[00:03] <rushed> rs
[00:59] <justizin> anyone have an auto-install image with sane defaults? i have a headless machine that's dead and needs a reinstall, and no effing keyboard around (personal machine ;d)
[00:59] <justizin> i guess i've been ssh-ing to this thing for like 3 years and i gave away all my keyboards
[00:59] <justizin> tried a trusty upgrade and i think it power failed installing glibc and it just keeps hanging trying to recover from a failed upgrade
[01:02] <MavKen> what is the easiest thing to install/setup for my php scripts to send out mail?  I dont need a full blown mail server
[02:24] <mwhudson> hm
[02:25] <mwhudson> i'm trying to run the latest trusty cloud image under kvm but i can't log in
[02:25] <mwhudson> neither taking the 'ubuntu with random password' or ubuntu:ubuntu options in grub work
[02:25] <mwhudson> is this known?
[06:59] <Amarino_Mallo_da> hi all
[09:32] <jamespage> Daviey, all of the 2013.2.2 updates are now in the saucy-proposed queue
[10:16] <Daviey> jamespage: ok, will process within the hour.
[10:16] <jamespage> Daviey, thanks
[12:11] <martin__> Hi, I'm trying to install php5-xsl in Ubuntu 12.04 but keep getting this error "E: Unable to locate package php5-xsl‎"
[13:35] <smoser> jamespage, is package list at http://status.qa.ubuntu.com/reports/ubuntu-server/cloud-archive/cloud-tools-next_versions.html still relevant ?
[13:35] <smoser> particularly i'm asking about mongodb and libv8-3.14
[13:36] <jamespage> smoser, right now yes
[13:36] <jamespage> I'm still waiting for upstream to switch over the juju-mongodb
[13:37] <smoser> ok.
[13:55] <smoser> rbasak, hooray, uvtool in cloud-tools-next: https://launchpad.net/~ubuntu-cloud-archive/+archive/cloud-tools-next/+build/5640087
[13:59] <rbasak> smoser: \o/
[13:59] <rbasak> smoser: thank you!
[14:24] <Daviey> hallyn: libvirt SRU candidate 1.1-0ubuntu8.6 seems to include many many more fixes than bug 1228977.
[14:25] <hallyn> Daviey: this is the saucy libvirtd?
[14:26] <Daviey> yah
[14:27] <Daviey> hallyn: i'd be happier if libvirt had an MRE, tracking the maint tree
[14:27] <hallyn> Daviey: yeah, if saucy was longer-lived I'd do that
[14:28] <hallyn> if 1.2.2 gets a -maint tree, we should request an MRE for trusty
[14:28] <hallyn> the problem is the -maint trees are rather randomly started after-the-fact.  Of course perhaps zul and I should always maintain one if nooen else does
[14:29] <hallyn> (I did start one for the precise tree;  noone uses it but me of course :)
[14:29] <zul> hallyn:  totally
[14:30] <hallyn> Daviey: anyway, I didn't feel comfortable that cherrypicking only the patches in comment #32 would suffice
[14:30] <hallyn> Daviey: but if you want to reject the upload I understand.
[14:30] <Daviey> hallyn: ah, i assumed there was a maint tree for all.
[14:30] <hallyn> nope :(
[14:31] <hallyn> I also need to look at bug 1264465 for saucy libvirt, once that SRU is either completed or rejected
[14:31] <zul> Daviey:  that would be too easy
[14:33] <Daviey> hallyn: no planned point release for 1.2.2?
[14:33] <Daviey> or point, point
[14:36] <hallyn> Daviey: that doesn't get announced in advance I don't think
[14:38] <hallyn> zul: qemu 2.0 candidate (misnamed) built fine in ppa:ubuntu-virt/canidate;  woohoo
[14:38] <zul> hallyn:  sweet..send out the info
[14:39] <hallyn> coffee
[14:39] <hallyn> Daviey caught me caffeine-less
[15:47] <caribou> jamespage: I Need your opinion on bug 1188126
[15:47] <caribou> https://bugs.launchpad.net/juju/+bug/1188126
[15:47] <caribou> Juju unable to interact consistently with an openstack deployment where tenant has multiple networks configured
[16:02] <eutheria> hi, is there a package i can install that will email notifications for package updates?
[16:03] <genii> eutheria: Instead, probably just to sign up for the mailing list of changes for the distribution you're running
[16:04] <eutheria> genii: i just want to know to install packages, i don't want to manually check a list
[16:04] <thedeeno> I'd like to use docker on some of my servers. Docker says I need at least version 3.8 of the kernel to use it.
[16:04] <thedeeno> What should I do with my 12.04 machines? Upgrade the kernel
[16:04] <thedeeno> or upgrade the whole system
[16:04] <thedeeno> to like 13.04 or something
[16:05] <thedeeno> ? I'm unsure of the pros and cons of each
[16:11] <eutheria> i wonder what docker is
[16:16] <caribou> thedeeno: you can install backported lts kernels on 12.04
[16:17] <caribou> thedeeno: either you use apt-get dist-upgrade but that will touch more than just the kernel, or look for linux-image-3.*lts* in the package list & install only the one you want
[16:18] <thedeeno> caribou: I'm new to this ecosystem. What are those? Are these the 12.04.X releases? They include new kernels?
[16:18] <caribou> thedeeno: since 12.04 is an Long Term Support (lts) release, newer kernels have been backported to 12.04
[16:18] <thedeeno> nice, that is my primary concern, keeping that long term release support
[16:20] <caribou> thedeeno: you can use "sudo apt-cache search linux-image | grep lts" to find them out
[16:20] <thedeeno> caribou: what do these packages look like? Is this one? `linux-generic-lts-raring`
[16:20] <thedeeno> caribou: great! I'll try that out
[16:21] <caribou> thedeeno: apt-cache show linux-generic-lts-raring | grep Version will tell you
[16:21] <caribou> thedeeno: for instance, the saucy-lts one is kernel 3.11-*
[16:22] <thedeeno> caribou: so these kernels are all considered safe then, because they're backported? I was nervous that I'd hit some serious bugs but just
[16:22] <thedeeno> replacing the kernel from underneath 12.04's feet
[16:22] <thedeeno> s/but/by/
[16:22] <caribou> thedeeno: safe & suported
[16:22] <patdk-wk> the only issue you should have, is if you need dkms modules
[16:22] <patdk-wk> they won't work
[16:23] <thedeeno> wow, excellent
[16:23] <caribou> thedeeno: patdk-wk don't the dkms mods get rebuilt ? hmm, maybe not on precise
[16:23] <patdk-wk> yes
[16:24] <patdk-wk> but no one backported the dkms module to support the new kernel
[16:24] <patdk-wk> so if it works, good, if it doesn't due to kernel changes, crash and burn
[16:24] <patdk-wk> xtables/open-vm-tools won't work
[16:24] <caribou> hmm, true
[16:24] <caribou> so thedeeno ^^^
[16:26] <patdk-wk> now, I have those two dkms modules built by me to work, on all except the newest one
[16:27] <patdk-wk> I opted to drop support and wait for 14.04 :)
[16:27] <patdk-wk> and I descided it was just too much of a pain, and only have one machine using them
[16:37] <thedeeno> yeah it sucks that I'm 2 months early, but i need to deploy these servers this week :(
[16:37] <thedeeno> i don't think I have any dependencies on dkms tho
[16:38]  * thedeeno crosses fingers
[16:38] <thedeeno> caribou: patdk-wk: ty ty
[16:38] <thedeeno> OH, one more thing, how does installing a new kernel effect my current packages.
[16:38] <thedeeno> ?
[16:39] <patdk-wk> it doesn't
[16:39] <thedeeno> sweet
[16:54] <thedeeno> Is there a way to get the actual kernel version from these packages without installing them? I'd like to do some automation and need
[16:54] <thedeeno> to work with the version number before reboot
[17:02] <patdk-wk> grub?
[17:08] <thedeeno> I'm installing this kernel on a rackspace node, and I need to edit the grub entry before reboot because rackspace apparently doesn't
[17:08] <thedeeno> like what grub-update is giving it
[17:09] <thedeeno> and I'd like to do it in an automated way using sed
[17:09] <thedeeno> but I'm not sure where to find the actual kernel version (formatted like 3.8.0-19-generic)
[17:10] <thedeeno> basically i'd like to, given one of these lts packages, figure out what the `uname -r` would report after installing it
[17:15] <patdk-wk> the only way is by parsing the grub file
[17:15] <patdk-wk> but if grub isn't there
[17:15] <patdk-wk> guess you just have to search /boot/
[17:16] <patdk-wk> ls /boot/vmlinuz-* | sort -r
[17:41] <hallyn> oh  noes - qemu upstream needs a seabios update
[17:42] <hallyn> ah no, only need a new symlink
[19:13] <shodan45> I'm running openvpn client on 12.04 & having problems when it loses connection to the VPN server - it tries to reconnect, but can't because it can't resolve the host name
[19:14] <shodan45> and it can't resolve because the routing is still set to use the VPN....
[19:14] <shodan45> anyone know a fix for that, aside from manually restarting things?
[19:15] <sarnold> I'm surprised it falls back to looking up an ip address when a connection drops rather than just try to re-handshake with the IP it already knew..
[19:19] <shodan45> sarnold: even if it did, because the routing table is still set to use the (now broken) VPN connection, it wouldn't be able to connect
[19:19] <shodan45> I have to manually stop & start openvpn to get it to work again
[19:20] <sarnold> shodan45: really? shouldn't the 'external' ip of the headend be routable over the internet? otherwise how do the vpn packets make it to the gateway?
[19:22] <shodan45> sarnold: I don'
[19:22] <shodan45> oops :)
[19:23] <shodan45> sarnold: I don't follow 100%.... the client is behind a NAT on my home cable internet
[19:24] <shodan45> it connects to the openvpn server just fine, but if my cable connection goes down, when it comes back up, I have to manually restart openvpn
[19:25] <sarnold> shodan45: hrm, this would be easier with a whiteboard or piece of paper, heh
[19:26] <sarnold> shodan45: so, when you initiate a ocnnection to your openvpn gateway, the first packet there has to travel on the open internet.. eventually a connection is established and new routing entries are installed ..
[19:26] <shodan45> for whatever reason, openvpn doesn't reset the changes it made to routing (etc.) when the connection to the remote VPN server drops
[19:27] <sarnold> shodan45: but all the data you send is still being sent to the globally routable ip address of the gateway, right? it can't migrate to an 'internal' ip, inside the tunnel, because the tunnel still needs to get packets to the remote host.
[19:27] <sarnold> that sounds annoying :) but perhaps they expected to be able to rebuild the connection transparently..
[19:28] <shodan45> I have 2 "gateways": my home router, and my VPN service's
[19:30] <sarnold> actually, I wonder if the problem is your NAT router..
[19:30] <shodan45> the only problem I have is when my home cable internet connection goes down (maybe once/week)
[19:30] <sarnold> if it is stateful (most are) it might tear down the connection when your connection drops; openvpv might assume it can just keep trying and it'll eventually work..
[19:30] <shodan45> sarnold: no... the problem is almost certainly with openvpn or its configuration
[19:31] <sarnold> ouch, weekly crashes? how is it so bad? o_O
[19:32] <shodan45> ok, let me try changing the scenario :)
[19:32] <shodan45> suppose I'm on a laptop with wifi & connect to a VPN
[19:33] <shodan45> then I go out of range of the wifi, and connect to a different network
[19:34] <shodan45> the problem is that openvpn needs to put all the changes (notably routing) back the way they were if it loses its connection to the VPN server
[19:45] <sarnold> shodan45: it may not have a good view of when it loses connectivity -- it may not know 'it's gone forever' from 'it's down for ten seconds'...
[19:46] <sarnold> shodan45: there may be a configuration knob to tweak to get it to tear down vpns that don't work, but they might have chosen to remain broken to make it clear that e.g. your connection to google or gmail would go over unsafe local networking
[19:46] <sarnold> shodan45: I suspect "require local teardown" is the more obvious and informative design decision, but when the local networking dies every week, that -would- get very old..
[19:47] <shodan45> sarnold: right... I'm investigating along those lines now
[19:48] <shodan45> sarnold: and it isn't "local networking", just the internet connection - and I suspect that's the real problem
[19:49] <sarnold> shodan45: to my mind that's local :)
[19:50] <shodan45> I think if I yanked the cat6 out of the box, I wouldn't have this problem :)
[19:50] <shodan45> (assuming I plugged it back in...)
[19:51] <sarnold> shodan45: probably, depends upon how quickly you replaced it :) hehe
[19:51] <sarnold> shodan45: off to lunch, have fun, and if you find something I'd be curious to know what your solution is
[19:52] <shodan45> #openvpn suggested a shell script to watch the connection.... :/
[19:54] <pmatulis_> jamespage: do we know what's going on here? --> http://tracker.ceph.com/issues/6726
[19:54] <jamespage> pmatulis_, refers to the upstream ceph repositories
[19:55] <pmatulis_> jamespage: i realize that.  i thought you might know why upstream is not making recent ceph releases available to recent ubuntu releases
[21:39] <ekosuhartonocom> Hay mwhudson .... welcome to this channel, more detail visit www.ekosuhartono.com
[21:39] <ekosuhartonocom> Hay axisys .... welcome to this channel, more detail visit www.ekosuhartono.com
[21:40] <ekosuhartonocom> Hay AndroidLoverInSF .... welcome to this channel, more detail visit www.ekosuhartono.com
[21:40] <ekosuhartonocom> Hay bogeyd6 .... welcome to this channel, more detail visit www.ekosuhartono.com
[21:41] <ekosuhartonocom> Hay jak2013 .... welcome to this channel, more detail visit www.ekosuhartono.com
[21:41] <ekosuhartonocom> Hay mgw .... welcome to this channel, more detail visit www.ekosuhartono.com
[21:41] <shauno> !ops
[21:41] <shauno> (sorry, but that's annoying)
[21:41] <Pici> indeed it is.
[21:41] <IdleOne> done
[21:41] <shauno> thanks
[21:42] <soren> shauno: Oh, thank *you*.
[21:42] <shauno> aw I feel special now :)  I don't like waking everyone up, but .. yeah
[21:43] <mgw> any suggestions on how to recover from a bad upstart script leaving the service in this state: start/killed, process 28539
[21:43] <mgw> I've manually killed the procs
[21:44] <mgw> but can't start the service (it hangs)
[21:47] <sheptard> lol what a twit