[00:34] <jrwren> anyone know how to use libapache2-mod_uwsgi ?  I enable the module, use apache diretive uWSGISocket but apache does not send requests to hte uwsgi process
[00:49] <Jeeves_Moss> how do I add secondary IP addresses to an interface?
[00:52] <sarnold> Jeeves_Moss: for temporary use, 'ip addr add <address> dev <device>', add new routes if necessary
[00:52] <Jeeves_Moss> sarnold, basically, I've setup Nagios, and I need to monitor a bunch of subnets internally
[00:52] <Jeeves_Moss> or is it just easier to add more NICs to the VM?
[01:07] <jrwren> what does one have to do with the other? Nagios doesn't need secondary IP addresses to work.
[01:10] <sarnold> jrwren: I assumed it was a matter of making it easy to communicate with different subnets..
[01:31] <zul> jamespage:  when you get in in the morning https://code.launchpad.net/~zulcss/nova/nova-novnc-regression/+merge/189201
[01:32] <pmatulis_> hallyn: do you know the state of lxc/dnsmasq on saucy?  i haven't used lxc in a long time and i also happened to upgrade to saucy this morning.  dnsmasq isn't giving out leases it seems
[01:41] <zul> hallyn:  policykit less libvirt perkulation
[01:41] <zul> https://launchpad.net/~zulcss/+archive/libvirt-testing
[01:42] <pmatulis_> hm, looks like the new maas package is having package dependency problems.  who can help?
[01:43] <pmatulis_> it appears to depend on 'python-curtin' which does not exist (until Saucy)
[01:53] <hallyn> pmatulis_: the containers need to be updated
[01:53] <hallyn> pmatulis_: do an lxc-create -t ubuntu -n u1 -- -F to flush the cache and build a new one
[01:53] <hallyn> zul: remind me tomorrow, i'll set up a test machine
[01:53] <pmatulis_> hallyn: hm, ok
[01:54] <zul> hallyn:  ack
[01:54] <hallyn> stgraber: i'm going to push the lxc-fedora update to git head unless you should in the next few mins
[01:57] <stgraber> hallyn: I think it's fine, I never use that template and if I did, it'd work better than the current one anyway
[01:58] <hallyn> hah, you THINK :)
[01:59] <hallyn> how much time do you really think you'd be willing to put into making systemd work in a container :)
[02:00] <pmatulis_> hallyn: dunno, still no address given out.  prolly a dnsmasq thing
[02:03] <pmatulis_> hallyn: i do see DHCP offers in the logs
[02:03] <hallyn> pmatulis_: you created a new container after flushing the cache, and it still doesn't work?
[02:04] <stgraber> hallyn: can I give a negative answer?
[02:04] <stgraber> because I feel I've already spent too much time on it ;)
[02:05] <hallyn> stgraber: :)
[02:05] <hallyn> stgraber: ^ what is the dnsmasq in containers issue again?
[02:05] <stgraber> hallyn: btw, do not ack the autodev stuff, we have /etc/systemd/system on Ubuntu and i don't want autodev by default, so he'll have to find a better way of detecting systemd
[02:05] <stgraber> hallyn: can't think of a dnsmasq issue, what are you reffering to?
[02:05] <hallyn> stgraber: ok.  i wasn't going to read that thread right now anyway, it's convoluted and uninteresting.
[02:06] <hallyn> stgraber: ^ pmatulis_'s containers aren't getting addresses.
[02:06] <hallyn> i thought that was a known problem with out-fo-date containers in most cases
[02:07] <hallyn> but i can't recall the details
[02:07] <stgraber> ah yeah, but that was isc-dhcp
[02:07] <stgraber> basically the 3.10 kernel changed a bit and started doing udp offloading on veth
[02:07] <stgraber> so the udp checksums are no longer filled in by the kernel
[02:07] <stgraber> which upset isc-dhcp
[02:08] <stgraber> to fix that, just make sure your container isn't out of date. I SRUed the fix to all supported releases
[02:09]  * hallyn scratches his head on how any sort of offloading on veth is supposed to work
[02:09] <hallyn> pmatulis_: ^ if that doesn't fix it, then i guess please file a bug so we can collect details
[02:17] <pmatulis_> hallyn: i'm at a container console.  dhclient causes OFFERS on the host but no address gets assigned
[02:18] <pmatulis_> hallyn: will do the bug thing tomorrow, it's late here
[02:20] <hallyn> pmatulis_: thanks - good night
[02:21] <stgraber> pmatulis_: yeah, that sounds like the thing I fixed two months ago, update your container and try again
[02:22] <stgraber> pmatulis_: chroot /var/lib/lxc/<container>/rootfs/ apt-get update && chroot /var/lib/lxc/<container>/rootfs/ apt-get dist-upgrade
[02:22] <stgraber> pmatulis_: if you see isc-dhcp-* in there, then that's probably it
[02:28] <pmatulis_> stgraber: the 'apt-get update' does not work.  DNS is busted.  http://paste.ubuntu.com/6190625/
[02:28] <sarnold> temporary copy over your /etc/resolv.conf from host to the chroot?
[02:29] <stgraber> pmatulis_: what he said ^ (cat /etc/resolv.conf > /var/lib/lxc/lxc_precise/rootfs/run/resolvconf/resolv.conf)
[02:30] <pmatulis_> ok, working
[02:31] <pmatulis_> stgraber: yeah, isc-dhcp-client isc-dhcp-common need upgrading
[02:32] <stgraber> ok, so hopefully that was the bug and the container will work after that
[02:32] <stgraber> I SRUed the fix for that quite a while back, I'm surprised people are still getting that issue. I guess not everyone has a package update OCD as I seem to have ;)
[02:33] <pmatulis_> stgraber: bingo
[02:33] <sarnold> stgraber: or once the kernel on the host is updated, the guests are a PITA to update :/
[02:33] <pmatulis_> stgraber: dunno, how could i *not* have it?  i don't quite understand
[02:34] <stgraber> sarnold: yep, exactly
[02:34] <pmatulis_> stgraber: this is a new guest.  me confused
[02:35] <pmatulis_> stgraber: how will the next guest be any different?
[02:35] <stgraber> pmatulis_: because you have a cache in /var/cache/lxc/ that you should flush or update from time to time
[02:35] <stgraber> the first container debootstraps the system in /var/cache/lxc, any subsequent lxc-create will just copy from the cache
[02:36] <pmatulis_> stgraber: hallyn had me flush that already
[02:36] <stgraber> so if you have the old isc-dhcp-* in there, all your new containers will be broken too
[02:36] <stgraber> hmm, odd then because the template is supposed to apply updates from -updates so the debootstrap was done in the last 2 months, you shouldn't have had the bug...
[02:37] <pmatulis_> i did this: 'lxc-create -t ubuntu -n u1 -- -F'
[02:37] <stgraber> that should have worked...
[02:37] <pmatulis_> another bug?  :)
[02:38] <pmatulis_> lemme try another guest...
[02:38] <stgraber> pmatulis_: ah, -F will only flush the cache for the version and architecture of your host, so as you are apparently on saucy, that wouldn't have flushed the precise cache
[02:39] <stgraber> lxc-create -t ubuntu -n u1 -- -r precise -a amd64 -F would do that
[02:39] <pmatulis_> stgraber: oh
[02:40] <pmatulis_> stgraber: retrying...
[02:41] <pmatulis_> stgraber: isn't there a command for all releases?
[02:41] <stgraber> pmatulis_: rm -Rf /var/cache/lxc/* would do that :)
[02:42] <stgraber> (I never use -F, I just wipe the cache manually)
[02:43] <pmatulis_> fair
[02:48] <pmatulis_> stgraber: ok, new precise guest is good!
[02:48] <pmatulis_> stgraber, sarnold: thanks
[02:48] <stgraber> cool, glad to hear we don't have another annoying bug to fix by release ;)
[02:52] <pmatulis_> stgraber: why when i try the same flush command for quantal does it say 'container already exists'?
[02:52] <pmatulis_> stgraber: it exists, but i want to flush/update it
[02:53] <pmatulis_> stgraber: gah, disregard
[05:11] <igalic> Hello happy people
[05:14] <igalic> We've had a couple of crashers recently, so I'm trying to implement https://help.ubuntu.com/12.04/serverguide/kernel-crash-dump.html
[05:15] <igalic> I've already tested this on my laptop, and now I'm testing it on a virtual machine, but my patience is running short. *How* long does it take until the dump/reboot happens?
[07:20] <rbasak> igalic: if you don't get an answer here, try #ubuntu-kernel.
[07:23] <igalic> rbasak: I was just looking at the https://wiki.ubuntu.com/IRC/ChannelList to find a channel more fitting.
[09:07] <gartral> ok all, i'm gearing up to dd one whole disk to a new disk, from /drv/sda to /dev/sdb, what caveats should I be awear of? I'm looking for any minute detail that would keep this from running smoothly, including any grub-related issues such as boot-drive uuid
[09:08] <gartral> /dev/sda to /dev/sdb*
[09:12] <TJ-> gartral: Use a large blocksize to make it fast as possible
[09:12] <gartral> TJ-: I plan on it
[09:12] <TJ-> I typically use BS=100M or more
[09:12] <gartral> TJ-: I'm just worried about other issues
[09:13] <TJ-> As long as the destination system won't have both drives in at boot... unless after the clone you change all the file-system UUIDs
[09:14] <gartral> TJ-: no, i'll be taking both drives and copying using another machne
[09:14] <TJ-> If drives are a different size, and the cloned drive is larger, you'll probably want to add or expand a partition to include the free space
[09:15] <gartral> TJ-: i'm dealing with a dying hdd in my server, both the dying drive and new drive are 320GB, I'm just hopin gto have a clean, if slow, swap..
[09:15] <gartral> hopeing*
[09:15] <TJ-> I'd recommend you not use 'dd' then, use 'ddrescue'
[09:16] <gartral> TJ-: there's no re-read errors or spin up errors on the dying drive, yet. So I'mma try straight dd first
[09:16] <TJ-> OK
[09:16] <rbasak> smoser: I no longer seem to get a "ci-info: eth0" line in my console output in Saucy. Is this expected?
[09:17] <gartral> TJ-: I'm not expecting this too take 15-30 minutes, I'm hopeing it doesn't take more than 6 hours
[09:17] <TJ-> Make sure you're on UPS!
[09:17] <gartral> TJ-: I wish I had one
[09:55] <gartral> where does mysql store it's databases? or is it safest to juse export them from the mysql client?
[09:57] <mardraum> you should use mysqldump if you are moving them around, the mysql data dir by default in ubuntu is in /var/lib/mysql
[09:58] <gartral> ok, I'll just export all databases from within phpmyadmin
[10:29] <jamespage> zul, reviewed, pushed and uploaded - thanks for fixing that up!
[11:07] <rbasak> utlemming: http://askubuntu.com/questions/353370/installing-ubuntu-13-04-64-bit-virtual-machine-with-vagrant-on-os-x-provides-o
[11:07] <rbasak> utlemming: looks like we/something is not expanding the partition to the size of the disk on boot.
[11:08] <rbasak> Maybe because cloud-init is not being used?
[11:34] <sander^work> How come du -csx /* gives 1GB used.. but df displays 10GB used?
[11:40] <rbasak> sander^work: did you try googling? Eg. http://serverfault.com/q/57098 and http://linuxshellaccount.blogspot.co.uk/2008/12/why-du-and-df-display-different-values.html
[12:03] <sander^work> rbasak, looks like Reserved block count is: 235916 and Block size is: 4096.. is this the problem?
[12:05] <sander^work> No. tune2fs -m 5 /dev/xvda1 gives me: Setting reserved blocks percentage to 5% (235916 blocks)
[12:06] <sander^work> df -i tells me only 5% is in use. but df tells me 96%
[12:09] <sander^work> Hum. lsof +L 1 / | grep -i deleted ..gives me two files
[12:11] <sander^work> How can I force delete those files I found with that command?
[12:15] <sander^work> Cool. Some app had it open. Fixed it :)
[12:19] <smoser> rbasak, i'll loook. you should. i'll lok.
[12:19] <smoser> rbasak, please file a bug
[12:38] <rbasak> smoser: filed bug 1235231
[12:41] <zul> jamespage:  https://code.launchpad.net/~zulcss/cinder/2013.3.rc1/+merge/189299
[12:49] <jamespage> zul, + * New upstream point release. -> New upstream release candidate
[12:49] <jamespage> zul, finish_volume_migration now tests OK right?
[12:50] <zul> jamespage:  yeah removed the skip and ran fine
[12:50] <zul> noticed the bug i opened was closed ;)
[12:50] <jamespage> other than the changelog comment +1 then
[12:51] <zul> jamespage:  fixed
[13:14] <jamespage> zul,  OK - dash theme refreshed
[13:14] <jamespage> man I hate doiing css
[13:14] <zul> hehe
[13:15] <zul> jamespage: CA nearly updated waiting for cinder and nova
[13:33] <smoser> rbasak, can you test one of the other saucy 'releases' ?
[13:33] <rbasak> smoser: which one? Do you have a simplestreams filter query for me? :)
[13:35] <smoser> you just need to get more than one
[13:36] <smoser> http://cloud-images.ubuntu.com/releases/saucy/
[13:38] <smoser> rbasak,
[13:38] <smoser> sstream-query http://cloud-images.ubuntu.com/releases/streams/v1/com.ubuntu.cloud:released:download.json release=saucy arch=amd64  "label~(alpha|beta)" ftype=disk1.img
[13:38] <smoser> that works for me
[13:39] <smoser> but uvtools just returns to quickly.
[13:48] <jamespage> zul, https://code.launchpad.net/~james-page/horizon/theme-refresh-perms-fix/+merge/189314
[13:48] <zul> jamespage:  +1
[13:53] <jamespage> zul, ta - pushed and uploaded
[13:53] <jamespage> phew
[13:53] <jamespage> I assume you reviewed every line of my css change?
[13:53] <jamespage> :-)
[14:00] <rbasak> smoser: it's because I'm using max_items: 1. Is that a bug, and if so, where?
[14:00] <rbasak> smoser: I wanted the latest single thing for each product that matched the query.
[14:00] <rbasak> smoser: it seems that I'm getting the latest single thing and then filtering that with the query instead.
[14:01] <smoser> right. that is how that works rbasak
[14:01] <smoser> max applies to the newest item
[14:02] <rbasak> smoser: which bit is right?
[14:02] <smoser> indepnedent of match.
[14:02] <rbasak> I'm lost. Are you telling me that I'm calling the simplestreams API wrong for what I want?
[14:03] <smoser> well, as it is right now you can't use max for that.
[14:03] <smoser> because 'label' is on the item
[14:03] <smoser> and max is counted before it gets down there to find the item
[14:03] <smoser> i think.
[14:03] <smoser> it is the way it is.
[14:03] <smoser> i dont know if its necessarily "right"
[14:04] <smoser> i think when i wrote it i thought i could go both ways on that.
[14:04] <smoser> i dontk now.
[14:04] <smoser> i'm willing to be convinced otherwise.
[14:06] <rbasak> I'd expect the filter to apply first, and then a limit, with the newest products coming out first.
[14:07] <rbasak> It doesn't make sense for a limit to apply before the filters.
[14:15] <rbasak> smoser: ci-info: eth0 disappeared between alpha3 and beta1. Also missing from beta2.
[14:16] <smoser> rbasak, thank you.
[14:17] <rbasak> smoser: alpha3: http://paste.ubuntu.com/6192406/; beta1: http://paste.ubuntu.com/6192407/; beta2: http://paste.ubuntu.com/6192408/
[14:24] <hallyn> zul: starting regression tests against your libvirt candidate
[14:24] <zul> cool thanks
[14:42] <Teduardo> Hi I am running Ubuntu 13.04 and I did apt-get install ubuntu-desktop and I need to start the desktop and a VNC server for the console at boot, can anyone tell me how to do this?
[14:42] <Teduardo> or even how to start the desktop/vnc from SSH
[14:43] <Teduardo> under rhel it's just init 5
[14:47] <smoser> jodh, can i 'start on mounted MOUNTPOINT=/run' and not block ?
[15:07] <hallyn> zul: http://paste.ubuntu.com/6192587/   4 errors... not sure why
[15:09] <zul> hallyn:  so that sucks
[15:13] <hallyn> zul: any ideas offhand?
[15:14] <zul> hallyn:  not really maybe the tests expect policykit?
[15:14] <hallyn> lemme finish this commit then i'll dig in
[15:15] <jamespage> zul, OK if I don't propose a merge for the swift fixes I did this week?  I'd like to push that out with rc1
[15:15] <zul> jamespage:  sure
[15:22] <jamespage> zul, ok pushed and uploaded
[15:22] <zul> jamespage:  yay!
[15:22] <zul> jamespage:  poking at glance with a pointy stick
[15:25] <smoser> rbasak, so this is i think not cloud-init's fault.
[15:25] <smoser> but plymouth or upstart.
[15:25] <smoser> rbasak, do you want to debug some ?
[15:25] <smoser> it'd be useful to see output with that 'my-logger.conf' added in.
[15:34] <rbasak> ack
[15:46] <rbasak> smoser: I'll need to look on Monday I think. It's pretty tedious doing this by hand; I might as well add some functionality to uvtool.
[15:48] <smoser> rbasak, mount-image-callback
[15:48] <smoser> is your friend on that.
[15:48] <smoser> mount-image-callback my.img sh -c 'tee -a $MOUNTPOINT/etc/init/my-logger.conf' < my-logger.conf
[15:48] <smoser> something like that.
[15:49] <rbasak> Yeah but this disk is a qcow2 overlay
[15:49] <smoser> give me some credit.
[15:49] <rbasak> I could use the nbd thing or whatever it is too.
[15:49] <smoser> why would i write a tool that  made you think about such things.
[15:49] <rbasak> It works with that?
[15:49] <smoser> :)
[15:49] <smoser> es.
[15:49] <smoser> yes
[15:49] <rbasak> "mount a file" is a bit vague :)
[15:50] <smoser> only because you are too smart
[15:50] <rbasak> Is it a problem if the debug results I give you isn't from first boot?
[15:50] <smoser> and think "oh, it can't bpossible do smart things"
[15:50] <smoser> i't be better if they were first boot i think.
[15:50] <smoser> but probably not
[15:50] <smoser> you could just patch the original libvirt volume
[15:50] <smoser> make a copy, patch, revert.
[15:51] <rbasak> Does mount-image-callback unmount after it's done? Or is the clue in the name?
[15:51] <smoser> it cleans up.
[15:51] <smoser> so after you do that it is unmounted.
[15:51] <smoser> it used to be called mount-callback-umount
[15:54] <rbasak> Aaargh.
[15:55] <rbasak> I used /mnt for temporary stuff
[15:55] <rbasak> Forgot I can't do that with cloud images since /mnt is used for stuff.
[15:58] <zul> jamespage:  was it neutron you unpatched, ran the testsuite, and then patched it?
[15:59] <jamespage> zul, yes
[15:59] <zul> jamespage:  ok i think i might have to do the same for glance
[15:59] <jamespage> zul, hallyn: ever seen one of these before?
[15:59] <jamespage> thats a 3.11 guest running on a 12.04/3.8 host
[16:00] <jamespage> and http://paste.ubuntu.com/6192811/
[16:00] <jamespage> from the host
[16:00] <jamespage> oh - and its nested KVM
[16:00] <jamespage> so its a 12.04/3.2 guest on a 13.10/3.11 host, which in turn is a guest on a 12.04/3.8 host
[16:01] <jamespage> blimey - thats alot of different versions
[16:01] <zul> why is it probing for an xfs filesystem?
[16:01] <hallyn> jamespage: smb is working on a nested kvm bug...
[16:02] <rbasak> smoser: it doesn't seem to have changed much. http://paste.ubuntu.com/6192819/
[16:02]  * jamespage shrugs
[16:02] <rbasak> smoser: I checked and I see a /etc/init/my-logger.conf
[16:02] <jamespage> zul,  - i have not idea - I suspect that something is trying to determine fs type maybe?
[16:02] <zul> right
[16:02] <zul> i havent seen that beofre
[16:02] <hallyn> jamespage: though really in this case it looks like you have very unfortunate garbage on nbd7
[16:02] <hallyn> it doesn't find ext3, and then xfs chokes on it
[16:04] <jamespage> hallyn, I think that is a problem but I don;'t think its the reason why the instance is in paused state
[16:04] <jamespage> hallyn, it actually started to boot
[16:05] <hallyn> jamespage: I think the guest is waiting for you to fix the fs...
[16:05] <jamespage> http://paste.ubuntu.com/6192827/
[16:05] <hallyn> jamespage: is http://paste.ubuntu.com/6192827 for someone else?
[16:06] <hallyn> (cause i don't see how it relates to the other paste)
[16:06] <hallyn> is the 5 hour jump in syslog from a tz switch during boot?
[16:06] <jamespage> hallyn, thats the console log from the highest level guest
[16:07] <hallyn> then what's the other one?
[16:07] <jamespage> is the kernel log from the host
[16:07] <jamespage> it problem did nothing for a few hours
[16:07] <hallyn> oh the HOST is choking on the xfs?
[16:08] <jamespage> yeah
[16:08]  * hallyn confused about why the host is trying to mount it then
[16:08] <jamespage> agreed
[16:08] <hallyn> seems like automount-gone-wild
[16:08] <jamespage> hmm
[16:08] <jamespage> its doing something with ndb
[16:08] <hallyn> udev sees new nbd device and says "aw shucks, lets mount it" ?
[16:11] <jamespage> maybe
[16:12] <jamespage> does libvirt do anything like that?
[16:12] <hallyn> not that I know of, no
[16:13] <hallyn> but maybe nova is pre-mounting it for the guest?
[16:19] <jamespage> hallyn, maybe - I'd hope not but who knows?
[16:19]  * jamespage digs further
[16:20] <hallyn> jamespage: is it possible to stop nova on the host and then run the test?
[16:20] <hallyn> zul: the 4 failures seem to be due to apparmor not allowing libvirt to read files under /tmp/*/
[16:20] <hallyn> what i can't figure out is why this would be new
[16:21] <zul> polkit version?
[16:21] <jamespage> hallyn, hmm
[16:21] <jamespage> zul, that might be something todo with it
[16:21] <hallyn> zul: libpolkit-gobject-1-0:amd64 is all i have
[16:21] <jamespage> I don't see this on the bare metal host
[16:21] <jamespage> which is running older libvirt from havana staging on 12.04
[16:21] <hallyn> zul: (or was that directed at jamespage )
[16:22] <hallyn> well lemme downgrade libvirt and see
[16:22] <zul> smb: https://launchpadlibrarian.net/152526461/buildlog_ubuntu-precise-i386.xen_4.3.0-1ubuntu1~cloud1_FAILEDTOBUILD.txt.gz
[16:27] <rbasak> EOD
[16:27]  * rbasak EOD
[16:27] <jamespage> hallyn, OK _ I see similar kernel log messages on the physical havana hosts; but no kernel stack trace
[16:29] <jamespage> for example - http://paste.ubuntu.com/6192922/
[16:31] <jamespage> hmm - instance key injection I think
[16:31] <hallyn> ?
[16:33] <jamespage> when nova starts an instance it can potentially inject an ssh key to the instance
[16:33] <jamespage> it does that using qemu-nbd
[16:33] <hallyn> ah.  but so that nbd7 is meant to be a valid fs?  I assumed it was not-yet-created fs attached as a secondary hd
[16:36] <jamespage> hallyn, I'm not sure but nova does not log the failure message so I'm assuming it succeeds
[16:36] <jamespage> hallyn, the http://paste.ubuntu.com/6192811/ paste was from libvirt/qemu
[16:36] <hallyn> jamespage: do you know what fs is actuall *on* nbd7?
[16:36] <jamespage> hallyn, well its a stock precise cloud image
[16:37] <hallyn> wouldn't that be ext4?  but it wasn't ext4
[16:52] <jamespage> hallyn, actually there is a qemu-nbd process running =- root     25190     1 99 15:53 ?        00:58:11 /usr/bin/qemu-nbd -c /dev/nbd7 /var/lib/nova/instances/3eac2a3b-52d3-4a9c-ba21-602f1cc5e220/disk
[16:52] <jamespage> so it started OK
[16:52] <jamespage> despite the error message
[16:53] <jamespage> but then the instances pauses
[16:53] <hallyn> that just sets up the block device.  it doesn't mount it
[16:53] <hallyn> you can do that with a completely garbage file
[17:25] <smoser> rbasak, sorry. slow reply. that paste http://paste.ubuntu.com/6192819/ shows that a *ton* of output got eaten
[17:44] <Phog> what are some good projects to help me learn ubuntu a bit better
[17:51] <sarnold> Phog: whatever will keep your interest long enough :) if you're at a complete loss of what to do next, you could read the server guide:https://help.ubuntu.com/12.04/serverguide/   or you could read all the manpages on your system (probably harder today than it once was, maybe focus on commands in /bin and /sbin for starting..)
[17:52] <Phog> hmm ok
[17:53] <Phog> i like projects because i learn as i implement
[17:53] <Phog> but
[17:53] <Phog> i can handle this
[17:53] <Phog> thanks
[17:54] <sarnold> Phog: you might find the perfect project while reading the guide :) it's got a lot of cool stuff...
[17:55] <Phog> you ever tunneled http through an ssh session
[17:55] <sarnold> sure, it's quite useful for traversing firewalls
[17:55] <Phog> yeah
[17:56] <Phog> does chrome do it?
[17:56] <Phog> on windows?
[17:56] <Phog> seems like all the guides i read are firefox
[17:57] <sarnold> Phog: sure, chrome will do it fine, if you can get the ports configured properly for ssh. I _think_ putty has those options, but I haven't looked in a decade or more.
[17:57] <Phog> ok
[17:58] <Phog> ssh is pretty nifty
[18:08] <hallyn> sarnold: yeah i've port fwded through putty for accessing security cam before, it does it fine
[18:08] <hallyn> zul: 1.1.1-0ubuntu8 fails the same way
[18:09] <hallyn> did we use to allow /tmp/** access to libvirt-qemu?
[18:09] <zul> hallyn:  really?
[18:09] <zul> im not sure
[18:09] <sarnold> hallyn: cool :) I figured MS would have made a Good Enough copy of BSD sockets layer to let it work fine..
[18:09]  * hallyn checks precise
[18:10] <hallyn> no...
[18:11] <hallyn> is this a regression in virt-aa-helper?
[18:11] <Phog> wow
[18:11] <Phog> i can't believe i've gone this long not realizing i could tunnel through SSH
[18:12] <Phog> jesus
[18:12] <Phog> how fucking insanely handy is that
[18:12] <hallyn> jobs-level insanely
[18:13] <hallyn> i suppose thats why plan 9 network ns  binding stopped being so cool
[18:15] <sarnold> plan 9 never stopped being cool, even if I never did get the hang of using the mouse so much.
[18:17] <hallyn> need a corded kbd
[18:22] <hallyn> grumble grumble thermal shutdown grumble grumble
[18:29] <tom[]> i foolishly broke a server's packages. i learned on #opensmtpd of a supported .deb. the only pkg available then was a sid but i didn't know what that meant. i added "deb http://http.us.debian.org/debian sid main" to sources.list on ubuntu 12.04. i soon learned how stupid that was. chaos ensued. the mess is beyond apt-get -f (no aptitude on the system). how can i get back to the versions from the precise distro?
[18:36] <hallyn> there's probably a better way, but you could just install a quick precise instance somewhere to get package versions of things you defiantely need, then grab them from archive.ubuntu.com/ubuntu/pool/
[18:37] <hallyn> (then dpkg -i them)
[18:37] <hallyn> or,
[18:37] <hallyn> you could install a rootfs (say a container) onto a usb stick;  mount that on that server;  copy stuff to bootstrap
[18:38] <tom[]> the servers are very distant.
[18:38] <tom[]> but i have two that are almost a matching pair. i screwed up only one of them
[18:38] <hallyn> rsync -va remote:/bin/ /bin/ ?
[18:39] <hallyn> well if you have wget and dpkg then i'd recommend the first option
[18:40] <sarnold> yeah, I like the sounds of the first option
[18:41] <sarnold> another option that I very nearly hesitate to mention... try replacing 'precise' with 'raring' in your sources.lists, and try to apt-get dist-upgrade your way forward. It won't be on an LTS release then, but going further forwards make work better than trying to downgrade packages..
[18:42] <tom[]> i have wget, dpkg. but i don't understand what you wrote. "install a quick precise instance somewhere" e.g. locally in virtualbox?
[18:44] <hallyn> or an amazon instance;  or a container (lxc-create -t ubuntu -n p1 -- -r precise;  lxc-start -n p1)
[18:44] <hallyn> or even just a debootstrap somewhere.  just so you can compare versions and sha1sums
[18:45] <hallyn> sarnold's raring suggestion is good depending on what this box does and your future plans for it
[18:47] <tom[]> its a webapp server (maria, php, nginx, sphinxsearch etc.)
[18:47] <tom[]> currently a hot standby and working but the packages are unmanagable
[19:05] <jrwren> i'm trying to chroot to a tree where most things (libc, bin/bash) are symlinks to another area in the same tree. strace shows chroot syscall succeed but execve /bin/bash fail with ENOENT   This works on an old FC system. Can anyone suggest where I should look?
[19:17] <jrwren> nevermind. of course after I ask, I find soemthing new and fix it.
[19:18] <hallyn> zul: all right, passes on raring.  i dunno...  lemme try a new cloud instance of saucy
[19:18] <zul> hallyn:  ack thanks for doing this
[19:53] <hallyn> zul: yea a new stock saucy instance has the same 4 failures.  not sure when this regression popped up.
[19:53] <hallyn> sigh - bisect i guess
[19:58] <hallyn> testing 1.1.1-0ubuntu4
[21:03] <delinquentme> OK so putting in alias commands ... ~/.bashrc or  ~/.bash_profile ??
[21:03] <jkyle> anyone have a working example of a multi-disk, multi volume group (lvm) preseed?
[21:03] <jkyle> or know if that's not possible.
[21:05] <TJ-> delinquentme: Neither - ~/.bash_aliases
[21:06] <delinquentme> TJ-, what are the other two files used for then?
[21:06] <jkyle> TJ-: shit, learn something every day. didn't realize *_aliases was in the init list
[21:06] <jkyle> for bash like shells
[21:07] <delinquentme> ^^^
[21:07] <delinquentme> ditto I'm characterizing another machine and I'm learning how "things should be done" bit by bt
[21:08] <TJ->  .profile will include .bashrc if the shell is bash, .bashrc will include .bash_aliases and other files
[21:09]  * TJ- is gone -  been a long bug-fix of a day
[21:11] <adam_g> smoser, http://status.qa.ubuntu.com/reports/ubuntu-server/cloud-archive/cloud-tools_versions.html
[21:11] <smoser> dude.
[21:11] <smoser> you rock!
[21:11] <adam_g> smoser, required minimal changes. adding the next -> ubuntu dev will be a little more work.
[21:12] <smoser> ubuntu -> next
[21:14] <ajhlinuxuser> Hi, using virt-manager, creating a VM remotely won't let me select an ISO file.  What is wrong??
[21:22] <GH0> Hello, I need some help installing Ubuntu Server on a UEFI motherboard, I am installing the 12.04.3 LTS on a Gigabyte 990FX UD3 AMD motherboard. Everytime I use the installer disc, I can get past the the partition management utility, but it seems to fail at "Installing the Base System". This fails at "The debootstrap program exited with an error (return value 1). Check /var/log/syslog or see vitual console 4 for the details"
[21:22] <GH0> When I switch over to that screen, I see three lines that are ext4-fs errors'
[21:23] <GH0> For device sda2 which would be the / mount point, as sda1 iis the /boot mount point
[21:24] <GH0> The error listed is: "EXT4-fs error (device sda2) in ext4_reserve_inode_write:4688: Journal has aborted
[21:24] <GH0> The second error listed is: "EXT4-fs error (device sda2) in ext4_evict_inode:243: Journal has aborted
[21:24] <GH0> This has happened twice, and I don't think it is a disc problem, so I was wondering if anyone could help me out with the install.
[21:49] <hallyn> zul: jdstrand: the qa-regression-tests have 4 failures using the raring libvirt package on saucy;  they all pass on raring.
[21:50] <hallyn> each failure is due to:
[21:50] <hallyn> Oct  4 21:39:46 ip-10-10-145-232 kernel: [40612904.124019] type=1400 audit(1380922786.415:533): apparmor="DENIED" operation="open" parent=1 profile="libvirt-7d781722-69b7-8801-fe96-caf37b7a8969" name="/tmp/tmpgvfz4m/device_disk.img" pid=17761 comm="qemu" requested_mask="rw" denied_mask="rw" fsuid=107 ouid=107
[21:51] <hallyn> i'm at a loss.  doesn't appear to be the policy shipped with libvirt that's the problem...
[21:52] <rgouveia> hi ... i have an old 32bit pentium-m without pae. is there a netboot of 12.04 server available ?
[21:55] <rgouveia> nevermind, i thought the server kernel was different from the desktop one
[22:48] <hallyn> yeah the entry   "/tmp/tmphzIH1Q/device_disk.img" rw,
[22:49] <hallyn> is not being added
[22:55] <hallyn> using the same version of libvirt
[23:00] <xibalba> is it just me or are the apt repos slow as snot
[23:01] <xibalba> always get around 80KB/sec
[23:02] <sarnold> xibalba: back when I used them, I'd get around 300KB/sec, it felt glacial.. I switched to http://mirror.anl.gov/ubuntu/ and now I get several megabytes per second now..
[23:03] <xibalba> where do i mod that again?
[23:03] <hallyn> zul: well anyway i see no regressions compared to the current saucy version.  back over to you with that.  but the apparmor problem has me stumped
[23:03] <xibalba> found it
[23:03] <xibalba> sarnold, what's the line you've got in sources.list?
[23:04] <sarnold> xibalba: here's just my deb lines: http://paste.ubuntu.com/6194382/
[23:04] <sarnold> xibalba: you may not need or want them all..
[23:04] <xibalba> thanks
[23:04] <xibalba> do i need to specify precise-*
[23:05] <xibalba> i never know the name of ubuntu releases
[23:05] <xibalba> on 12.04.3 though
[23:06] <sarnold> xibalba: ah right, precise for that.
[23:07] <xibalba> holy crap
[23:07] <xibalba> exceptionally faster
[23:07] <sarnold> yeah :D
[23:07] <xibalba> i remember back in my freebsd boxes you could run like fastest_cvs and it would auto add the fastest mirrors for you
[23:07] <xibalba> ubuntu should do that post install on first apt-get updat
[23:09] <sarnold> I tried to find one of those when I got tired of the slow ubuntu.com speeds but the ones I thought I remembered from debian a decade ago just weren't there. but anl.gov is bloody fast, faster than my connection can go anyway, so it works well enough for me. :)
[23:16] <hallyn> heh, remember, kids, root's cron-fired jobs do not have /sbin/ in $PATH by default
[23:17] <sarnold> heh, if I've learned one thing about cron, it's that it is just best to use full pathnames for everything always.
[23:17] <jkitchen> yes.
[23:18] <jkitchen> or just ... not use cron.
[23:18] <xibalba> not use cron?
[23:18] <xibalba> blasphemy