[00:44] <mtx_init> How do you do a debootstrap of ubuntu server?
[01:12] <hallyn> stgraber: heh, it was fine - cgroup-bin just refused to actually remove itself
[01:12] <hallyn> (and as a result libvirt was not starting, so my containers couldn't attach to virbr0)
[01:13] <hallyn> stgraber: the package at http://people.canonical.com/~serge/cgroup-lite  should work
[01:14] <The_Rufus> for the love of all that is holy, please, can someone please help me get Ubuntu Server 11 working with a RocketRaid 2320 RAID card?
[01:17] <hallyn> The_Rufus: (i probably can't help, but someone probably can) where precisely is it failing
[01:19] <The_Rufus> I've installed Ubuntu server 11 x64 to a USB flash drive, I have 5x2Tb SATA drives on a RocketRaid 2320. US-s is installed on the drive and working fine
[01:19] <The_Rufus> but no matter what I try, I can't get RR2320 drivers loaded
[01:20] <The_Rufus> i've follwed every howto, downloaded all the drivers (ubuntu specific AND open source) from the HighPoint website
[01:21] <The_Rufus> the ubuntu specific drivers are too old (ubuntu 9) and the open source generic drivers are missing a header file and won't compile
[01:21] <tarvid> what is the best way to do dyndns like services using bind on ubuntu server?
[01:22] <hallyn> The_Rufus: where do you get the open source drivers, and can you pastebin the output showing msising header file
[01:22] <hallyn> I'm wondering whether we can find the header file for you
[01:23] <The_Rufus> http://www.highpoint-tech.cn/BIOS_Driver/page/rr232x.htm <--drivers aquired from there
[01:33] <The_Rufus> how can I pastebin the error?
[01:33] <hallyn> The_Rufus: which part is failing?  'install.sh'?  pre-inst?
[01:33] <hallyn> The_Rufus: if it's install, for instance, you can do
[01:34] <hallyn> 'install.sh > output 2>&1'
[01:34] <hallyn> then 'pastebinit output'
[01:34] <hallyn> and paste the resulting url here (or apt-get install pastebinit and do it again :)
[01:34] <The_Rufus> well, i'm doing a make
[01:35] <hallyn> of what?  did you grab the ubuntu 9.04 package?
[01:35] <The_Rufus> so that would be : make > output 2>&1
[01:35] <The_Rufus> ?
[01:35] <The_Rufus> no, open source driver
[01:36] <The_Rufus> when I try the Ubuntu driver, it just says that ....insert is for linux kernel 2.6
[01:36] <The_Rufus> .....ot the system to use the new driver module
[01:36] <hallyn> ok - yeah make > output 2>&1 works
[01:36] <hallyn> if you haven't yet, 'apt-get install linux-headers' would hopefully give you waht you need.
[01:36] <The_Rufus> i say .... because I can't see the first 5 characters of my terminal because bash doesn't work with my 22" widescreen lcd
[01:38] <The_Rufus> ok, pastebinit pasted to something......ubuntu.com/668763
[01:40] <hallyn> uh, that looks successful.  do you have any .ko files in that tree?  ('find . -name "*.ko"')
[01:40] <The_Rufus> nothing
[01:41] <The_Rufus> can you give me the url to my pastebin please?
[01:47] <hallyn> http://paste.ubuntu.com/668763/ ?
[01:47] <The_Rufus> i worked it out
[01:47] <The_Rufus> but that's NOT what's on the screen
[01:47] <The_Rufus> I get (among other things): fatal error: linux/config.h: No such file or directory
[01:48] <The_Rufus> [/home/rufus/rr/rr232x-linux-src-v1.10/product/rr232x/linux/.build/os_linux.o] Error 1
[01:49] <The_Rufus> [_module_/home/rufus/rr/rr232x-linux-src-v1.10/product/rr232x/linux/.build] Error 2
[01:49] <The_Rufus> rr232x.ko] Error 2
[01:50] <hallyn> hm, it says latest supported kernel is 2.6.25
[01:51] <hallyn> i would contact them and ask for any updated drivers
[01:51] <The_Rufus> damnit
[01:51] <The_Rufus> so if I install ubuntu server 10, would that work?
[01:51] <hallyn> 10.04  was based on 2.6.32
[01:51] <The_Rufus> gah
[01:51] <The_Rufus> no wonder people use windows
[01:52] <hallyn> 2.6.25 was april 2008
[01:52] <hallyn> nonsense, i could use bios update for my 1 year old vaio under windows 7 bc that's too new
[01:53] <hallyn> s/could/couldn't/
[01:55] <hallyn> d'oh!  My attempts at reproducing were foiled by 3.0 kernel switch :)
[01:56] <twb> Dump the lspci -n into kmuto.jp's HCL
[02:01]  * hallyn wonders who twb was talking to
[02:02] <mtx_init> How do you do a debootstrap of ubuntu server?
[02:03] <twb> hallyn: re rr232x, whatever that is
[02:04] <hallyn> mtx_init: debootstrap lucid destdir
[02:06] <mtx_init> hallyn: How is that any different than regular lucid and when regular lucid becomes EOL, what happens as lucid server is supposed to keep going.
[02:06] <hallyn> mtx_init: i don't understand the question
[02:06] <hallyn> twb: The_Rufus: oh yeah, this rings a bell - linux/config.h did in fact disappear at one point
[02:07] <hallyn> (or maybe rather, was there at one point :)
[02:09] <The_Rufus> hallyn: and....? how can I get it going?
[02:11] <hallyn> The_Rufus: well, you can comment out line 12 in osm/linux/osm_linux.h, but then you run into plenty of other compile failures.
[02:11] <hallyn> userspace ABI is consistent, but in-kernel is not :(  ask vendor for new source
[02:12] <mtx_init> hallyn: lucid desktop is supported to 2012, server to 2015.
[02:13] <The_Rufus> that didn't work
[02:16] <mtx_init> hallyn: so if they have different end of life dates, how will the system know if not installed from a specific media
[02:22] <hallyn> mtx_init: good question.  My understanding is that so long as you do a minimal install and stick to software from server set, you'll be supported.
[02:23] <hallyn> in other words, if you file a bug against postfix in lucid, it will be addressed.
[02:23] <Corey> hallyn: Is that an example, or is there a postfix bug?
[02:23] <hallyn> Corey: just an example :)
[02:24] <hallyn> I'm sure it is bug-free
[02:24]  * hallyn runs out the door
[02:24] <Corey> *snicker*
[02:57] <jcastro> kirkland: RoAkSoAx: woo, provisioning my first orchestra client!
[02:58] <bkerensa> jcastro: cool
[05:02] <_aaron_> any one have information on creating a RIS server in ubuntu
[05:03] <_aaron_> !RIS
[05:05] <_aaron_> going to ask in #ubuntu
[05:08] <twb> Maybe if you told us what a RIS is
[05:08] <twb> Reparto Informazioni e Sicurezza, a military intelligence agency of Italy, see SIOS ?
[05:10] <twb> Routing Information Service project of RIPE NCC?
[05:10] <_aaron_> a pxe server for installing os's over network
[05:10] <twb> Oh, that
[05:10] <twb> SUre, we can help with that
[05:11] <twb> A lightweight solution is to roll your own with dnsmasq, or to use di-netboot-assistant.
[05:11] <twb> Heavyweight solutions include FAI (old) and Orchestra (new)
[05:13] <_aaron_> where can I find information on orchestra
[05:13] <twb> Dunno
[05:13] <twb> It's some new-fangled thing that the locals are gushing over
[05:13] <twb> Probably because they're writing it
[05:14] <twb> https://launchpad.net/orchestra or something
[05:15] <_aaron_> can I get the info for the di-netboot
[05:15] <twb> apt-get install di-netboot-assistant
[05:15] <twb> And read the appendices of the install guide
[05:18] <_aaron_> i'm looking at something like FAI but I need it to also install M$
[05:19] <twb> Sorry, I don't do Windows
[05:19] <_aaron_> I don't like to either but this is for work and well execpt for the servers it's all M$
[05:20] <twb> Shrug
[05:21] <twb> You can talk to ##windows about it; I don't know anything about Microsoft systems.
[05:28] <_aaron_> ##windows ?
[05:28] <twb> Er, you know how to use IRC, right?
[05:29] <_aaron_> kinda i know how to join channels
[05:29] <_aaron_> lol
[07:57] <twb> So guess what I get to do today
[07:57] <twb> Make slapo-ppolicy work with xdm
[07:58] <twb> That's right, we signed off on delivering password expiry under lucid at fixed price, forgetting that we had to switch from gdm to xdm when we migrated from hardy to lucid (I forget why).
[07:58] <twb> This is going to be about as fun as home trepanation
[09:55] <philipballew> does ubuntu still support floppy disks?
[10:03] <soren> philipballew: It should, yes.
[10:03] <soren> I've never ever tested it, though.
[10:03] <philipballew> would i mount that any differently?
[10:03] <soren> ..but the drivers are there.
[10:03] <philipballew> i just found a floppy i need to read
[10:03] <philipballew> so 90's....
[10:03] <soren> philipballew: Differently from what?
[10:04] <philipballew> any other command to mount anything
[10:04] <soren> No.
[10:04] <soren> Back in the old days, you'd "sudo mount /dev/fd0 /mnt"
[10:06]  * smb thinks there also used to be mdir, mcopy... loooong way back
[10:30] <_ruben> i usualy use(d) the mtools as well
[11:22] <ivoks> only mtools :)
[11:39] <reisi> does anyone have a clue on how to recover when a dpkg package list file goes corrupted?
[11:44] <Daviey> soren: Am i correct in saying nobody has started packaging keystone?
[11:45] <Daviey> oh wait
[11:51] <Daviey> soren: Would you mind pushing your branch to the team namespace?  Anything else need doing before uploading it?
[12:09] <soren> Daviey: I thought zul took that and worked some more on it.
[12:09] <soren> Daviey: But sure, I can move it.
[12:10] <soren> Daviey: https://code.launchpad.net/~openstack-ubuntu-packagers/keystone/ubuntu
[12:10] <soren> Daviey: Enjoy.
[12:16] <lynxman> soren: btw sent a merge request your way for upstartified swift init scripts
[12:17] <lynxman> that one :)
[12:17] <lynxman> does somebody know adam_g launchpad profile name?
[12:19] <lynxman> found it
[12:28] <Daviey> soren: Yeah, zul mentioned he was working on it - but he isn't around at the moment, and i didn't know if he knew you had started it
[12:28] <KM0201> how do i put a users home folder, on my NAS?
[12:28] <KM0201> i've already created the user.
[12:29] <KM0201> let me clarify, i know how to make a users home directory on my NAS,  but i have a folder on my NAS, for this user already, and don't want to lose the data in there
[12:30] <KM0201> sudo useradd -d /media/NAS/username -m username    --- i get an error the home directory already exists
[12:31] <Pici> KM0201: How about using usermod -d if the user already exists?
[12:32] <KM0201> Pici: never used that.. so you're saying create the user normally... then use that, and point it at the folder on my NAS, right?
[12:32] <Pici> KM0201: I thought you said that you already created the user?
[12:32] <KM0201> no...
[12:32] <KM0201> when i tried to create the user (using the command i created above)... .. that creates the user, and puts their home directory on my NAS
[12:33] <KM0201> then i just give them a password, and thats that...
[12:33] <KM0201> but i think what you've said will work, lemme add the user the normal way, and try that
[12:36] <Pici> KM0201: If you specify -m with useradd then it will fail, just use -d
[12:36] <KM0201> ok.
[12:39] <KM0201> that seems to have done the trick
[12:39] <KM0201> thanks
[12:40] <Pici> np
[12:44] <KM0201> i think it worked anyway
[12:44] <KM0201> i'll know in a minute.
[12:44] <KM0201> is there a way i can tell if that's the home directory assigned to that user.
[12:45] <Pici> KM0201: getent passwd theusername
[12:45] <Pici> (same as grepping for theusername in /etc/passwd
[12:46] <KM0201> looks like it worked
[12:46] <KM0201> i'll know for sure in a minute.
[13:05] <hggdh> Daviey: so, how is ti now?
[13:05] <hggdh> (and good morning)
[13:15] <KM0201> Pici: seems to have worked perfectly, thanks.
[13:16] <Pici> KM0201: great
[13:28] <kim0> smoser: hey there, is ubuntu-vm-builder discontinued ? should I remove it from docs?
[13:31] <kim0> Daviey: I see your name on the blueprint for dropping vmbuilder .. is it going to be dropped for 11.10 ?
[13:36] <soren> Daviey: I've pointed him at my PPA with the packages.
[13:36] <soren> Daviey: Twice :)
[13:45] <Daviey> soren: good stuff!
[13:55] <lynxman> smoser: ping whenever you're around
[14:09] <Daviey> lynxman: smoser is out today
[14:12] <Daviey> kim0: will now not be removed this cycle.
[14:13] <Daviey> hggdh: good morning, will look shortly
[14:13] <Daviey> Thanks for working on that, it is most apreciated!
[14:15] <lynxman> Daviey: aah k, ty :)
[14:16] <kim0> Daviey: thanks
[14:29] <cjs226> how can i forward non-syslog logs from remote Ubuntu servers to a central rsyslog server?  I've tried snare's epilog but it's not logging correctly
[14:31] <patdk-wk> non-syslog logs?
[14:31] <patdk-wk> just have rsyslog read them
[14:31] <cjs226> logs for applications to don't write to syslog
[14:32] <patdk-wk> http://www.rsyslog.com/doc/imfile.html
[14:32] <cjs226> thx!
[14:33] <patdk-wk> normally though, I will pipe the log output from the prorgam to something like logger though
[14:33] <patdk-wk> actually, I have some perl scripts I made to do that, for logging apache log files
[14:34] <cjs226> why would you prefer to do that vs. using imfile?
[14:35] <patdk-wk> cause I don't care to have the log exist on the drive, and mainly don't wan the i/o overhead required to save the logfile in the first place
[14:35] <cjs226> ah, understand
[14:36] <patdk-wk> anything to help my 5gig disk space vm's run without using disk space, and keep i/o down so I can load up more vm's at a time helps :)
[14:36] <patdk-wk> it's a pain if they do anything to the local filesystem, that means making sure they don't overfill, clean them selfs up, and everything else
[14:37] <cjs226> yup, makes sense
[14:42] <_ruben> yet it sucks when the remote syslog server is unavail for whatever reason :)
[14:45] <patdk-wk> ruben, haven't had that issue :)
[14:46] <_ruben> hehe :)
[14:46] <patdk-wk> but then, I told syslog to hold a queue of 12hours in it's cache if it is unavailable
[14:46] <patdk-wk> never had it unavailable for >2hours
[14:46] <_ruben> ah nice, didn't know that was an option
[14:46] <patdk-wk> and that was cause of harddrive crash
[14:46] <patdk-wk> dunno in rsyslog, it is in syslog-ng
[14:47] <_ruben> interesting
[14:47] <patdk-wk> I still haven't done any tests with rsyslog to see if I want to switch to it or not
[15:15] <utlemming> Ping Daviey
[15:15] <utlemming> Ping zul
[15:16] <skrite> hey all
[15:17] <bluethundr> hey guys.. I need to install xinetd / xinetd-logrotate on jaunty (9.04) server but I dont see any jaunty packages on the ubuntu site
[15:18] <bluethundr> no longer supported? how do I get jaunty packages.. maybe backports?
[15:19] <bkerensa> bluethundr: Just find the .deb file wget it and install it?
[15:21] <skrite> every tutorial i see about building an apache and mysql cluster shows me how to set up an load ballancer haproxy, but i am wondering about a hardware load ballancer. Are they as straightforward to configure?
[15:22] <bluethundr> bkerensa: ok just need to google the .deb for the jaunty package
[15:23] <bkerensa> bluethundr: Yeah that would be the easiest
[15:24] <bluethundr> bkerensa: cool tx
[15:24] <bkerensa> bluethundr: no problem
[15:35] <kim0> any idea why vmbuilder downloads g++ and build-essentials ?
[15:38] <bkerensa> probably is prereq
[15:52] <hggdh> Daviey: my pleasure
[15:52] <bkerensa> =o
[15:56] <hallyn> Daviey: I filed bugs 828782, 828785, and 828789 for spice sync+enablement.  all of it compiles fine for me in oneiric
[16:08] <Daviey> hallyn: I will have a look at them shortly.
[16:10] <bkerensa> hallyn: I'm looking too
[16:14] <hallyn> thx
[16:16] <hallyn> oh, i didn't realize I could combine those.
[16:19] <stgraber> hallyn: good morning. Grabbing you cgroup-lite now.
[16:20] <stgraber> hallyn: first note (before installing it), I think it should be "arch: all" and not "arch: any" as it doesn't seem to contain any binary
[16:28] <bkerensa> hallyn: No problem
[16:37] <hallyn> stgraber: ok
[16:41] <stgraber> hallyn: ok, it installs fine, does what it's supposed to and I can still suspend/resume. So looks good to me.
[16:51] <hallyn> stgraber: cool.  so what's next, file a [FFE] [ITP] bug?
[16:53] <hallyn> (biab, lunch)
[16:53] <stgraber> hallyn: yes, next step would be to prepare a debdiff against libcgroup (I think it makes sense to have both cgroup-bin and cgroup-lite built from the same source)
[16:54] <stgraber> hallyn: having them conflict with each other
[16:54] <stgraber> then attaching that debdiff to the FFe bug
[16:54] <stgraber> the FFe bug should also mention changing the Recommends on the lxc package from cgroup-bin to cgroup-lite
[16:56] <hallyn> that would still include a second FFE bug against lxc though right?
[16:57] <hallyn> stgraber: ok, thx.  I'll get on that after finish qemu and libvirt merges.
[17:02] <hggdh> Daviey: what else can I do?
[17:44] <zul> hello
[17:44] <zul> hallyn: around?
[17:45] <utlemming> zul: hey do you want partitioned arm images?
[17:46] <zul> utlemming: still in vancouver but next week sure! :)
[17:46] <utlemming> zul: I'll have non-partitioned images being spit out automagically starting tonight
[17:47] <zul> utlemming: sweet ill have a look when i get back home
[17:47] <utlemming> zul: but I'll get the partitioned images going along in a bit. The code that makes the partitions assumes some _very_ x86ish assumptions, like you want grub, etc.
[17:48] <zul> utlemming: grrr ick....talk to the arm guys if you need help
[17:50] <utlemming> zul: its actually our cloud-image building code and our desire to keep a similar code tree that is causing the headaches. arm images are built 90% of the same, and the remaining 10% is where all the pain is coming from. Also, I'm installing the linaro-vexpress which should work with panda-boards (if I'm not mistaken)
[17:51] <utlemming> zul: I get my panda board whenever Robbie sends it, so these images haven't been tested on real ARM hardware.
[17:51] <utlemming> (not that I would expect it to work real well anyway)
[17:52] <Daviey> hggdh: Where do we start? :)
[17:52] <Daviey> hggdh: give me 10 mins, there are a few things to tackle
[17:52] <hallyn> zul: what's up?
[17:52] <hggdh> Daviey: heh. BTW, kees is accepting the ajaxterm patch
[17:54] <Daviey> hggdh: hah, just checked that :)
[17:54] <Daviey> hggdh: that is rocking.
[17:54] <Daviey> i love it when a plan comes together
[17:55]  * utlemming hears the A-Team theme song
[17:55] <hggdh> indeed
[17:55] <Kingsy> can someone in here help me with rebuilding a raid mirror?
[17:56] <zul> hallyn: so whats the problem with libvirt/lxc?
[17:57] <zul> utlemming: cool hopeuflly ill get libvirt on pandaboard next week
[17:57] <medberry> Daviey, thanks.
[17:57] <hallyn> zul: there's a problem?
[17:57] <hallyn> must be not enough ppl using it
[17:57] <zul> hallyn: i dunno smoser says there is a problem
[17:58] <hallyn> oh.  yes.  i don't yet know anything more than what he's said.
[17:58] <hallyn> except it's not an lxc/libvirt problem,
[17:58] <medberry> Hi Daviey, when firing up a bunch of cloud instances, we exhaust our IP pool as some machines end up with three distinct leases (although only one MAC id is involved.)
[17:58] <hallyn> it's an openstack/libvirt/lxc problem
[17:58] <zul> hallyn: did you see upstream libvirt has got block device support and loopback support fox lxc now?
[17:58] <medberry> We'd like to find a way to force isc-dhcp-server (natty) to hand out the same lease.
[17:58] <hallyn> zul: no i didn't.  that's cool.
[17:58] <Daviey> medberry: Are you seeing this in dnsmasq aswell?
[17:59] <Daviey> medberry: Also, is this a regression from dhcp-server3?
[17:59] <medberry> Daviey, no idea.
[17:59] <Daviey> medberry: If you know the mac addresses that will be needed, they can be statically defined?  Does that help?
[17:59] <Daviey> medberry: I can pastebin a config for an example, if that use case is ok?
[17:59] <medberry> it doesn't seem to help as there is already a host definition with the mac.
[18:00] <medberry> sure, I'll take a look.
[18:01] <medberry> Daviey, where/why would dnsmasq come into play? I think a full DNS server is in place.
[18:02] <Daviey> medberry: so we are doing, http://pb.daviey.com/wii2/
[18:02] <Daviey> Does that help, or are you already doing that?
[18:02] <medberry> already doing that (on a much bigger scale)
[18:02] <Daviey> medberry: I just wondered if dnsmasq handled this differently to isc-dhcpd
[18:03] <Daviey> (dnsmasq can do dhcp aswell)
[18:03] <medberry> ah. no idea.
[18:03] <Daviey> medberry: And isc-dhcp is ignoring those fixed addresses?
[18:03] <medberry> I can run a test with dnsmasq and see if it exhibits the same problelm.
[18:04] <Daviey> medberry: That would be interesting to compare.
[18:04] <Daviey> medberry: You mentioned the other day that it might be part of the spec to supply a different address?
[18:05] <medberry> it is part of the spec: varying the client-id is treated as a distinct request (per RFC). I was looking for a flag/etc to ignore client-id and always use hardware (aka MAC).
[18:05] <Daviey> wow, i didn't know that.
[18:05] <medberry> so those hardware type entries get ignored if the dhclient provides a client-id
[18:05] <Daviey> Ooooooo! i see.
[18:06] <medberry> so you can end up with multiple leases per machine as PXE's dhclient and Installer's dhclient and running OS dhclient all behave distinctly.
[18:06] <medberry> and the RFC is written so that client-id field is first/primary key and only falls back to mac if client-id isn't sent.
[18:07] <medberry> at any rate--I was just poking you as I saw your name in the changelog. Thanks for the help, I'll look at an alternative (dnsmasq) dhcp server.
[18:07] <medberry> (the closest thing I've found but is the "master hardware;" and "deny duplicates;" dhcpd.conf setting and I'm still investigating in that area.)
[18:08] <medberry> s/but//
[18:08] <Daviey> medberry: Does "deny unknown-clients;" change what happens?
[18:08] <medberry> haven't tried that one....
[18:08]  * medberry reads the man
[18:10] <Daviey> med_out: a short dynamic-bootp-lease-length XX might also be helpful?
[18:10] <Kingsy> has anyone used mdadm before?
[18:11] <Daviey> Kingsy: many people have used mdadm
[18:11] <Daviey> med_out: one-lease-per-client looks interesting
[18:12] <Kingsy> Daviey: well, I need some advice on a raid drive I have, I am sure its setup but on boot my raid controller says it needs to be rebuilt. the drive I can see on my operating system has files on it, so I am assuming that just means I need to resync the mirror? right?
[18:12] <Daviey> Kingsy: raid controller and mdadm is sort of at odds surely?
[18:13] <Kingsy> Daviey: sorry, its fake raid
[18:13] <Kingsy> onboard raid..
[18:13] <Kingsy> raid controller was the wrong choice of words there
[18:13] <Daviey> Kingsy: Why use fakeraid?
[18:14] <Kingsy> Daviey: because its onboard
[18:14] <Daviey> Kingsy: Can you disable it?
[18:14] <Daviey> Kingsy: mdadm is, unless i am mistaken, purely for entirely software raid - handled entirely from linux.
[18:15] <Daviey> Kingsy: dmraid is for fakeraid.
[18:15] <Kingsy> oh ok
[18:15] <Daviey> fakeraid can burn in a fire, it's a waste of time.
[18:15] <Daviey> You get no extra perf' and it's usually less reliable/good than linux mdad.
[18:16] <Daviey> Also, it's not hardware dependant - you can pull out the disks from a mdad based system and throw them into another box.  Good luck with that using fakeraid :)
[18:17] <Daviey> If the raid support your motherboard has doesn't steal cpu cycles, then there is some benefit.. but unless it's a good end card, i wouldn't use it.
[18:18] <Kingsy> Daviey: well I have installed dmraid, I just wanna see if I can get this rebuilt first then maybe look at some other options. you used this before?
[18:18] <CrazyGir> don't use onboard RAID
[18:18] <CrazyGir> it's a shame
[18:18] <CrazyGir> *sham
[18:18] <CrazyGir> though shame works too
[18:29] <hallyn> Daviey: gah, a little qemu-kvm+ipxe issue, are you around?
[18:30] <hallyn> Daviey: in qemu-0.14.1, which we have now, it looks for /usr/share/qemu/pxe-*.bin
[18:30] <hallyn> in qemu-0.15.0, that is changed to /usr/share/qemu-*.rom
[18:30] <hallyn> Daviey: that's the only issue I'm seeing with the otherwise clean 0.15.0 qemu merge.  What is the best way to handle that?  A simultaneous update of ipxe?
[18:34] <hallyn> well i guess ipxe update is the only thing that makes sense.
[18:34] <Daviey> hallyn: o/
[18:35] <Daviey> hallyn: ipxe is new this cycle, so we can largely do what we want.
[18:35] <Daviey> hallyn: the symlinks is a ubuntu delta anyway, that i am *about* to upload.
[18:35] <Daviey> So.. it seems OK to morph the symlinks :)
[18:36] <hallyn> Daviey: ok.  Morphing them before I upload qemu 0.15.0 is wrong too, we have time it :)  so you can upload what's there while i test 0.15.0 some more
[18:36] <hallyn> thanks!
[18:36] <hallyn> Daviey: I'll upload a new merge proposal when I'm ready?
[18:36] <Daviey> I am wondering if a "for *.bin do symlink /foo/*.bin /bar/*.rom" in debian/rules
[18:36] <hallyn> or maybe i should just take the plunge and uplaod qemu-kvm
[18:36] <Daviey> But a static list is also ok, as you have done
[18:37] <hallyn> yeah and trying to get the order right of dh_links and the rest of debian/rules can get flaky :)
[18:37] <Daviey> hallyn: Does qemu have any new features?
[18:38] <Daviey> hallyn: I think, considering how vital qemu is - we probably should have a FFe where the merge is discussed.
[18:39] <hallyn> oh ye...
[18:39] <hallyn> bug 827831
[18:39] <hallyn> feature list is at http://wiki.qemu.org/Planning/0.15.  Now I don't know if usb 2.0 actually made it.  If so that'd be killer.
[18:40] <Daviey> hallyn: Can you attach an upstream changelog diff to the bug?
[18:41] <hallyn> no
[18:41] <Daviey> meanie
[18:41] <hallyn> Changelog is un-updated,  and git log is huge
[18:41] <Daviey> bah
[18:41] <hallyn> It does look like usb2 is supported...  haven't tested though
[18:42] <hallyn> changelogs are so pre-millenial
[18:43] <hallyn> Daviey: actually http://wiki.qemu.org/ChangeLog/0.15   looks like it
[18:43] <tdn> I am trying to enable connection sharing as described here: http://blogs.perl.org/users/smylers/2011/08/ssh-productivity-tips.html -- I have added "ControlMaster auto" and "ControlPath /tmp/ssh_mux_%h_%p_%r" to my .ssh/config, however, I do not see any file created in /tmp, nor do I notice any speedup in connections. I do not think this setting has any effect. If I check netstat -tn, I see that there is created a new tcp connection each time I ss
[18:43] <Daviey> hallyn: throw that on the bug.
[18:44] <hallyn> yup
[18:44] <hallyn> and attaching a debdiff
[18:45] <Kingsy> yeah it turns out it is fakeraid.. and I have started with it, so can someone give me some advice on dmraid?
[18:45] <Daviey> hallyn: good stuff.
[18:45] <Daviey> Kingsy: My advise is "don't" :)
[18:45] <Daviey> !fakeraid Kingsy
[18:46] <Daviey> !fakeraid
[18:46] <geekbri> I've installed perl and a ton of perl modules all through apt-get on ubuntu server 10.04 LTS, but when i try to use the zlib module, i get the following error Compress::Raw::Zlib version 2.037 required--this is only version 2.023
[18:46] <geekbri> Any clue why that would be coming up if ive installed everything through apt?
[18:46] <Daviey> geekbri: Purely installed via apt-get and not cpan?
[18:46] <Kingsy> basically dmraid -help|grep rebuild
[18:46] <geekbri> Daviey: yes, i have some modules installed through CPAN but they are not these compression modules
[18:47] <Kingsy> gives me the command I need but what does it mean by RAID-set ? how do I find out the paths of my raid set?
[18:47] <Daviey> geekbri: If you are certain this is packages at fault, can you raise a bug against the offending one?
[18:47] <Kingsy> is that whatever is in /dev/mapper/ ?
[18:47] <geekbri> Daviey: i am pretty sure, but is there a way for me to verify?
[18:48] <Daviey> geekbri: raise a bug and provide an example how to reproduce.  Someone can then verify in a clean lucid instance.
[18:48] <geekbri> Daviey: k thanks, will do.
[18:48] <Daviey> geekbri: Ie, a simple perl script and what packages it needs.
[18:56] <Kingsy> ok this is a new one, my hd light has been on contstantly since I booted about 15 mins ago.. how can I see what its doin?
[18:56] <Kingsy> is it possible its resyncing the raid array automatically or something?
[18:59] <patdk-wk> kingsy, possible
[18:59] <RoAkSoAx> Daviey: ping?
[19:00] <Kingsy> something called kworker/0:0 is using the hd.. what is that?
[19:09] <Daviey> RoAkSoAx: hola
[19:09] <patdk-wk> hmm, I have no kworker's
[19:10] <Kingsy> patdk-wk: according to google its a "kernel bug" whatever that means, so I just need to either go back to windows or just let this random wakeup prog abuse my hard drive ?
[19:11] <patdk-wk> why not use a more stable verson?
[19:11] <hallyn> Daviey: why did you think that "Does this require a new build dep of librbd-dev (universe)?"
[19:11] <hallyn> Daviey: it compiled and installed fine on regular oneiric
[19:11] <Kingsy> patdk-wk: I am using Linux mars 2.6.38-10-generic #46-Ubuntu SMP Tue Jun 28 15:07:17 UTC 2011 x86_64 x86_64 x86_64 GNU/Linux
[19:11] <Kingsy> thats stable
[19:12] <patdk-wk> depends on how you define stable then
[19:12] <patdk-wk> I would consider when using ubuntu, LTS to be stable :)
[19:12] <Kingsy> patdk-wk: well I downloaded the official stable iso from ubuntu.com
[19:13] <Kingsy> but this really doesnt help.. does this mean that ubuntu have allowed me to download install and configure an operating system that is broken?
[19:13] <Kingsy> pretty crazy
[19:13] <patdk-wk> depends on your definition of broken also
[19:13] <patdk-wk> I wouldn't really call that broken
[19:14] <Kingsy> patdk-wk: fair enough, but its not exactly great
[19:15] <Kingsy> saying as though its a brand new installation and now I am going to have to try and "fix" the kernel
[19:15] <patdk-wk> much better than what karmic did to my laptop
[19:15] <patdk-wk> it would wipe everything on the harddrive, if you used the wifi
[19:16] <patdk-wk> had to install and upgrade it with wifi disabled, then reboot, and I could turn on the wifi
[19:17] <Kingsy> maybe i should try and apgrade the kernel
[19:17] <Kingsy> upgrade
[19:18] <RoAkSoAx> Daviey: hola!! and never mind found the solution to what I was gonna ask you :)
[19:21] <sponzor> hi. i m installing ubuntu server and i got unable to install grub in dev sda. so i change it to /dev/md0 i got the same problem ( i put boot to raid 1) what should i do?
[19:23] <Daviey> RoAkSoAx: you can't just say that.... what was the problem, dammit! :)
[19:52] <RoAkSoAx> Daviey: postinst wouldn't return when restarting apache2 and there wasn't apparent error messages. so I just tried with apache2ctl and it restarts correctly and post-inst returns
[20:09] <jcastro> RoAkSoAx: I think I ran into that mirror/squid problem you guys were having
[20:09] <jcastro> where it can't find a mirror
[20:09] <jcastro> RoAkSoAx: the other stuff appeared to work well though, the PXE menu with the proper ISOs, etc.
[20:16] <RoAkSoAx> jcastro: yeah Idk if that's squid or not cause today i imported and ISO. cleaned squid's cache, and still failed to install with not finding the kernel modules
[20:17] <RoAkSoAx> jcastro: so I'm guessin that today's ISO has a older upload of the kernel than the archives
[20:17] <jcastro> ah
[20:17] <jcastro> well, shouldn't like the natty ISOs work?
[20:17] <RoAkSoAx> s/upload/revision
[20:17] <RoAkSoAx> jcastro: those should
[20:17] <jcastro> hmm I got the same problem on any iso I tried
[20:17] <RoAkSoAx> jcastro: those not working?
[20:18] <RoAkSoAx> uhm interesting... are ytou using squid or squid-deb-proxy
[20:18] <RoAkSoAx> ?
[20:22] <RoAkSoAx> jcastro: unless the issue is with some corrupts Release files or similar
[20:29] <jcastro> RoAkSoAx: I was using whatever is stock in oneiric
[20:30] <RoAkSoAx> jcastro: you mean orchestra?
[20:30] <jcastro> right
[20:36] <RoAkSoAx> jcastro: uhmm check that squid is running cause i think there's a bug a in the squid conig
[20:36] <Daviey> jcastro: keep up, keep up.
[20:38] <bernhard2> cant seem to enable tls on exim4 anybody want to shed some light on my issues ?
[21:01] <Kingsy> can someone tell me what application rebuilds/updates raid1 in ubuntu by default?
[21:02] <fhh> mdadm ?
[21:03] <Kingsy> fhh: and if the raid is fake?
[21:03] <Kingsy> dmraid?
[21:04] <Kingsy> fhh: cos in this case the raid controller is onboard and the drives are in /dev/mapper/ so they are fakeraid
[21:07] <fhh> so dmraid
[21:09] <Kingsy> fhh: ok, so if dmraid is syncing my raid mirror how can I tell ? are you aware of any dmraid --status type commands?
[21:10] <Kingsy> fhh: basically the hdd on my pc is on constantly, I need to find out if thats dmraid syncing my drives or if its some rogue process that needs to be fixed...
[21:10] <Kingsy> can yo think of a way of finding that out?
[21:10] <Kingsy> you**
[21:12] <Kingsy> fhh: iotop says that kworker is read/writing 70-100MB/s but that is the only information I can find, it doesnt trace ot na
[21:12] <Kingsy> it back to an application**
[21:13] <fhh> im not specialist in fakeraid...
[21:13] <RoyK> Kingsy: just use software raid
[21:14] <RoyK> works well
[21:14] <vraa_> is everything in /var/log in logrotated by default, or i have to add it in the conf manully
[21:14] <Kingsy> RoyK: lol it might but I am trying to solve this problem first, not look for an alternative solution
[21:15] <RoyK> Kingsy: fakeraid sucks performance-wise
[21:15] <RoyK> better use software raid
[21:15] <RoyK> or get a decent raid controller
[21:15] <fhh> Kingsy: like RoyK i prefer soft raid or real hard raid...
[21:15] <Kingsy> RoyK: ok, i agree but are you aware of how to solve this problem?
[21:15] <RoyK> fhh: I like good hardware raid systems for the root
[21:16] <RoyK> Kingsy: is this a system you're trying to fix, or is it a new install?
[21:17] <Kingsy> RoyK: its something I am trying to fix, but tbh I don't even know if it needs *fixing* I just need to know what the hell the hdd is doing at 70-100MB/s constantly
[21:17] <RoyK> if it's a new install, never mind the so-called raid controller, just use software raid
[21:17] <Kingsy> RoyK: naa its not a new install
[21:17] <RoyK> no idea about fakeraid stuff
[21:17] <RoyK> too many variables there
[21:18]  * RoyK uses hardware raid or softraid for mirrored roots or zfs for storage
[21:18] <RoyK> the latter on non-linux
[21:18] <Kingsy> RoyK: ok, let me suggest this then, say I wanted to move from fake to soft raid. can I move to that WITHOUT formatting the hd I want to raid?
[21:19] <RoyK> I daresay that would be hard
[21:19] <RoyK> backup/restore would be the easiest
[21:19] <Kingsy> RoyK: see I did try to get a soft raid working but I couldnt seem to do it
[21:20] <Kingsy> RoyK: I take it hdd need to be firstly set to IDE in the bios right?
[21:20] <Kingsy> then you boot into the operating system so you can see two hard drives?
[21:20] <RoyK> Kingsy: that depends on the hardware etc
[21:21] <RoyK> Kingsy: for those quasi raid systems, linux will see two drives
[21:21] <RoyK> if you don't have much data on it, reinstall with sw raid
[21:21] <RoyK> less hassle
[21:21] <Kingsy> RoyK: well basically I am new to this, and I cant seem to get any specific help so I just enabled the onboard raid and booted. it seemed to mirror great..
[21:21] <Kingsy> RoyK: well the operating system isnt installed on tehe raid drive
[21:22] <Kingsy> the raided drive is a data disk
[21:22] <RoyK> Kingsy: if the system doesn't hold much data, reinstall
[21:22] <Kingsy> thats not the point, the raid drive has 700GB worth of data I need to first backup.. in order to format the drives to setup something I don't know how to do
[21:23] <RoyK> dmraid is quite bad
[21:23] <RoyK> jus don't use dmraid
[21:23] <RoyK> just, even
[21:24] <Kingsy> RoyK: ok, but do you know where I am coming from? I am looking for advice and all ppl say are don'y use fakeraid, and I cant find any solid information on how to setup any other type..
[21:25] <vraa_> dont use fakereaid lmao
[21:30] <RoyK> Kingsy: please pastebin lshw
[21:30] <RoyK> !pastebin
[21:30] <RoyK> !pastebinit
[21:33] <Kingsy> RoyK: http://paste.ubuntu.com/669568/
[21:35] <patdk-lap> hmm, go into your bios, setup acpi without raid
[21:35] <RoyK> Kingsy: looks like a raid controller to me
[21:36] <Kingsy> oh so its not fake after all?
[21:36] <Kingsy> I assumed it was cos it was onboard
[21:36] <RoyK> doesn't look that way
[21:36] <Kingsy> doh
[21:36] <RoyK> did you configure raid in the bios or the controller's bios?
[21:37] <patdk-lap> heh? it's intel ICH5 southbridge fakeraid
[21:37] <Kingsy> RoyK: I enabled RAID in the bios and configured it in a raid controller menu which is accessed later on in the boot process
[21:38] <patdk-lap> exact same as my workstation: product: 82801 SATA RAID Controller
[21:39] <RoyK> Kingsy: if you boot up on an ubuntu cd, what does /proc/partitions tell you?
[21:39] <RoyK> if /proc/partitions show both drives, it's fakeraid
[21:40] <Kingsy> RoyK: hmm I don't know I cant really do that right now, is there another way of finding out?
[21:40] <RoyK> if it shows a single drive (sda etc), it's ok
[21:40] <patdk-lap> it's fake raid
[21:40] <RoyK> Kingsy: pastebin /proc/partition
[21:40] <RoyK> Kingsy: pastebin /proc/partitions
[21:40] <RoyK> patdk-lap: ok
[21:40] <Kingsy> RoyK: command not found
[21:40]  * RoyK chooses to trust patdk-lap 
[21:41] <Kingsy> np
[21:41] <RoyK> Kingsy: cat /proc/partitions
[21:41] <patdk-lap> it's the intel ICH thing
[21:42] <patdk-lap> just go into the bios, disable the raid options and just set it to achi mode
[21:42] <patdk-lap> install ubuntu
[21:42] <patdk-lap> then when you go to setup the partitions
[21:42] <RoyK> use sw raid
[21:42] <Kingsy> RoyK: yeah it appears as twi drives --> http://paste.ubuntu.com/669570/
[21:42] <patdk-lap> you just set those 1gig drives to use for raid
[21:43] <RoyK> Kingsy: bingo
[21:43] <RoyK> hw raid is nice, if it's true hw raid
[21:43] <patdk-lap> and kicks ass if it has the battery backup to turn on writeback mode
[21:43]  * RoyK hands patdk-lap a beer
[21:44] <RoyK> patdk-lap: or if you use zfs with ssd-caching :P
[21:44]  * RoyK  rather likes zfs
[21:44] <patdk-lap> I've been thinking about trying flashcache
[21:45] <RoyK> with what system?
[21:45] <patdk-lap> that is my issue
[21:45] <patdk-lap> I have no point to try it at the moment :)
[21:45] <RoyK> I have a system with 7TB net storage - 14 1TB drives - striped mirrors with some SSDs
[21:46] <RoyK> performance is better than anything I've seen so far
[21:46] <patdk-lap> I would like to try it for a esx san backend
[21:46] <RoyK> patdk-lap: that system cost about NOK 43k
[21:46] <patdk-lap> but really need like 10g nics or something to make use of it, to stress it
[21:46] <RoyK> so not a lot :P
[21:47] <RoyK> 14 drives in striped mirrors, two spares, two for the root, the remaining slots for SSDs
[21:47] <w00> Link :)
[21:47] <RoyK> s/14/16/
[21:48] <RoyK> w00: just a normal 24-bay supermicro machine with a bunch of drives
[21:49] <dannf> hallyn: can #697690 be marked fixed-released now?
[21:50] <patdk-lap> yay, I might get ipv6 this month
[21:50] <dannf> er bug 697690
[21:51] <Daviey> adam_g: around?
[21:55] <hallyn> bug #697690
[21:55] <hallyn> dannf: yeah, it's in oneiric right?  it's even in debian
[21:55] <dannf> hallyn: right
[21:56] <hallyn> does it need to be SRUd to lucid, or is that done?
[21:56] <dannf> hallyn: does adding a dbg package make sense for an SRU?
[21:56] <hallyn> marked it fixed
[21:56] <dannf> hallyn: thx
[21:56] <adam_g> Daviey: here
[21:56] <hallyn> dannf: i guess not, bc it's a new pkg
[21:57] <hallyn> thx, ttyl
[21:59] <Daviey> adam_g: ah! i just opened bug 829055
[22:01] <adam_g> Daviey: ah, yea. i just ran that pkg by chuck before sending it to PPA. was gonna let it build while im in this talk
[22:01] <Daviey> ** Public Service Announcement **  - Are there any bugs that people know of that should be tracked on http://status.qa.ubuntu.com/reports/ubuntu-server/release-bugs.html that are nt
[22:01] <Daviey> not. Thanks :)
[22:01] <Daviey> adam_g: Ah cool, you are working on it atm.
[22:01] <patdk-lap> I just submitted one for ubuntu desktop :)
[22:01] <Daviey> adam_g: You need to jump through the FFe hoops, sadly.
[22:02] <adam_g> Daviey: figured as much.
[22:03] <Daviey> hallyn: did you see https://bugs.launchpad.net/ubuntu/+source/qemu-kvm/+bug/827831/comments/3 ?
[22:11] <Daviey> adam_g: If you are able to test that update on nova, that would rock my world.
[22:18] <utlemming> maybe we need a new build host for making arm cloud image...build time is about an hour on our current build host
[22:19] <utlemming> Daviey: do you want to chat tomorrow around 3ish your time?
[22:21] <bernhard2> Cant seem to enable TLS in EXIM4 file 03_exim4-config_tlsoptions .ifdef MAIN_TLS_ENABLE howto set this ? MAIN_TLS_ENABLE = yes
[22:21] <bernhard2> MAIN_TLS_ENABLE = 1 MAIN_TLS_ENABLE = true In this order ? .ifdef MAIN_TLS_ENABLE MAIN_TLS_ENABLE = 1
[22:23] <Daviey> utlemming: sounds good to me
[22:24] <Daviey> utlemming: can you fire me across across a google invite?
[22:24] <utlemming> Daviey: yup, can do
[22:24] <Daviey> super!
[22:24] <utlemming> Daviey: assuming this build goes through, arm images are now automagic.
[22:25] <hallyn> Daviey: yes i did - scroll up, i asked you what you meant by that
[22:25] <hallyn> Daviey: it compiles and runs just fine on standard oneiric
[22:28] <Daviey> hallyn: ah, sorry - so 0.15 RC is currently in Debian experimental with that added build dep.
[22:28] <Daviey> I wondered if that is a requirement / want.
[22:37] <hallyn> Daviey: interesting.  but that's different from the debian unstable version right?
[22:37] <hallyn> oh
[22:37] <hallyn> i see
[22:38] <hallyn> i've got libvirt on the brain :)
[22:39] <hallyn> Daviey: I guess we're just supporting without librbd.
[22:40] <hallyn> seems like something worth pursuing MIR for for 12.04
[22:40] <hallyn> Alright, in theory i'm done with libvirt merge.  I'll save testing for later so i can keep this good feeling during dinner and not get all cranky :)
[22:51] <Daviey> hallyn: Have a good one.. Have you pushed the libvirt merge anywhere?
[22:56] <hallyn> Daviey: /win 6
[22:56] <hallyn> bleh
[22:56] <hallyn> Daviey: not yet
[22:57] <hallyn> Daviey: it's whining at me about libnl3-dev, and i don't understand
[22:57]  * Daviey wonders what is more interesting on window 6
[22:57] <Daviey> hallyn: I'm too tired to worry about it :)
[22:58] <hallyn> Daviey: np - good night
[22:59] <Daviey> hallyn: nn
[22:59] <adam_g> Daviey: what tests were failing on amqplib being out of date?
[23:01] <Daviey> adam_g: amqplib >1.0 is required for a newer kombu package.  A newer kombu package is required to throw away carrot from nova (or at least, they way i started working on it)
[23:01] <Daviey> also, the MIR required a newer version
[23:06] <adam_g> Daviey: are there nova packages anywhere that depend on kombu instead of carrot?
[23:08] <Daviey> adam_g: no.. it's work in progress.
[23:08] <Daviey> adam_g: What i mean is, check that a newer version of amqplib doesn't break nova, regardless of kombu vs carrot
[23:10] <Daviey> and on that note, i'm going to bed.  Nighty-night.
[23:11] <adam_g> Daviey: 10-4. night.
[23:11] <utlemming> have a good one Daviey
[23:31] <Kiall> Cam anyone tell me whats wrong with these rules? getting "Can't specify HL option twice" on all the commented rules ... http://pastie.org/private/g0gxhepmeox3woqkx6ouma
[23:31] <Kiall> these iptables rules*
[23:34] <bkerensa> hmm
[23:34] <bkerensa> Kiall: Strange... If they are commented why not just save the commented rules in another file until you need them?
[23:35] <Kiall> Sorry .. I've only commented them out because they dont work :)
[23:35] <bkerensa> oh
[23:35] <Kiall> Basically I have this .. `-m hl --hl-eq 255 -m comment --comment "Embedded Rule comment"`
[23:36] <Kiall> This works: `-m hl --hl-eq 255` ... as does this ... ` -m comment --comment "Embedded Rule comment"`
[23:36] <bkerensa> Kiall: Well they look like duplicates
[23:36] <Kiall> but combined, they fail :/
[23:36] <bkerensa> hmm
[23:37] <Kiall> I've also tested that `-m conntrack ... -m comment ...` works as expected (in case the -m twice was causing issues)
[23:41] <Kiall> hah .. i found the issue .. when --comment is used with hop limit stuff, the comment option must come before the hop limit option -_-
[23:43] <bkerensa> yeah