[01:44] <ScottK> kirkland: How'd the server team meeting go today?
[01:44] <kirkland> ScottK: hey, it happened :-)  Koon led it, as mathiaz was absent
[01:45] <ScottK> Who is Koon?
[01:45] <kirkland> ScottK: Koon is Thierry Carrez
[01:46] <ScottK> Ah.  He sent the announcement mail.
[01:46] <kirkland> ScottK: he's leading our Java efforts, among other things
[01:46] <ScottK> Has he been with Canonical long?
[01:46] <kirkland> ScottK: started just before UDS
[01:46] <kirkland> mathiaz was on vacation and needed someone to run the meeting, Koon volunteered ;-)
[01:46] <ScottK> Was he there?  I don't remember meeing?
[01:46] <kirkland> Yeah, he was there
[01:47] <PrivateVoid> sorry I missed the meeting... I was in the BT meeting
[01:49] <ogra> ScottK, he had dinner with us at the naked fish the first evening (werent you there ??)
[01:49] <ogra> (at the sprint)
[01:50] <ogra> err
[01:50] <ScottK> Nope.
[01:50] <ogra> sorry, mixed you up with Steven ... i should really go to bed
[01:50] <ScottK> Not at any sprints.
[01:50] <ScottK> Bah.  I'm older and wiser.
[01:50] <ogra> that didnt happen to me since half a year or so, dang
[01:53] <kirkland> ScottK: anything in particular you were curious about, from the Server Meeting?
[01:54] <ScottK> on the phone ....
[01:54] <ScottK> I was hoping there was a hord of MIR writers suddenly appeared.
[01:55] <ScottK> ... off the phone
[01:55] <kirkland> ScottK: :-)
[01:55] <kirkland> ScottK: I wrote 4 for ecryptfs-utils, and friends...  really really really tedious work
[01:55] <kirkland> ScottK: and I have one more to write, for keyutils
[01:55] <ScottK> Yeah.  I did about a dozen for amavisd-new in Hardy.
[01:56] <kirkland> ScottK: there was consensus that AV in Intrepid is important
[01:56] <ScottK> Right, so someone else please help ...
[01:56] <ScottK> If I sound a bit grumpy about no one other than sommer actually doing any of the work, I am.
[01:57] <kirkland> ScottK: understood.  Unfortunately, I'm falling a bit behind on my items, atm.
[01:57] <kirkland> ScottK: iSCSI is kicking my butt
[01:57] <ScottK> I can imagine.
[01:58] <ogra> you can always fall back to nbd :P
[01:58]  * ogra hides
[01:59] <zul> hah
[02:00] <zul> ogra: so helpful ;)
[02:13] <Crewsr3> is anyone around to answer an ubuntu server question?
[02:14] <PrivateVoid> I can try...
[02:15] <Crewsr3> what i want to do is install ubuntu server 64 and then install vmware server and then run 32 gui of ubuntu amount other things
[02:15] <Crewsr3> do you think it would be hard to do
[02:15] <Crewsr3> amout=amoung
[02:16] <Crewsr3> among
[02:16] <PrivateVoid> no...
[02:16] <Crewsr3> Ok I can't spell
[02:16] <PrivateVoid> are you going to leave the DE/WM off the server?
[02:16] <PrivateVoid> or are you going to install a DE/WM on the server?
[02:16] <Crewsr3> Your going to half to help me out what is DE/MM
[02:16] <PrivateVoid> Desktop Environment and WIndows Manager
[02:17] <PrivateVoid> there are ways to do VMs with no GUI on the server, but you can access the VM Guest through VNC etc
[02:17] <Crewsr3> No I did not want to I wanted to have it run only the command line and then have the vm with a GUI
[02:17] <PrivateVoid> yes that can be done... there are some FAQs on doing it
[02:17] <PrivateVoid> through the FAQs I have seen are for KVM not VMWare server
[02:17] <Crewsr3> so when I start up and ubuntu sever is ready can I just start the vm from the command line
[02:18] <PrivateVoid> yep
[02:18] <Crewsr3> and then will the gui kick start
[02:18] <PrivateVoid> not sure if you get a GUI on the server... or if you will need to VNC from a client
[02:18] <PrivateVoid> I have not actually completed the task on my server, but read the FAQs and heard other people claim to have had success
[02:19] <Crewsr3> so if my main box is headless and I have a client ssh into the server, then I can access the gui of the VM
[02:19] <PrivateVoid> or use VNC
[02:19] <PrivateVoid> yes
[02:19] <PrivateVoid> and you can GUI through SSH from what I have been told as well.
[02:20] <Crewsr3> How powerful do I need the client to be in order to enjoy good compfuzion
[02:20] <Crewsr3> or can I even use comp fuzion
[02:20] <PrivateVoid> well you are not going to get comp-fusion on a VM
[02:20] <PrivateVoid> no hardware accel
[02:21] <Crewsr3> right now I have a fairly powerful GPU and would like to make use of it
[02:21] <Crewsr3> so it sounds like this is not an ideal setup
[02:21] <PrivateVoid> Then you would want to install the client on the machine directly
[02:21] <PrivateVoid> VMs are really for testing or other uses...
[02:21] <PrivateVoid> and do not feature full hardware usage
[02:22] <Crewsr3> that makes sense, that is what I use them for now
[02:22] <Crewsr3> I have my windoz VM for windoz only apps and then I use it to play around with other distros
[02:23] <Crewsr3> are you a command line pro?
[02:23] <PrivateVoid> not really
[02:23] <PrivateVoid> I do what I need to and Google the rest
[02:23] <Crewsr3> Google is my best friend
[02:24] <Crewsr3> thanks for your help
[02:24] <Crewsr3> take care and have a good one
[02:24] <PrivateVoid> you too
[04:30] <ishkur_> quick question, im trying to setup a VM in Virtualbox. what is the kernel that comes with Ubuntu SE 8.0.4.1 ?
[04:54] <fujin> Hi, is there a preferred method of upping the number of file descriptors for an application?
[04:54] <fujin> I'm getting: general: error: socket: too many open file descriptors
[04:54] <fujin> with BIND
[07:59] <kraut> moin
[09:21] <uvirtbot`> New bug: #255251 in amavisd-new (universe) "One line is not commented out in 40-policy_banks" [Undecided,New] https://launchpad.net/bugs/255251
[09:44] <edmoore> So I have two 250GB hdds, and i want to do software raid 1. This is a largely personal headless server, though doing a bit of webserving and providing some file storage to the outside world. But not a full-time webserver or anything. How best to partition things up?
[09:50] <thefish> edmoore: look in to lvm, it will give you some flexibility to grow/shrink volumes
[09:50] <thefish> (depending on file system)
[09:51] <thefish> also gives you some nice benefits on top like snapshots
[09:52] <edmoore> thefish: thanks for the tip, I am looking now. But say for the sake of arguement, roughly what do you think I should put where, and it what proportions?
[09:55] <hads> I usually put most of the disk into /data save a few GB for / and /var although it really depends what you're doing and where you store things.
[09:55] <thefish> ^ faster typist :)
[09:55] <uvirtbot`> thefish: Error: "faster" is not a valid command.
[09:55] <thefish> you may also want about 100M /boot
[09:56] <thefish> i also use the same mount point /data :)
[09:56] <hads> e.g. / = 2GB ext3 /var = 2GB ext3 /data = * JFS
[09:57] <hads> JFS is my preference, a lot of people like XFS
[09:57] <hads> Or just ext3 depending on what you're up to.
[09:57] <hads> If you've got multiple users then you may want a separate /home with some space.
[09:58] <hads> I don't tend to use LVM much myself unless it's a special circumstance.
[09:58] <hads> It is nice though.
[10:00] <soren> I never set up a machine of any kind without lvm.
[10:01] <hads> My /data is usually RAID5 though which you can extend these days.
[10:01]  * soren dislikes stuff like /data
[10:02] <hads> Where do you put things like virtual mail or virtual web sites?
[10:02] <soren> /srv somewhere, usually.
[10:03] <soren> /data is pointless. Everything you have you your disk is data in some form.
[10:03] <hads> srv == data really
[10:03] <henkjan> /mnt/data
[10:04] <nxvl> soren: yeah, but the name of the mount point doesn't change much
[10:04] <soren> And I *really* don't like having the several hundred GB /somedirectory for general storage.
[10:04] <nxvl> i use /usr or /var
[10:04] <nxvl> but that's a matter of tastes
[10:05] <hads> Yeah, same thing, /var or /srv will end up with several hundred GB if you decide to use that so it's six of one half a dozen of the other really.
[10:07] <soren> nxvl: Well, the name of the mount point should give some hint about what you expect to find in there. You don't rename /home to /files, for instance.
[10:07] <edmoore> so what's the purpose of /home for a largely single-user machine?
[10:07] <soren> edmoore: That's where you keep your files?
[10:07] <edmoore> i thought that was my general purpose store for everything
[10:07] <nxvl> soren: /etc is not just random and non-important stuff, but etc sounds like it is
[10:08] <edmoore> ok, just got confused by people using /usr and /var for seemingly that purpose
[10:08] <Deeps> big /usr, ln -s /usr/home /home :D
[10:08] <soren> nxvl: Yes, but there are historical reasons for that, which we cannot easily change. That's a really bad excuse for making up other bad names for mount points.
[10:08] <soren> Deeps: Please, no.
[10:09] <Deeps> the freebsd way \m/
[10:09] <hads> /srv and /data aren't any different at the end of the day. It's a preference really.
[10:09] <nxvl> soren: yes, true. But if you want to put your things in /personal i really don't care, i try to use the tree that comes with my system and put my files where they fit in there
[10:10] <hads> The FHS is weird to start with so there's no point trying to argue about it.
[10:10] <nxvl> soren: at the end, you will not touch my servers and i hardly doubt someone else will
[10:10] <Deeps> isn't freedom great?
[10:11] <soren> Besides, thinking of all of this in terms of mount points is silly. They're directories. That's all. If a particular directory is bound to have a lot of stuff in it, or if you want to contain it, make it a separate file system.
[10:11] <nxvl> yeah, directories, sorry i named it bad
[10:11] <hads> Deeps: You said it :)
[10:11] <soren> nxvl: That's beside the point, really. We're discussing what we find to be good and bad practice.
[10:12] <nxvl> i have just come from the airport and haven't sleep well
[10:12] <nxvl> soren: oh! in that case yes, you are right
[10:12] <nxvl> :D
[10:13] <hads> I probably should use /srv rather than /data - that was a convention I had before I know that /srv existed in the FHS and wasn't used for anything else.
[10:14] <nxvl> btw, does the systems creates a /data directory?
[10:14] <hads> No
[10:14] <nxvl> then why to use it?
[10:14] <soren> What I usually do is this: I make /boot a primary partion of a few hundred MB. The rest of the disk, I make a big LVM physical volume.
[10:15]  * hads goes back to work
[10:15] <soren> I make a VG out of that, and create a "root" lv of about a 1.5GB.
[10:15] <nxvl> i usually create a 10 Gb partition for / and the rest for /var
[10:16] <soren> And a "var" lv of a couple of GB, a "usr" lv of a couple of GB, and a "home" lv of about the same.
[10:16] <nxvl> and in some cases 8 Gb for /usr
[10:16] <nxvl> depend on the server
[10:16] <nxvl> i don't use lvn
[10:16] <nxvl> haven't try it yet
[10:16] <soren> If something big comes along, like a large new website I want to host, I add a new lv for it and mount it somewhere under /srv.
[10:16] <nxvl> lvm*
[10:17] <nxvl> well
[10:17] <soren> Oh, I use ext3 for the /boot partition and XFS for everything else.
[10:17] <nxvl> time ti sleep
[10:17] <nxvl> read you!
[10:17] <soren> Good night, Nicolas.
[10:17] <nxvl> have a nice day soren!
[10:17] <soren> Otherwise, I just put stuff where it belongs and keep an eye on disk usage and crank it up or divide it as necessary.
[10:19] <soren> Of course, depending on the server, the disks underneath migth be RAID1 or RAID5 or something, but in terms of allocation of the disk space I have available, that's all I do.
[10:20] <soren> LVM and XFS provides me with the tools I need to be able to manage this "on-line", so I don't have to sit around being nit-picky about my disk space, when I'm installing.
[10:20] <hads> Why have a boot partition these days?
[10:21] <_ruben> cant boot lvm
[10:21] <hads> Oh, LVM root.
[10:22] <soren> hads: Right. To be able to have root on lvm.
[13:15] <crypted> Can someone please help me settup a bridge for internet charring across many machine?
[13:15] <crypted> Please :(
[13:19] <zul> yay /usr for everything
[13:19] <sommer> morning
[14:19] <ArtimusDeathhole> hello all
[15:37] <KB3NZQ> any thoughs setting up a proxy on ubuntu 8.04 desktop or should i ask this in #ubuntu
[15:38] <crypted> My DNS are not resolving. But I can ping ip's. Please help
[15:41] <ArtimusDeathhole> crypted,  what is in your /etc/resolv.conf?  also, do you have nscd installed?
[15:47] <BUGabundo> hi there
[15:47] <BUGabundo> I'm having trouble with dovecot
[15:48] <BUGabundo> http://pastebin.ubuntu.com/34783/
[15:57] <sommer> BUGabundo: try creating the directory sudo mkdir /var/run/dovecot, then try sudo apt-get -f install
[15:57] <sommer> BUGabundo: did you manually remove that directory at some point?
[15:58] <BUGabundo> not that I remember
[15:59] <BUGabundo> sommer: nops. that didn't do it!
[15:59] <BUGabundo> chmod: cannot access `/var/run/dovecot/login': No such file or directory
[16:00] <BUGabundo> should I create that too?
[16:00] <sommer> BUGabundo: worth a try
[16:00] <BUGabundo> who should be owner?
[16:00] <BUGabundo> root or mail?
[16:00] <sommer> BUGabundo: shouldn't matter to apt-get
[16:02] <BUGabundo> but for the configure set of dovecot it might
[16:03] <BUGabundo> update-rc.d: warning: multiuser is deprecated; specify runlevels manually
[16:04] <sommer> BUGabundo: what are you trying to do?  from your pastebin I assumed you had an error when trying to install dovecot?
[16:05] <BUGabundo> I'm trying to start it
[16:05] <BUGabundo> it seems to days updates mess it up
[16:05] <BUGabundo> and it was left uncofigured
[16:05] <BUGabundo> https://bugs.launchpad.net/ubuntu/+source/dovecot/+bug/255367
[16:05] <uvirtbot`> Launchpad bug 255367 in dovecot "Package dovecot-common is not configured yet. " [Undecided,New]
[16:06] <sommer> BUGabundo: so you did have a working configuration?
[16:06] <BUGabundo> yep
[16:06] <BUGabundo> let me see if it now start
[16:06] <BUGabundo> *starts
[16:06] <uvirtbot`> New bug: #255367 in dovecot (main) "Package dovecot-common is not configured yet. " [Undecided,New] https://launchpad.net/bugs/255367
[16:06] <BUGabundo> still it gives that error / warning now
[16:07] <BUGabundo> imap(s) is now up
[16:07] <BUGabundo> is that message important at all ?!
[16:08] <sommer> BUGabundo: which release are you using?
[16:08] <BUGabundo> intrepid
[16:09] <ArtimusDeathhole> please add that to you bug filing as well please
[16:09] <BUGabundo> I'll leave it there, in case any dev stumble on it
[16:09] <BUGabundo> its already there
[16:09] <ArtimusDeathhole> k, thanks
[16:10] <BUGabundo> # dovecot --version \n 1.1.2
[16:11] <BUGabundo> https://bugs.edge.launchpad.net/ubuntu/+source/ebox/+bug/255368
[16:11] <uvirtbot`> Launchpad bug 255368 in ebox "ebox: Depends: libapache-authcookie-perl but it is not installable " [Undecided,New]
[16:11] <BUGabundo> next bug
[16:11] <BUGabundo> any tips?
[16:11] <BUGabundo> I would like to give ebox a go...
[16:14] <sommer> BUGabundo: you'll probably have to wait for a new version of ebox to be uploaded to intrepid... that's to be expected using an alpha release
[16:48] <jmedina> morning
[16:57] <lukehasnoname_> hi
[17:36] <edmoore> hi - am installing server right now, but the box is w/o network connection (non available) and therefore has obviously just failed its DHCP search
[17:36] <edmoore> is this a bad thing or can I carry on anyway?
[17:38] <Deeps> assuming you're not doing a net install, its not an issue
[17:50] <jmedina> edmoore: which version are you istalling?
[18:05] <spiekey> Hello!
[18:05] <spiekey> i there a live cd for 64bit xeon machines?
[18:07] <jmedina> I think there is not server edition livecd
[18:08] <spiekey> iam getting: chroot: cannot run command `/bin/bash': Exec format error
[18:08] <spiekey> when i try to do a chroot in knoppix
[18:08] <spiekey> any cd will do ;)
[18:09] <jmedina> spiekey: why dont you download the 64bits server edition?
[18:09] <jmedina> http://ubuntu.osuosl.org/releases/hardy/ubuntu-8.04.1-server-amd64.iso
[18:09] <jmedina> what works with intel 64bits
[18:10] <Nafallo> spiekey: the cd just mentioned have a recovery mode which uses d-i to guide you up until it leaves you in a chroot
[18:21] <spiekey> thanks!
[18:22] <spiekey> i just successfully booted :)
[18:22] <spiekey> now i can view my lvm disks with lvdisplay. But they do not exist as /dev devices :-/
[18:22] <spiekey> do i need a kernel module
[18:23] <spiekey> ?
[18:26] <soren> spiekey_: They live under /dev/<name of volume group>/
[18:26] <soren> spiekey_: Also, you'll find the corresponding device-mapper devices under /dev/mapper
[18:26] <spiekey_> yes. but they are not there when booting with the live cd
[18:26] <soren> spiekey_: "lvm vgchange -ay"
[18:29] <jmedina> that is you need to activate tehem
[18:31] <spiekey_> hmm..looks like the live cd does not have the device mapper in the kernel :-/
[19:08] <spiekey_> soren: thanks alot!
[19:09] <spiekey_> it seems to work ;)
[19:10] <zul> nxvl: ping
[19:11] <kees> kirkland: raidy goodness?
[19:11] <kirkland> kees: yeah, so i'm attacking the multi-MBR aspect of this now
[19:11] <kees> cool
[19:11] <kirkland> kees: i found a few things, but I'm missing a few others
[19:11] <kirkland> kees: so, in no particular order.....
[19:12] <kirkland> kees: who/what/when creates device.map?
[19:12] <kees> grub-install, IIRC
[19:12] <kirkland> kees: hmm, if that's the only place, it doesn't do it very smartly
[19:13] <kirkland> kees: also, i'm a bit confused by the shell script "grub-install" which is in the "grub" package, and then some other package, "grub-installer"
[19:13] <kees> no idea what "grub-installer" is.  :P
[19:14] <kees> here's the method for doing a raid1 MBR: http://www.dirigo.net/tuxTips/avoidingProblems/GrubMdMbr.php
[19:14] <kirkland> kees: ah, interesting....
[19:15] <kirkland> kees: i was trying to decide if sdb should be hd1, or hd0 also
[19:15] <kirkland> kees: if that article is correct, my hunch for hd0 is verified
[19:15] <kees> right -- grub-install doesn't like the idea of marking hd0 as the secondary device, so you have to effectively trick it.
[19:16] <kees> by claiming that the secondary drive is hd0.
[19:16] <kees> so, detecting the raidness, the secondary drive, and the boot partition number is what's required to automate that recipe.
[19:16] <kirkland> kees: yeah, but i probably don't want to write that to device.map
[19:16] <soren> kirkland: The grub-installer in the installer is different from grub-install in that it also handles foreign OS'es.
[19:16] <kees> absolutely you don't.
[19:17] <kirkland> kees: okay, then i've been grossly over-thinking this
[19:17] <kirkland> soren: good, that gels with what ogra told me :-)
[19:17] <kirkland> soren: well, that it's the one in the installer
[19:17] <kees> kirkland: I would imagine that update-grub should not be aware of any of this -- I think only grub-install (and I guess grub-installer from what soren says) need awareness.
[19:18] <kirkland> kees: yeah, my changes, so far, have been in grub-installer
[19:18] <Fenix|work> Greetings... when I SSH to an ubuntu box on the DMZ I get this huge lag as it's trying to authenticate me, where when I was on the same subnet it would be lightning fast... what's causing the hangup?
[19:18] <kees> (which reminds me, I need to do this to my newly install hardy server...)
[19:18] <kirkland> kees: in there, there's support for a couple of unique situations, like dmraid, and multipath
[19:19] <kirkland> kees: i was going to add one more for "mdadm"
[19:19] <kirkland> kees: the test, though, is *weak* in my opinion....
[19:19] <kirkland> if type dmraid >/dev/null 2>&1; then
[19:19] <kirkland> ...
[19:19] <kirkland> elif type multipath >/dev/null 2>&1; then
[19:19] <kirkland> ...
[19:19] <kirkland> i can add a
[19:19] <nxvl> zul: pong
[19:19] <kirkland> elif type mdadm  >/dev/null 2>&1; then
[19:19] <soren> Fenix|work: I'd look at DNS settings.
[19:20] <kirkland> but just searching for the command seems flimsy
[19:20] <kees> what does "type" do?
[19:20] <kees> oh
[19:20] <kees> right, checks path
[19:20] <zul> nxvl: are you working n the multiuser for dovecot?
[19:20] <kees> (I'm used to "which")
[19:20] <kees> kirkland: that does seem flimsy
[19:20] <kirkland> kees: i'll stick with type for the first pass
[19:20] <kirkland> kees: but meh
[19:20] <kees> kirkland: well, are you doing this in grub-installer or grub-install?
[19:20] <nxvl> zul: yep, i'm on it right now
[19:20] <kees> (which does the server CD use/)
[19:21] <kees> s|/|?|
[19:21] <Fenix|work> soren, I had a feeling that may be the problem.  Question is, how do I avoid it to do a DNS lookup on SSH connect.  resolv.conf is pointing to external servers and will never be able to look up my name.
[19:21] <kirkland> kees: grub-installer
[19:21] <soren> Fenix|work: "UseDNS no" in sshd_config
[19:22] <nxvl> zul: why?
[19:22] <kirkland> kees: i don't understand your question about the server CD
[19:22] <nxvl> zul: you want have any suggestions or plans on it?
[19:22] <Fenix|work> soren, thank you very much
[19:22] <Fenix|work> works like a charm
[19:22] <kees> kirkland: I just know there are differences between ubiquity and d-i
[19:23] <soren> kees: The server CD uses grub-installer, I belive.
[19:23] <soren> believe, even.
[19:23] <kirkland> kees: i'll have to check, but I have a gut feeling it's grub-installer
[19:23] <kirkland> kees: gimme a couple of minutes to clean up some proof-of-concept code
[19:24] <kirkland> kees: and i'll pastebin it to you
[19:24] <zul> nxvl: i thought you might were working on a patch or something
[19:24] <kees> cool
[19:24] <kirkland> kees: see if it's on the right track
[19:24] <kirkland> kees: not tested yet, but worth reading
[19:24] <nxvl> zul: it's just changing the line on debian/rules as i understand it
[19:25] <soren> kirkland: What happens to the disk ID's if you boot from the second disk (ie. if the first disk is dead)?
[19:25] <lukehasnoname> what is d-i, again?
[19:25] <soren> debian-installer
[19:25] <lukehasnoname> figured
[19:25] <lukehasnoname> thx
[19:26] <kirkland> soren: i think "tricking" it like kees suggested handles that, making each of sd[a,b] think they're hd0
[19:27] <kees> soren: right -- grub really does not like booting up on the failed drive if it thinks that drive is mapped to hd1.  :)
[19:27] <soren> kirkland: Well, we'd be writing the same MBR to both disks, right?
[19:28] <soren> And that mbr expects (usually) to find the grub config and stuff on "hd0" (0x80). The question is: Is the second drive suddenly magically 0x80 if that's what it's booting from.
[19:28] <soren> If so, what is there to be done, really?
[19:29] <zul> nxvl: yep it is
[19:29] <soren> I've never looked much at this problem. It's very likely that I'm missing something.
[19:29] <zul> and updating the init script
[19:30] <nxvl> zul: indeed
[19:38] <kees> soren: if the failed drive is present enough that the BIOS attempts to boot from it, there is no way to boot.  in the case of being totally failed, the second drive becomes 0x80, and grub requires that it's mbr info matches that expectation.  I haven't actually checked to see if the mbr contents are identical on both drives.  I prefer just doing both mbrs writes via grub.
[19:40] <kirkland> kees: you have to do the mbr write through grub...  raiding the devices sda1 and sdb1 keeps the data in those partitions identical, but doesn't handle the mbr, which is at the head of sda and sdb
[19:40] <soren> kees: Ah, yes. If you install them separately, then you need to be careful. I somehow imagined the entire drive to be RAID1'ed, in which case it'd be a no-brainer.
[19:40] <kirkland> kees: sorry, kees, that was not intended to be instructional :-)
[19:40] <kirkland> kees: i meant to confirm what you were saying to soren :-)
[19:41] <kees> kirkland, soren: cool, yeah.  I think we're all on the same page.  :)
[19:43] <kirkland> kees: have a look at http://pastebin.ubuntu.com/34856/
[19:43] <kirkland> kees: i'm reading the grub docs now, for doing the grub shell bit in an automated manner
[19:44] <kirkland> kees: but see if the accounting for the drive/partitions/devices looks acceptable to you
[19:44] <kirkland> kees: i'm using sed b/c no one else in the script is using awk :-)
[19:52] <kirkland> kees: did I lose you?  :-)
[19:55] <Koon> kirkland: I may have scared him in another channel.
[19:56] <kirkland> Koon: yeah, he's reeling
[19:56] <kirkland> :-)
[20:01] <nxvl> zul: the init doesn't need to be modified
[20:02] <zul> nxvl: no but the information in the init script should be updated
[20:03] <nxvl> zul: of dovecot-common?
[20:03]  * nxvl is lost
[20:04] <zul> nxvl: just a sec
[20:04] <nxvl> zul: i have reverted the change and use the debian line on debian/rules
[20:05] <zul> nxvl: https://lists.ubuntu.com/archives/ubuntu-devel-announce/2008-June/000430.html
[20:05] <r00tintheb0x> Hey, im trying to automatically download an attachment out of an email from our server.
[20:05] <r00tintheb0x> Anyone know of a simple way to do that?
[20:07] <nxvl> zul: oh! now i see
[20:07] <nxvl> zul: thank you!
[20:07] <zul> nxvl: np
[20:13]  * delcoyote hi
[20:34] <nxvl> zul: patch uploaded, sponsorship will be really apreciated :D
[20:39] <zul> nxvl: sure
[21:28] <kirkland> kees: did you get a look at that proof-of-concept code
[21:33] <nxvl> zul: did you look at the patch?
[21:33] <nxvl> zul: Bug #255367
[21:33] <uvirtbot`> Launchpad bug 255367 in dovecot "update-rc.d: warning: multiuser is deprecated; specify runlevels manually" [Undecided,Confirmed] https://launchpad.net/bugs/255367
[21:37] <nxvl> kirkland: i can't open https://wiki.ubuntu.com/EncryptedPrivateDirectory
[21:37] <nxvl> kirkland: can you?
[21:37] <kirkland> nope
[21:37] <nxvl> kirkland: i'm getting 500
[21:37] <kirkland> nxvl: me too
[21:37] <nxvl> so it's a wiki problem
[21:37] <nxvl> :S
[21:37] <kirkland> nxvl: something about a wiki upgrade
[21:37] <nxvl> i was trying to get in to test it :(
[21:38] <kirkland> that sucks...  right when i do my blog post
[21:38] <jpds> kirkland, nxvl: I suggest asking in #canonical-sysadmin, just they may still be fixing upgrade problems.
[21:38] <kirkland> jpds: ahhhh..... so you *are* having wiki problems :-)
[21:38] <jpds> kirkland: Am now.
[21:38] <kirkland> jpds: :-P
[21:38] <nxvl> kirkland: one time i did a blog post with some information on it calling for helpers and my server go down, that was horrible
[21:39]  * jpds has had a Squid error and an Internal Server Error so far.
[21:39] <kirkland> back up, perhaps?
[21:39] <kirkland> quick, cache all of the wiki!!!!
[21:40] <jpds> Yep, seems to be having fun falling off the net.
[21:41] <jpds> kirkland: I read that page yesterday, will ~/Private be hard-coded or will we be able to change the name?
[21:41] <kirkland> jpds: at the moment, it's #define'd
[21:42] <kirkland> jpds: needs to be, for security purposes
[21:42] <kirkland> jpds: mount.ecryptfs_private is a setuid binary
[21:42] <jpds> kirkland: Shame... I really dislike files/folder with Capital letters.
[21:42] <nxvl> kirkland: wiki alive again
[21:42] <kirkland> jpds: there are some nasty things that a user can do, if they can choose the mountpoint
[21:43] <kirkland> jpds: we might be able to provide a system-wide, root-configurable setting in /etc
[21:44] <kirkland> jpds: you can add your usecase/comments to https://bugs.launchpad.net/ubuntu/+source/ecryptfs-utils/+bug/247421 if you wish
[21:44] <uvirtbot`> Launchpad bug 247421 in ecryptfs-utils "ecryptfs-utils private directory should support translations of "Private"" [Wishlist,Triaged]
[21:45] <jpds> kirkland: Can it not be defined in xdg-user-dirs ?
[21:47] <kirkland> jpds: it's on my list of things to do
[21:47] <kirkland> jpds: unfortunately, i've been buried in iSCSI hell for the last few weeks
[21:48] <kirkland> jpds: ultimately, as long as the root user controls the name of the dir, i think it'll work
[21:48] <jpds> kirkland: OK, great. Thanks for the great work. :)
[21:49] <kirkland> jpds: it's really just a matter of keeping non-priv users from choosing the string arbitrarily
[21:49]  * nxvl is installing encryptfs
[22:00] <kees> kirkland: yeah, sorry, had to run off to other stuff, let me read it and get back to you in a bit
[22:00] <kirkland> kees: no worries, i'm running with it and testing it
[22:00] <kees> kirkland: I get "Filesystem" out of that df
[22:01] <kees> kirkland: does $ROOT include /boot/ ?
[22:02] <kees> kirkland: probably skip the if mdadm --detail test and go straight to the processing, if it fails, the list will be empty
[22:04] <kirkland> kees: oh, sorry, $ROOT is something in that script
[22:04] <kirkland> kees: okay, that was set -e proofing, i though
[22:05] <kees> grub_partnum=$(( partnum - 1 ))
[22:05] <kees> I'd use that instead of expr (saves an exec of expr)
[22:05] <kees> also, I like seeing $() over 11
[22:05] <kees> er
[22:05] <kees> ``
[22:05] <kirkland> kees: k
[22:05] <kees> looks like the right stuff to me.  :)
[22:07] <spader3d_2> hi all
[22:07] <spader3d_2> how do i open port 53 domain on my machine?
[22:07] <kees> kirkland: I'd be curious about possible failure conditions, but I'm not feeling very creative at the moment.  I think best failure mode to catch is "if type mdadm" but nothing in mdadm --detail (i.e. not on md, but mdadm installed)
[22:07] <kees> kirkland: I'll be back on in a bit...
[22:09] <kirkland> these wiki problems are absolutely killing my productivity
[22:09] <kirkland> kees: doh....
[22:09] <kirkland>         md=`df "$ROOT/boot/grub" | sed "s/ .*//"`
[22:10] <kees> okay, cool, that's what I thought.  :)
[22:10] <kees> kirkland: btw, where is the source for grub-installer, just so I can peek at it?
[22:10] <kirkland> kees: yeah, sorry, got lost between the 4 cut and pastes around terminals and pastebins
[22:10] <kees> hehe, no problem.
[22:10] <kirkland> kees: apt-get source grub-installer
[22:10] <kees> heh.  go figure.  ;)
[22:11] <kirkland> kees: let me get something working before you waste much time on this
[22:11] <kees> kirkland: yeah, I just wanted to see the basic routine
[22:11] <kirkland> kees: i just wanted a checkpoint that these are the things that need to be done
[22:14] <thenewguy_> anybody  have any links/tips on setting up seaside / squeak with apache on ubuntu server? i dont know where to start
[22:37] <kirkland> kees: grub-installer....  gross....  there are a bunch of bash functions, followed by 340 lines of shell code, followed by a bunch more functions
[22:38] <kirkland> kees: I think I'm going to need to bust up that middle blob into functions, such that we can call them iteratively
[22:46] <Fenix|work> if I want to download packages from packages.ubuntu.com... where do I place these downloaded packages on my ubuntu server
[22:47] <Fenix|work> this box as of yet doesn't have a direct http/ftp connection to the outside
[23:32] <zul> nxvl: yep will upload it tomorrow
[23:37] <nxvl> zul: :D