
[20:04] <neonfreon> fyi ;)
[20:05] <jumbers> I remember this same sort of problem I had once on a Windows machine. Whenever I would play a certain game that used a lot of RAM, I'd be playing and then suddenly the thing would BSOD and die. Turned out the stick was faulty
[20:05] <neonfreon> what changed when it started happening
[20:05] <jumbers> It's been 2 months, I'm honestly not sure
[20:05] <jumbers> A kernel issue wouldn't spit out any sort of message before dying?
[20:05] <neonfreon> not necessarily no
[20:06] <Sam-I-Am> neonfreon: i think they're rare these days
[20:07] <Sam-I-Am> i havent had a linux machine lock solid due to a kernel problem in years
[20:07] <Sam-I-Am> and when they have, i've gotten kernel spew
[20:07] <neonfreon> perhaps you haven't been trying enough kernels! =)
[20:08] <Sam-I-Am> neonfreon: yeeeah...
[20:08] <jumbers> I just use the ones released via the repos :p
[20:08] <Sam-I-Am> i havent run bleeding edge stuff in a long time
[20:08] <Sam-I-Am> nor on bleeding edge hardware
[20:08] <neonfreon> ah
[20:36] <jumbers> I guess we shall see if they can find a hardware problem
[20:51] <AndyGraybeal>  /part
[21:15] <jiboumans> ScottK: regarding server-lucid-more-mail-integration, I agree with ttx's review. For clarity I've added that to the blueprint.
[21:16] <ScottK> jiboumans: That's fine, I just want to get it marked approved so I don't get caught up in feature approval deadline issues.
[21:16]  * ScottK will fix it or convince ivoks to do it.
[21:16] <jiboumans> scottk: you're fine in that regard, although both me and ttx would love to see how we can avoid the pain from the issues we saw
[21:17] <ScottK> OK.
[21:17] <ScottK> Thanks.
[21:18] <jiboumans> no problem; sorry for not being clearer on that point earlier. i asked ttx to take a look and add the review as soon as you requested it, but i failed to let you know that.
[21:18] <ScottK> jiboumans: He did give me some feedback, I've just been slammed with $WORK.
[21:18] <jiboumans> scottk: i feel your pain :)
[21:19]  * ScottK has been wanting to do this since Intrepid, so I think I've got it worked out how, even if I didn't explain it very well in the spec.
[22:01] <unit3> heh, was just about going to ask about exim4 hubbed_hosts behaviour because it wasn't working, and then saw a big typo. whoops.
[22:02] <unit3> Works fine now!
[22:03] <sponzor> will this work? http://pastebin.com/m8365263
[22:04] <Sam-I-Am> anything works, it whether or not it works in your situation :)
[22:04] <unit3> Not gonna do the math, but if those disk sizes add up, it should be fine.
[22:04] <Sam-I-Am> which from a random pastebin, we cannot tell
[22:04] <unit3> Just make the partitions the sizes you want, add then to the raid devices.
[22:05] <sponzor> ok i will give it a try :P
[22:21] <slide> How do I see what the current dns lookup ip is?
[22:23] <unit3> errr... do you mean "how do you see what name servers your computer is currently using to do dns lookups"?
[22:24] <unit3> that'll be in /etc/resolv.conf.
[22:25] <slide> is there anyway to refresh it?
[22:26] <unit3> err, it's a live file. the values in there are the values the system uses.
[22:26] <unit3> I'm not sure what you mean.
[22:27] <slide> ok i guess is there anyway to update that file based off a dhcp server but still use a static ip?
[22:29] <unit3> maybe... I'm unsure why you'd want to do such a thing, though. If your IP is static, then certainly your DNS servers are static.
[22:30] <slide> well, i move this server between my home in tx and la, and since they both use 192.168.1.x ips i just set its ip to be .99 so that i can always access it the same way
[22:33] <unit3> ahhh... well, it might be easier to just give it a static IP via the DHCP server at each location, and then just use DHCP as normal.
[22:34] <slide> one of the routers doesnt have the ability to assign static ips based on the mac =\
[22:34] <unit3> oh...
[22:35] <unit3> well, the other thing you could do is use Google's public DNS servers, 8.8.8.8 and 8.8.4.4.
[22:35] <unit3> Those'll work everywhere.
[22:35] <unit3> and then you can leave it with a static IP.
[22:35] <slide> hrm true, let me see
[22:37] <slide> i know my isp here blocks outside mail servers, it may do the same with dns
[22:37] <unit3> shouldn't do, since there's no spam related problems to using external dns.
[22:37] <unit3> but try it.
[22:37] <unit3> you can do something like "dig yahoo.ca @8.8.8.8"
[22:37] <unit3> and see if it works. ;)
[22:38] <slide> worked
[22:38] <slide> cool thanks
[22:40] <slide> awesome now my downloads are going at 1.2MB/s haha
[22:40] <unit3> heheh there you go.
[22:42] <sponzor> is there any faster way to delite all partitions in installation? becose one by one is long... :P
[22:43] <sponzor> erasing data on... it takes ages and i have like 20 partitions to modify.. :/ i will be deliting this like 10  days :P
[22:44] <ruben23> hi any simple but reliable backup system for enterprise setup..
[22:46] <unit3> sponzor: in the installer, if you select the disk it should ask you if you want to remove all partitions.
[22:46] <unit3> I believe. I can't remember exactly.
[22:46] <unit3> but there's an option for sure.
[22:46] <KismetGFX> ya that works
[22:46] <sponzor> ok i will check it
[22:49] <sponzor> ok i did it.. tnx
[22:50] <unit3> np
[22:51] <KismetGFX> only reason i was able to verify that is coz i've reinstalled/repartioned ubuntu over 20 times it seems past 4 days str8 lol still no success
[22:52] <unit3> heh
[22:52] <KismetGFX> windows sadly installs just fine. i've ran fixmbr install looks successful boot up and i get non-system disk or disk error replace and strike any key when ready as if theres no OS installed
[22:52] <KismetGFX> when booting from the 9.10 cd and executing the shell after a complete install, i can see my install in /target
[22:54] <Aison> I installed snmpd on 3 ubuntu servers
[22:54] <Aison> all are using the same snmpd.conf
[22:54] <Aison> but I can do snmpwalk only on one
[22:54] <Aison> im really confused what's wrong :(
[22:56] <unit3> hard to say without seeing your snmpd.conf.
[22:56] <Aison> I only enabled com2sec readonly  default         public
[22:57] <Aison> besides that, i didnt change anything
[22:57] <Aison> ah yes, snmpwalk on the host itself works on all machines
[22:57] <Aison> just over network it isn't
[22:59] <KismetGFX> i think i've ran fdisk/mbr, fixmbr and reinstalled lilo and grub so many times it is so hosed up can't figure it out :|
[23:00] <KismetGFX> cant tell if grub or lilo is loaded no trace of either on startup
[23:18] <sponzor> when the installation ask me how many active devices for the raid array? i have 5 disks so i should put there  4? or 3? or 5? :P
[23:19] <unit3> uhhh... however many you want active, so probably 5?
[23:22] <sponzor> so the installation will configure raid 5? that 5 disks are active and one is for failsafe?
[23:23] <DrUnKnMuNkY> i'm running apache 2.2.8 on ubuntu 8.04 and apache won't shut down cleanly anymore (apache2ctl, init.d, killall). i need to do a "killall -KILL" to stop it. nothing in error.log. any ideas?
[23:24] <sponzor> uh i understand now.. is this it.. 5 disks are active and if one failes you have to change it by your self. and if you put there 4 and one dies the spare disk then become the active one?
[23:25] <unit3> if we're talking raid5, then yes, that's what'll happen.
[23:26] <sponzor> ok i was a little confused with the part active.. tnx
[23:29] <KismetGFX> super grub is truly super found out i have a mismatched or corrupt version of stage1/stage2
[23:31] <Aison> how can I setup network device with NO ip, subnetmask ,etc.. aplied? just a NULL device
[23:32] <Aison> of course in /etc/network/interfaces
[23:34] <unit3> you use the "manual" interface method, instead of dhcp or static or whatever.
[23:34] <Aison> ok :)))
[23:39] <czajkowski> kirkland: thanks for making your testdrive much easier to follow and test :)
[23:44] <kirkland> czajkowski: thanks