[00:00] <amelie> hi
[00:04] <amelie> [18:53] <amelie> I made a RAID 1 with 2 500gb hard drives
[00:04] <amelie> [18:54] <amelie> i mean (2)  500 gb hard drives
[00:04] <amelie> [18:54] <-- Dragonmaster_Dan has left this server (Remote closed the connection).
[00:04] <amelie> [18:54] <amelie> formatted the partitions with ubuntu server 8.10
[00:04] <amelie> [18:54] <amelie> then, downloaded the kubuntu-desktop
[00:04] <amelie> [18:54] <amelie> my system was working right during a month
[00:04] <amelie> [18:54] <amelie> and suddenly, I've got this error:
[00:04] <amelie> [18:55] <amelie> /dev/md6: clean, 37/24096 files, 34600/96256 blocks
[00:04] <amelie> [18:55] <amelie> /dev/md9 contains a file system with errors, check forced.
[00:08] <Kamping_Kaiser> fsck did a check. so what?
[00:57] <cjwatson> dayo__: I see RediXe's gone, but for the record I believe our standard documented method is 'sudo deluser <user> admin', just as adduser is our standard documented method for adding users
[01:02] <amelie> /dev/md9: File ... (inode #14246082, mod time Tue Jan 27 18:53:46 2009)
[01:02] <amelie>   has 1 multiply-claimed block(s), shared with 1 file(s):
[01:02] <amelie> /dev/md9: 	/amelie/.mozilla/firefox/2d4jsiml.default/Cache/41628DC2d01 (inode #14246794, mod time Tue Jan 27 23:01:42 2009)
[01:02] <amelie> /dev/md9:
[01:02] <amelie> /dev/md9: UNEXPECTED INCONSISTENCY; RUN fsck MANUALLY.
[01:02] <amelie> 	(i.e., without -a or -p options)
[01:02] <amelie> fsck died with exit status 4
[01:02] <amelie> what does it mean?
[01:09] <jmarsden|work> amelie: It means that filesystem has an issue that fsck cannot fix automatically... umount it and then do what the error msg says... run fsck manually :)
[01:13] <twb> Where can I read about the differences between the linux-image-server and linux-image-generic packages?
[01:13] <twb> "server" is rather more nebulous than the -march variations I'm used to under Debian.
[01:15] <maxb> twb: -server vs. -generic is more about the selection of kernel config options, than the cpu architecture
[01:15] <maxb> Things like is the scheduler tuned for desktop responsiveness, or server throughput
[01:21] <twb> maxb: indeed; so where can I read about who decides what things to tweak, and how?
[01:22] <maxb> who? The kernel team, I suppose. What exactly? Not sure, though you could always resort to diffing the relevant files in the source package
[01:24] <twb> I was hoping there'd be a policy document (say, in the wiki) about it.
[01:26] <amelie> thanks jmarsden|work, but the thing is i don't know how to use the fsck or mount a partition; i'm pretty newbie on this
[01:26] <jmarsden|work> amelie: Are you the administrator of a machine running Ubuntu Server?
[01:27] <maxb> twb: Well, you could hop on #ubuntu-kernel, ask, and hang around for a while.
[01:27] <amelie> yes
[01:28] <jmarsden|work> OK.  Is the filesystem on /dev/md9 mounted now, or not (can you get at the files on it)?
[01:28] <amelie> i don't know
[01:28] <amelie> i just hit control D to boot up normally
[01:30] <jmarsden|work> OK, open up a Terminal and in it do      mount | grep md9    and tell me what it displays
[01:33] <twb> maxb: thanks.
[01:37] <amelie> "/dev/md9 on /home type ext3 (rw,relatime)"
[01:38] <amelie> i suppose it means is mounted right?
[01:38] <jmarsden|work> OK, so it is mounted.  We need to undo that so we can fix it... but it is your /home partition, which could make life fun... are there other users currently logged into or accessing this server right now.
[01:38] <jmarsden|work> Or just you?
[01:40] <jmarsden|work> (Basically, it may be simplest to reboot it to the point where it shows the error and drops you into a recovery shell... but that will affect any other users!
[01:44] <jmarsden|work> amelie: I only have about 15 minutes before I have to leave here... is it OK with you if we reboot your server as part of fixing this issue??
[01:46] <amelie> well, you mean restart the computer completely?
[01:46] <amelie> i don't know to use the fsck command
[01:47] <jmarsden|work> Yes; once we restart and you get to where you did ctrl-D last time, I can help with the fsck part...
[01:48] <amelie> how do i do that?
[01:48] <amelie> log as another user?
[01:48] <jmarsden|work> ?  You don't know how to restart your server??  Something feels odd here.
[01:48] <amelie> i told you
[01:48] <amelie> i'm a newbie....
[01:49] <jmarsden|work> Then who configured the server with software RAID...?
[01:49] <amelie> me
[01:49] <jmarsden|work> OK... if you are sure you are the only user using the server, do     sudo shutdown -r now
[01:49] <amelie> i just format 4 partitions on each disk as "raid" and later tied each one with the menu
[01:50] <amelie> and asign the mount points and system type
[01:51] <jmarsden|work> Oh... she was IRCing from the server itself?!  OOPS.
[02:18] <amelie> jmarsden|work: I made the fsck
[02:18] <amelie> and hit the default options
[02:18] <amelie> the /dev/md9, ended with errors I think
[02:19] <amelie> whe the system restarted, it says the same thing about unclean shutdown, but with partition md7
[02:20] <amelie> then, the automated fsck, worked
[02:20] <amelie> and reboot normally
[02:20] <amelie> i don't know if is working good now...
[02:20] <amelie> still there?
[02:21] <amelie> well, i supposed you are busy
[02:21] <amelie> i'll come tomorrow is late for me now
[02:21] <amelie> thanks for the help
[02:21] <amelie> bye
[02:40] <kansan> anything after exit 0;  in /etc/rc.local will NOT get executed... correct?
[03:48] <looseparts> Hello. Probably more of a windows question, but deals with Ubuntu and samba, so. I have finally setup my smb.conf to allow me to map a drive to a share on my server from my windows box. I want it to reconnect at boot up but it is requiring a password where i want it to connect automatically. thanks.
[04:34] <SudoKing> hi guys, my server (on LAN) is trying to initiate a connection to the gateway.  This computer I'm using is attached to the same LAN and can connect successfully.. the eth0 device doesn't recieve a DHCPOFFER, how can I identify the problem? :\
[05:19] <jtmoney> hey guys, i have software raid and updated my kernel... now my machine won't boot back up... how can i go about fixing this?
[08:42]  * soren realises that in order for the opennebula driver to get into the libvirt package, OpenNebula needs to be in main.
[08:50] <Scix> Is this the place for solving  gpgv problems?
[09:12] <tarimari> hi guys
[09:12] <tarimari> i want to setup hosts so that *.loc go to localhost. is it possible?  then from apache2 configuration, i ll direct to different website-folder
[09:37] <domas> ... can do with dns not with hosts
[09:37] <domas> but as he quite
[09:37] <domas> quit, he won't find that out
[10:14] <Chipzz> are the base differences between a debian and ubuntu install documented somewhere?
[10:16] <Chipzz> read: if I install debian, with no tasks selected in tasksel, and ubuntu server alternate with no tasks selected in tasksel, what are the differences I should expect?
[10:16] <Chipzz> I'm aware of init vs upstart, and different kernel packages
[10:17] <Chipzz> but any other issues where ubuntu deviates from debian?
[10:27] <kaushal> hi
[10:27] <kaushal> is there a step by step guide to install Oracle 9i Release 2(9.2.0.4) on Ubuntu Server 8.10 ?
[10:32] <domas> we already told you!
[10:33] <kaushal> domas, you said it for MySQL DB :)
[10:35] <domas> yes yes, but that solves your issue!
[11:03] <Scix> How can i get something like "server:/home/* /home/*/Server nfs noauto,user" to work in fstab?
[11:04] <Scix> Or it here an other way to mount a NFS dir, without beeing root
[12:52] <SmokeyD> hey people. Does anyone have tips for software I can use to monitor the amount of bandwidth consumed by postfix,dovecot and apache2 all running on the same machine, split up per domain name?
[12:54] <SmokeyD> I am thinking of doing it with iptables, but I don't know if iptables can distinguish between the virtual domainnames of smtp and pop3/imap traffic
[12:56] <soren> It can't.
[12:56] <SmokeyD> soren: ok, you have any other recommendations?
[12:56] <soren> No.
[12:56] <SmokeyD> or maybe monitor web traffic through iptables and other data through other means
[12:57] <soren> The web traffic case is clearly the simplest. Apache logs the size of each request and can be told to log to separate files for each domain.
[12:57] <soren> No need to get iptables involved.
[12:57] <SmokeyD> soren: :)
[12:57] <SmokeyD> good one
[12:57] <soren> The others are much more difficult.
[12:58] <soren> Neither postfix nor dovecot log sizes of requests.
[12:58] <soren> For dovecot you also might have very, very long-lived connections, which makes it quite difficult.
[12:58] <lamont> soren: postfix does log sizes
[12:58] <soren> lamont: Of SMTP sessions?
[12:58] <soren> lamont: ...or message sizes?
[12:59] <lamont> though you also get long-lived connections with connection caching now...
[12:59] <lamont> message sizes
[12:59] <soren> Ok.
[12:59] <SmokeyD> ok
[12:59] <lamont> OTOH, message size should roughly correlate to session size.  with a fudgefactor for turning DATA into $SESSION
[13:00] <soren> Also, what if a message is addressed for foo@example.com, but is delivered to foo@example2.com? Who should this count against?
[13:01] <SmokeyD> lamont: yes indeed, I can assume that since pop3 is mainly used, twice DATA=$SESSION (once arriving through postfix, once downloading through dovecot pop3)
[13:01] <soren> Logging of e-mail traffic like this is much less clear-cut.
[13:01] <soren> SmokeyD: Unless they decide to keep mail on the server.
[13:02] <SmokeyD> soren: indeed
[13:02] <SmokeyD> but I can assume they don't I think. I am not talking about a major commercial hoster, just a few domainnames of people I know well and trust
[13:10] <soren> If that's the case, why are you monitoring this?
[13:43] <chimp> When connecting to a webserver via its ip address, is it possible to pass it the host name you intend to use externally so that it sends you to the correct apache virtual webserver?
[13:44] <sommer_> chimp: sure, just add the hostname to your /etc/hosts file
[13:44] <chimp> I am a massive fool
[13:45] <chimp> cheers
[13:45] <chimp> must be too much revising of physics, ruins ones brain
[13:45] <sommer> I wouldn't say "massive" :-) ... it's one of those things that until you need to do it there's not much reason to know how
[13:46] <chimp> yeah but ive done it plenty of times before
[13:46] <chimp> my hosts file is full of stuff :S
[14:44] <DogWater> Does anyone know of any single command way to remove all partitions in Ubuntu?
[14:44] <domas> that sure must be single command
[14:45] <DogWater> hmm?
[15:01] <frippz> DogWater: I think you'll have to use fdisk for that (no single command as far as I know, unless you get someone to write a script for you :P)
[15:06] <soren> DogWater: Just nuke the partition table?
[15:06] <domas> random dd could work, yes :)
[15:06] <soren> DogWater: "dd if=/dev/zero of=/dev/nameofthedevice bs=512 count=1"
[15:06] <soren> Sounds like a horrible idea, though.
[15:06] <domas> mkswap /dev/nameofthedevice %)
[15:07] <domas> back when I was evil kid, I used to suggest 'mkswap' instead of usual 'rm -rf'
[15:07] <domas> would work quite well
[15:09] <lionel> soren: could you have a look at bug #223024 please. MOTU-SRU is interested by a ~ubuntu-server member POV on the patch before hacking
[15:10] <uvirtbot`> Launchpad bug 223024 in apt-cacher "apt-cacher / Use of uninitialized value in concatenation / line 169" [Medium,Confirmed] https://launchpad.net/bugs/223024
[15:25] <alexrussell> Hi folks. I've recently installed a new (64 bit) Ubuntu Hardy Server and did the LAMP option. Thsi has installed PHP as an apache module (as far as I can tell), but I want to install PEAR. If I do a 'sudo apt-get install php-pear' it talks about wanting to install php5-cli. While I'm happy to do so, will this ruin the default LAMP set-up (or even not give my non-cli php5 installation no access to PEAR which was my original intention in
[15:25] <alexrussell> the first place)? Also, I hear if one uses the cli version of php with apache, files need execute permission. Is this true and if so I'd like to avoid this situation as chmodding all php files to +x is a pain when the server's being used on a samba network...
[15:25] <alexrussell> ...sorry for the big lump of text there
[15:26] <_ruben> php-cli is just an addition to the apache modules .. you can have apache use the php modules and also have the php commandline tool installed at the same time
[15:26] <_ruben> iow: installing pear wont break your php+apache config
[15:26] <_ruben> unless something freaky happens
[15:26] <alexrussell> excellent, thanks
[15:44] <axisys> can I install ubuntu 64bit server on T1000 (sun4v) ?
[15:46] <jrwren> yes.
[15:46] <axisys> can I install the 8.10? it shows 7.10 certified
[15:47] <jrwren> if you can find a version for the ultrasparc T1 processor.
[15:48] <axisys> jrwren: partners/sun shows 7.10 as the latest certified for t1000
[15:48] <axisys> i wonder if i can install 8.10
[15:48] <axisys> ofcourse i can give it a try.. but wondering if anyone did it
[15:49] <jrwren> 7.10 looks like the last sparc release
[15:49] <jrwren> i'm just browsing here: http://osmirrors.cerias.purdue.edu/pub/ubuntu-releases/
[15:49] <ScottK> 7.10 was the last release that had official Canonical support for Sparc.  I know people who use the later releases.
[15:49] <jrwren> notice how in teh 7.10 directory there is a "sparc" iso image.
[15:49] <jrwren> but 8.04 there is not.
[15:49] <ScottK> It was moved to ports.
[15:49] <ScottK> There's a different URL.
[15:50] <ScottK> They aren't (I don't think) mirrored so look at ports.ubuntu.com (IIRC).
[16:02] <binspace> Is there an ubuntu universe/multiverse repository on ec2?
[16:03] <Deeeps> umm, i've not used ec2 before, but surely you just add them to your sources.list like anywhere else?
[16:06] <binspace> Yeah, I was just wondering if there was a public mirror. The reason is to make installs/updates faster.
[16:27] <axisys>  ScottK: i dont see the iso in ports.ubuntu.com
[16:27] <ScottK> I'm not sure then.
[16:28] <axisys> i see mini.iso, netboot, cdrom
[16:28] <axisys> they are as recent as jan 24 2009
[16:28] <ScottK> I except the mini.iso would be enough to get a booting system, but i don't know.
[16:29] <ScottK> Those will be for Jaunty then.
[16:29] <axisys> i saw an article on how to install using mini.iso and netboot..
[16:29] <ScottK> There was a period where CD builds for ports weren't happening, so that may be why you don't see them.
[16:29] <axisys> can't find it now.. the how to
[16:30] <lionel> axisys: you should find what you want on http://cdimage.ubuntu.com/ports/releases/
[16:32] <axisys> lionel: sweet! that's exactly what I was looking for
[16:32] <axisys> lionel: thanks a lot
[16:33] <lionel> you're welcome
[16:59] <mathiaz> soren: dendrobates- told me there was a need to come up with a PKI infrastructure for your cloud work
[17:00] <mathiaz> soren: is there a specific document that outlines the requirements?
[17:00] <dendrobates-> ha, soren document that goes against his religion.
[17:01] <mathiaz> soren: how does it work from the client side?
[17:01] <dendrobates-> right now certificates are ssh'd around by a shell script
[17:01] <dendrobates-> mathiaz: ^^^
[17:01] <mathiaz> dendrobates-: well - I've looked into openldap and all of the bits to support an X509 PKI are there
[17:02] <dendrobates-> mathiaz: I'm not sure how far we will get into this this release, but be prepared to talk to the euacalyptus team next week,
[17:02] <mathiaz> dendrobates-: support for rfc 4523 is there
[17:03] <mathiaz> dendrobates-: which means we should be able to store all of the certificates in an ldap tree directly
[17:04] <mathiaz> dendrobates-: ok- is there at least somewhere I can look at to understand how certificates are used in the overall architecture?
[17:04] <dendrobates-> mathiaz: now all they have to do is be able to get them out.
[17:05] <dendrobates-> mathiaz: I can give you an email addess to send a request to.
[17:06] <mathiaz> dendrobates-: I guess that would already be better than digging in the code
[17:06] <dendrobates-> mathiaz: I'm sure they have docs.
[17:21] <soren> mathiaz: I'm not familiar with the intimate details of it, I'm afraid.
[17:43] <ivoks_> https://wiki.ubuntu.com/Testing/Cases/UbuntuServer-drbd
[17:43] <ivoks_> i hope everyone can follow this one
[17:45] <mathiaz> ivoks_: seems like a good page! Thanks for putting it together
[17:46] <ivoks_> np
[17:48] <ivoks_> mathiaz: testing primary-primary would require setting up cluster, so i decided to leave that out
[17:48] <ivoks_> it's too complicated for testing
[17:49] <ivoks_> if someone doesn't understand how it works
[17:49] <ivoks_> it's complicated to explain, not to test :)
[17:50] <mathiaz> ivoks_: right. That's a good start.
[17:50] <ScottK> Step 1: Go learn all about brbd.  Step 2: The rest is obvious.
[17:50] <ScottK> ;-)
[17:51] <ivoks_> drbd :)
[17:51] <jmedina> good, here another infor about brdb and lvm resize
[17:51] <jmedina> p://www.asplund.nu/xencluster/xen-extend-domu-disk.html
[17:51] <ivoks_> drbd! :D
[17:52] <ivoks_> resizing...
[17:52] <ivoks_> i could add that to a test
[17:52] <jmedina> :D, they have to change their products name :D, looks like brdb is easier
[17:53] <ivoks_> first d is from disk
[17:53] <ivoks_> so... :D
[18:20] <axisys> ok i downloaded the iso from cdimages.. now how do I jump it? never jump installed ubuntu.. only solaris
[18:21] <ivoks> jump install?
[18:21]  * ivoks googles
[18:22] <ivoks> google shares my feelings :D
[18:22] <ivoks> could you explain what's jump install?
[18:24] <axisys> ivoks: hehe.. i meant jumpstart .. (solaris term)
[18:24] <axisys> ivoks: how do I network install
[18:24] <axisys> ivoks: the server has no CDROM or USB
[18:25] <axisys> ivoks: t1000
[18:25] <ivoks> oh
[18:25] <ivoks> https://wiki.koeln.ccc.de/index.php/Ubuntu_PXE_Install
[18:25] <ivoks> i find this tutorial perfect
[18:26] <ivoks> just watch out when downloading image
[18:26] <ivoks> 'Get and install the files'
[18:26] <orudie2> how can i find out which version of php and mysql server is currently installed ?
[18:26] <ivoks> apt-cache show php5
[18:26] <ivoks> apt-cache show mysql-server
[18:27] <ivoks> axisys: sparc?
[18:27] <axisys> ivoks: yes
[18:27] <yann2> axisys > hi
[18:28] <yann2> axisys > could you write an article once you get it to work? i got some t1000 too
[18:28] <axisys> yann2: sure
[18:28] <yann2> you plan to install which version?
[18:28] <ivoks> axisys: http://ports.ubuntu.com/dists/hardy/main/installer-sparc/current/images/sparc64/netboot/
[18:28] <yann2> hey I didnt know there still were some isos for sparc for hardy
[18:29] <ivoks> ports.ubuntu.com has it all
[18:29] <ivoks> :D
[18:30] <axisys> yann2: for isos you go to cdimages.ubuntu.com/ports
[18:30] <axisys> ivoks: i was not looking for images
[18:31] <ivoks> axisys: those aren't images
[18:31] <ivoks> axisys: that's kernel and initrd for netboot
[18:31] <axisys> ivoks: true
[18:32] <axisys> ivoks: this looks a good starting point for how to for me
[18:32] <axisys> ivoks: https://help.ubuntu.com/community/Installation#Server and network installations
[18:33] <ivoks> i've never netbooted sparc
[18:33] <ivoks> i don't know how that works...
[18:34] <axisys> ivoks: boot net - install at ok prompt.. it looks for the tftp
[18:35] <ivoks> great
[18:35] <ivoks> i would just follow the link i gaved you
[18:36] <ivoks> and instead of serving i386 initrd and kernel, serve sparc kernel and initrd
[18:36] <ivoks> from the other link i pasted
[18:36] <genii> Another great link is https://wiki.koeln.ccc.de/index.php/Ubuntu_PXE_Install for netboot installs
[18:38] <ivoks> :)
[18:38] <ivoks> genii: i agree :D
[18:39] <genii> ivoks: Ah, when I check backscroll I see it was posted already, apologies for my redundancy :)
[18:40] <ivoks> redumdacy is good
[18:40] <ivoks> we are on #ubuntu-server
[18:40] <genii> Hehe, yes
[18:40] <ivoks> we like redundacy :D
[18:40] <ivoks> bah...
[18:40] <ivoks> too many typos today
[19:13] <grindking> if i wanted to change the desktop machine i've been using with 8.10 intrepid to more of a server setup, would it be a bad idea to just install the packages i need, and remove the ones i don't rather that installing the actual ubuntu server iso? i'm trying to save a step if i can :D
[19:14] <asterisk_user2> Hello
[19:14] <asterisk_user2> we want to replace our Windows 2000 server (active directory)
[19:14] <asterisk_user2> and we would like to use UBuntu server
[19:15] <asterisk_user2> Is that possible with Ubuntu ? (to use as a file server with AD features)
[19:15] <ivoks> mathiaz: i've added test D, primary-primary
[19:16] <jmedina> asterisk_user2: depends what AD features you want
[19:17] <jmedina> with samba3 you can implement a NT4 Domain Controller
[19:17] <asterisk_user2> NT4 domain controller ? OK
[19:17] <asterisk_user2> will have a look at that
[19:18] <ivoks> AD will be possible with samba4
[19:18] <asterisk_user2> I would like to block internet settings
[19:18] <asterisk_user2> and other configuration options
[19:18] <asterisk_user2> for security reasons
[19:18] <asterisk_user2> OK nice to hear
[19:18] <ivoks> it is not possible to replace windows AD with non-windows AD
[19:18] <ivoks> at the momment
[19:19] <ivoks> samba4 will be able to do that, but it's not stable yet
[19:19] <asterisk_user2> ok thank you ivoks
[19:19] <asterisk_user2> do you know when samba4 is stable enough for production use ?
[19:19] <asterisk_user2> maybe 2010 ??
[19:20] <ivoks> maybe even this year
[19:21] <asterisk_user2> i just found http://wiki.samba.org/index.php/Samba4
[19:21] <asterisk_user2> sounds very cool !!
[19:21] <ScottK> We have alpha grade Samba4 packages in Intrepid and Jaunty.
[19:21] <asterisk_user2> who tried SAMBA4 ?
[19:22] <asterisk_user2> SAMBA 4 seems to become much better than 3
[19:22] <ivoks> it's a rewrite
[19:23] <jmedina> asterisk_user2: but samba4 is not yet complete
[19:24] <asterisk_user2> yes i know...
[19:24] <asterisk_user2> looking forward to samba4
[19:24] <jmedina> asterisk_user2: you can make your donations to the samba fundation to speed up development XD
[19:25] <asterisk_user2> ok
[19:25] <asterisk_user2> i think i will !
[19:25] <kansan> do things in /etc/rc.local get run as root?
[19:26] <kansan> i.e. if i want to do:  "sudo start runsvdir"  how do i make sure that happens on every startup?
[19:26] <jmedina> kansan: yeap
[19:26] <kansan> jmedina, so i dont need the sudo there huh
[19:26] <jmedina> dont think so, try it with a simple echo
[19:27] <ivoks> it runs as roon
[19:27] <ivoks> root
[19:28] <ivoks> brb
[19:37] <axisys> i am reading the mailing list ... i get the impression ubuntu will drop the ball on sparc very soon.. is that general belief here?
[19:38] <ivoks> axisys: i have couple of sparc servers
[19:38] <ivoks> axisys: it's not supported by canonical
[19:38] <ivoks> axisys: but it still gets builded and isos are created
[19:39] <axisys> ivoks: is the support dropped or was never there ?
[19:39] <ivoks> it was there for couple of releases
[19:40] <ivoks> axisys: but, for example, bug in openssh is fixed in source, and all binaries are builded, including those for sparc
[19:40] <axisys> ivoks: my work wont let me install it unless there is some support available.. i guess i rather go with solaris then :-(
[19:40] <axisys> hmm. what a predicament
[19:42] <ivoks> i've choosed ubuntu for my sparcs
[19:42] <ScottK> ivoks: Offer him a support contract.
[19:43] <ivoks> :)
[19:43] <axisys> heh
[19:44] <ivoks> i just love drbd :)
[19:45] <kees> ivoks: I really want to play with it again.  I tried to set it up back around 2003 or 2004 or so.
[19:45] <ivoks> kees: it's way cooler these days :D
[19:45] <ivoks> it even supports three nodes
[19:45] <kees> !!
[19:45] <kees> nice
[19:46] <kees> ivoks: yeah, it had Issues(tm) back then.
[19:46] <kees> I wanted it for making our mailman server into a hot-spare cluster.
[19:46] <kees> mailman really didn't like the idea of being on 2 filesystems.  :)
[19:46] <ivoks> :)
[19:47] <ivoks> i'm looking how to integrate it better with cluster-suite
[19:47] <ivoks> now it's just too much manual work
[19:48] <ivoks> it should start from rcS.d not rc2.d, before cluster-suite
[19:48] <ivoks> so that one can have gfs on top of drbd
[19:48] <ivoks> and still be able to boot :D
[19:50] <kees> cool
[19:50] <ivoks> i just fixed it :D
[19:51] <ivoks> yay!!! :D :D
[19:53] <ivoks> i'll upload the change
[19:53] <ivoks> oh, i can't do that :D
[19:55] <ivoks> i think sshd should be the first service that starts after network is up
[19:55] <ivoks> not mounting filesystems
[19:56] <Jeeves_> hi all
[20:04] <ivoks> zul: good catch
[20:04] <ivoks> zul: on bacula
[20:04] <zul> dendrobates caught it
[20:05] <genii> ivoks: How can it know where sshd is until it mounts the dir containing it?
[20:06] <ivoks> genii: you are talking about the case when /usr is different partition?
[20:06] <ivoks> anyway
[20:07] <ivoks> network is up after local filesystems are mounted
[20:07] <ivoks> so, there is no problem
[20:07] <ivoks> problem is that GFS/GFS2/NFS are mounted and some services are started before SSH is up, before you have access to remote system
[20:07] <Deeeps> ivoks: unless sshd is on a network file system
[20:07] <ivoks> this is great when everything works
[20:08] <ivoks> but when you have faulty drbd/gfs or some service
[20:08] <ivoks> you are in big trouble
[20:08] <ivoks> Deeeps: well, that's a corner case... in that case, everything is possible
[20:09] <ivoks> i had a case when my system didn't come up
[20:09] <jrwren> ssh in initrd!
[20:09] <jrwren> :)
[20:09] <Deeeps> likewise recently, if you rememeber ;)
[20:09] <ivoks> it was thousands of kms away, and it just had IP
[20:09] <jrwren> maybe dropbear sshd would fit :)
[20:09] <Deeeps> with sendmail and bind9-host hehe
[20:10] <kees> ivoks: dropbear, d'oh jrwren beat me
[20:10] <ivoks> hm... no, maybe someone doesn't want ssh?
[20:10] <Deeeps> thats when you kick yourself for not having an IPKVM / real servers with iLO and the like
[20:10] <ivoks> desktops, for example
[20:10] <ivoks> Deeeps: luckly, i had IPMI :D
[20:11]  * jmedina loves IPMI
[20:11] <Deeeps> atta boy
[20:11] <ivoks> i would just move ssh from rc2.d to rcS.d
[20:11] <ivoks> before portmap
[20:11] <ivoks> and other network stuff
[20:11] <ivoks> what's with upstart anyway?
[20:11] <ivoks> upstart would solve this problem
[21:06] <uvirtbot`> New bug: #319850 in mysql-dfsg-5.1 (universe) "mysql-*-5.1 still depends on mysql-common-5.1." [Undecided,Fix released] https://launchpad.net/bugs/319850
[21:15] <jrwren>   What would you like to do about it ?  Your options are:
[21:15] <jrwren>       D     : show the differences between the versions
[21:15] <jrwren> would be cool if dpkg would invoke a 3way diff tool.
[22:44] <sektor1952> evening
[22:45] <sektor1952> what's the best way to installing gnome without the extra fluff like kernels, bluetooth etc
[22:46] <Kamping_Kaiser> pardon?
[22:47] <sektor1952> I would like to install gnome to make it easier build some stuff out but aptitude install ubuntu-desktop installs way too much
[22:49] <Kamping_Kaiser> try gnome-desktop-environment or gnome-desktop. (however, i feel your question is somewhat OT for here)
[22:49] <sektor1952> well I installed ubuntu server from iso
[22:50] <ScottK> As soon as you say Gnome, it's OT here.  Doesn't matter what you installed.
[22:51] <sektor1952> my apologies
[22:51] <sektor1952> thx
[22:52] <binspace> Hello, I need some help with debconf with mysql-5.0. It doesn't seem to work.
[22:52] <binspace> I've tried echo "mysql-server mysql-server/root_password select" | sudo debconf-set-selections
[22:53] <binspace> and echo "mysql-server-5.0 mysql-server/root_password password password" | sudo debconf-set-selections
[22:53] <binspace> and I still get prompted
[22:53] <binspace> erichammond: Hey Eric, can I pick your brain for a bit?
[22:57] <erichammond> binspace: If you ask questions I might be one of the people who answers :)
[22:58] <erichammond> binspace: If it relates to EC2 you might find a wider variety of experts on #ubuntu-ec2 and ##aws
[22:58] <binspace> ok, thanks
[22:58] <erichammond> (including me)
[23:06] <uvirtbot`> New bug: #322952 in nagios3 (main) "several reports will not allow submitting parameters" [Undecided,New] https://launchpad.net/bugs/322952
[23:06] <binspace> *sigh* Ubuntu is driving me crazy
[23:18] <binspace> Ok. I found the solution. I was doing apt-get install mysql-server
[23:18] <binspace> I needed to do apt-get install mysql-server-5.0
[23:18] <binspace> sheesh
[23:19] <rgotten> has anybody good experience with software raid
[23:20]  * ogra points binspace to man debconf and the fset command (with seen = true) to avoind questions after preseeding
[23:23] <binspace> ogra: Honestly I was staring at man debconf, and its pretty anemic. I don't understand what you mean by fset and seen = true. I know that the default answer is to RTFM, but TMFS (the manual fucking sucks).
[23:26] <binspace> And relevant examples go much further than mountains of configuration documentation. Simply put the signal to noise ratio of man pages tends to be tiny.
[23:37] <ogra> binspace, echo -e "set mysql-server/root_password select\nfset mysql-server/root_password seen true" | debconf-communicate mysql-server
[23:38] <ogra> that line would tell debconf to never ask that question again
[23:40] <ogra> i agree that man debconf isnt actually of much help if you want to do preseeding ... man 7 debconf is though
[23:41] <binspace> ogra: Dude. You are a force of nature :-) Thanks alot.
[23:41] <binspace> I noticed that man debconf-devel has the fset info
[23:41] <ogra> oh, its actually debconf-devel :)
[23:42] <ogra> not debconf
[23:43] <binspace> yeah. Its a weird naming convention. I figured it was for people hacking on debconf. I guess the rule is follow the see also links.