[00:04] <hallyn_> smoser: your patch intorducing the ubuntu-cloud-prep hook removed code doing rm $rootfs/etc/hostname in the ubuntu-cloudtemplate, and didn't replace that anywhere
[00:04] <hallyn_> smoser: when I re-add that, ubuntu-cloud containers get their correct hostname.  Otherwise, they are always called ubuntu, as sidnei found
[00:26] <kantlivelong> ive got an interesting question maybe someone can help with.. i have an img of a single partition on a disk. is there a way i can make a partition table in a file and move the contents of the file into it? i basically wanna rebuild the disk as a file
[00:26] <SpinningWheels> recent development. my ubuntu servers don't use the configuration in /etc/network/interfaces on bootup. i have to ifdown ifup the interface or restart networking service for the configuration to be used.
[00:27] <sarnold> kantlivelong: why do you need the partition table?
[00:28] <kantlivelong> sarnold: i wanna make it bootable in vmware/virtualbox
[00:28] <kantlivelong> sarnold: so basically take a single partition dump made using dd
[00:28] <kantlivelong> sarnold: then recreate the partition table and put it in
[00:29] <kantlivelong> oh and i have a dump of the mbr too
[00:30] <kantlivelong> my only idea is to make a blank file using dd and treat it as a block device, partition it w/ parted then copy it over.. i just dont know how to have it handled as a block dev
[00:31] <sarnold> kantlivelong: if cfdisk or fdisk won't just work on a file, you can use a loop device to get there (man losetup)
[00:32] <sarnold> kantlivelong: but before putting too much effort into this yourself, look at the qemu-img --help and manpage and see if it can get you there without having to do the dd dirty work yourself
[00:32] <kantlivelong> sarnold: thanks :D didnt know of losetup
[00:32] <sarnold> (or, if you can just re-dd the disk, dd the whole thing rather than just the one partition.. :)
[00:32] <kantlivelong> sarnold: yeah theres alot of other partitions though :P its a multi boot system
[01:31] <smoser> hallyn_, that was by design. that should'nt really be necessary.
[02:31] <hallyn_> smoser: but it seems to be.  if i leave $rootfs/etc/hostname in place, it always is 'ubuntu' after starting the container
[02:31] <hallyn_> if i remove it, it has the correct hostname after lxc-start
[06:24] <MACscr> ok, i cant for the life of me get LXC to work with a normal bridge so that it gets an ip address from my routers dhcp server. Nat works great out of the box, but its not what i want =(
[06:25] <MACscr> seems all the intructions i have tried not only work, but are somehow different than each other.  Shouldnt something like this be pretty basic and not require a bunch of workarounds?
[06:26] <MACscr> also, should i use 12.04 or should i maybe use a newer version that might have better LXC support?
[07:59] <kami> Good morning.
[07:59] <kami> I have problems with NFS on a fresh install of 12.04.02.
[08:00] <kami> In /etc/exports, I have '/export *(rw,async,no_root_squash,no_subtree_check)'
[08:00] <kami> hosts.allow and hosts.deny are both empty
[08:02] <kami> When I try to mount the share, I get http://paste.ubuntu.com/6013157/
[08:06] <kami> or better: http://paste.lisp.org/display/138579
[08:06] <kami> I wonder why it says "qword_eol: fflush failed: errno 2 (No such file or directory)"
[08:07] <kami> There IS such a directory as ls shows.
[08:32] <josephtacos> is the a new way to enable mod_rewrite in apache2.4.6? on ubuntu 13.10
[08:46] <josephtacos> is the a new way to enable mod_rewrite in apache2.4.6? on ubuntu 13.10
[09:03] <Quest> what choices do I have for type 1 virtualization and which one is better for installing  solaris and windows 2008 server r2 with multiple user access by remote computers (for both OS simultaneously)?
[09:29] <jamespage> smb, OK - I now have a functional dkms module based on the 1.11 branch of openvswitch
[09:29] <jamespage> only 12 patches to get to 3.11 support
[09:29] <smb> so simple? :-P
[09:31] <smb> jamespage, Do you want me to cross-check or are you just uploading it?
[09:31] <jamespage> smb, review would be nice
[09:32] <smb> jamespage, sure, just point me at the bzr branch
[09:32] <jamespage> smb, lp:~james-page/ubuntu/saucy/openvswitch/snapshot-1.11
[09:32] <smb> jamespage, k, looking
[10:06] <smb> jamespage, Only patch 1 sounds a bit like it might rely on some kernel implementation without adding compat elements but maybe that is not true. The rest is a lot of churn to look at. So I just try how the dkms module compiles and works on S and P.
[10:06] <jamespage> smb, I think thats a good idea!
[10:07] <jamespage> smb, my cherry picking was based on having to pick the GRE restructuring patch to support 3.11 usage
[10:07] <jamespage> otherwise I'd end up implementing a hacked compat patch for the piece
[10:11] <smb> jamespage, Oh yes, that sounds plausible. From 0012 it looks like even more moved into upstream headers. And in theory the openvswitch source tries to have the compat layer. So it is quite possible that some of the later patches actually does that. Looked like gre files might have been pulled in. Just hard to decide on a patch by patch base
[10:11] <jamespage> smb, it has - infact the upstream kernel module for openvswitch in 3.11 supports GRE
[10:11] <jamespage> oh - I cc'ed you didn't I
[10:11] <jamespage> doh!
[10:12] <smb> jamespage, You may or may not. Have not looked into my inbox for a while
[10:12] <jamespage> smb, yeah - upstream confirmed that was the case
[10:13] <jamespage> support for VXLAN tunnelling should come in 3.12
[10:13] <jamespage> smb, so hopefully we can ditch DKMS for 14.04 for openvswitch
[10:13] <jamespage> that would be nice!
[10:13] <smb> Heh, yep
[10:14] <jamespage> smb, bah - I have a system I've been testing with used ovs on the primary nic
[10:14] <jamespage> rmmod openvswitch was not a great idea over the network was it....
[10:14]  * jamespage faceplants
[10:16] <smb> Thats many systems to you. I also have often to be very careful about what I do where
[10:18] <jamespage> smb, OK _ I'm testing with the 3.2 kernel in precise as well
[10:19] <smb> jamespage, seems to have at least built in 3.2
[10:20] <jamespage> smb, thats good
[10:20] <jamespage> smb, just installing a 3.2 kernel on my precise system - normally run the hwe kernel
[10:22] <smb> jamespage, Oh, hm, need to make sure this is the right kernel...
[10:22] <jamespage> smb, well it built for 3.2
[10:22] <jamespage> thats good
[10:24] <smb> jamespage, Ok, yeah mine was a 3.2, too but one never can be sure
[10:24] <smb> The openflow test ran successful too
[10:25] <smb> jamespage, only the common part of loading unloading looks like it might suffer from upstart script problems
[10:26] <jamespage> smb, oh yes
[10:26] <jamespage> service openvswitch-switch force-reload-kmod
[10:26] <jamespage> that don't work with my latest updates
[10:28] <jamespage> smb, I'll update the tests
[10:31] <jamespage> smb, OK - updated tests pushed
[10:32] <smb> jamespage, Ok, that works on P
[10:36] <jamespage> smb, I've submitted the patch for 3.11 upstream and asked whether the patchset for 1.11 branch might be accepted - I don't think its technically released as yet!
[10:38] <smb> jamespage, Yeah I think we at least would have to wait for comments about the patchset from upstream. At least practical testing seemed good so far if that does not miss something
[10:41] <jamespage> smb, I'm going todo a quick gre tunnel test
[10:44] <smb> jamespage, Ok, I would try to go back finding any sanity in nested kvm
[10:50] <jamespage> smb, thanks for your help
[10:51] <jamespage> smb, GRE seems OK on 3.11 kernel but busted on 3.2
[10:51] <jamespage> which is a bit of a problem as I want to backport this for the cloud archive
[10:55] <smb> jamespage, Was that with only new dkms or userspace replaced too?
[10:55] <jamespage> smb, just dkms
[10:56] <smb> Hm, yeah that may get into trouble when the interfaces are different. :/
[10:58] <jamespage> smb, oh - hold on
[10:58] <jamespage> this might be something else interring with it on my precise machine
[10:58] <jamespage> I'm using vlan package as well
[10:58]  * jamespage tried on a different machine
[11:21] <jamespage> smb, nope - its bust
[11:21]  * jamespage debugs some more
[11:22] <jamespage> probably something in the gre compat layer I suspect
[11:24] <smb> jamespage, In some sense good that its not "just" some other dependency. I would suspect that some user-space facing part changed from their old implementation to those picked from the kernel.
[11:25] <jamespage> smb, I'll see
[11:25] <smb> Those are the cases where one wonders whether not the whole source should be pushed into older releases... Of course that will make any AA beat the foo out of me
[11:25] <jamespage> smb, well for the cloud archive you will get the userspace tools as well
[11:25] <jamespage> smb, rebuilding for precise now to see if that makes a difference
[11:26] <jamespage> smb, I doubled checked the lts-raring dkms backport we are providing for 12.04.3 - that works OK with gre
[11:26] <jamespage> and the 12.04 userspace tools
[11:27] <smb> I would think that the raring version still had their old gre code
[11:31] <jamespage> smb, hmm - interesting
[11:31] <jamespage> needs new userspace tools
[11:31]  * jamespage stops flapping
[11:44] <jamespage> smb, hmm - bug in there somewhere - as soon as I do stateful TCP connections over GRE I pretty quickly kill the machine
[11:45] <smb> jamespage, You mean even with Saucy and new tools? ... that would be bad
[11:45] <jamespage> smb, yes
[11:46] <jamespage> and I do mean that!
[11:46] <smb> darn
[11:46] <jamespage> smb, nothing in the kern.log
[11:48] <smb> jamespage, If a VM reacts the same way and does not take down the host, maybe a dump can be taken
[11:48] <smb> Though it seems 3.11 may have changed the dmesg format and crash has to play catch up
[11:48] <jamespage> smb, VM do  the same thing
[11:53] <jamespage> smb, OK - master of upstream with my 3.11 support enabled does not crash
[11:54] <jamespage> so I'm either missing a fix or I broke the backport
[11:55] <smb> Hm, neither particularly pleasant to find
[12:02] <jamespage> smb, also native kernel module is OK
[12:02] <jamespage> smb, so probably my backport - suspect I'm missing something
[12:02] <jamespage> I think I see
[12:06] <smb> Wished I could be a bit more helpful but I am not very good at multitasking :/
[13:56] <Jeeves_Moss> I'm trying to recover e-mails that were orignally in a Dovecot server.  I was able to recover all but one's e-mails from .PST files on the local system.  I currently have the directory where the e-mail files were orignally, but I'm lost hot to make a simple install so I can install JUST those e-mails, so I can connect with outlook and pull them into Exchange.
[13:59] <greppy> Jeeves_Moss: it may be simpler ( or harder ) than you think, if it was an IMAP server, you may be able to just convert from Maildir to Exchange, a google search may help you there.
[14:01] <Jeeves_Moss> greppy, I think (if memory serves), I've tried that.  I'm going to try spinning up the VM, and uncompressing the 6Gb file to the drive
[14:01] <Jeeves_Moss> greppy, if you're going to be around for a bit, I'd apericate any help you can offer
[14:09] <plars> jamespage: I'm having trouble getting my maas to get as far as a juju deploy for the 12.04.3 testing. If you have a simple way to do it in your environment again, it would be nice to have that result filled in on the iso tracker
[14:09] <jamespage> plars, sorry - I dont
[14:09] <plars> jamespage: I'm getting further than before, but it's one of the last remaining things we need before release
[14:10] <jamespage> plars, have you pinged matsubara?
[14:10] <plars> jamespage: no, not yet
[14:10] <plars> jamespage: doesn't seem to be on
[14:11] <zetheroo> just was looking at this (http://www.serenux.com/2009/11/howto-fix-a-missing-eth0-adapter-after-moving-ubuntu-server-from-one-box-to-another/) since I am experiencing the same issue .... but I am wondering if there was any way to make it so that the disks can be removed from one host and placed into a backup host and not have to manually fix the MAC address issue .... !?
[14:36] <TJ-> zetheroo: If you make /etc/udev/rules.d/70-persistent-net.rules "chmod o-w" (read-only) then "/lib/udev/write_net_rules" will not change it, instead writing the updated rules to /dev/null
[14:44] <zetheroo> TJ-: would that allow a disk from host1 to be placed into host2, booted up and have network up and running?
[14:45] <TJ-> zetheroo: If you prevent *any* rules ever being written to "/etc/udev/rules.d/70-persistent-net.rules"
[14:45] <zetheroo> ok, and I can leave /etc/network/interfaces alone?
[14:46] <TJ-> You could just make /lib/udev/write_net_rules non-executable, too
[14:47] <TJ-> You could even add a custom rule that fires on the same triggers and blanks 70-persistent-net.rules if it detects that all the MACs are now different to those recorded
[14:49] <TJ-> I have something similiar in an emergency USB flash drive so I can use it in multiple-interface servers (some have 10 i/fs) and it won't change assignments
[14:51] <zetheroo> what would be the safest and simplest method in your opinion?
[14:53] <TJ-> Well, if you know the HDD is going to move then prevent write_net_rules ever running from day 0. If not, and it has already written to 70-persistent-net.rules you've got to edit that file. So then you need to decide whether to do it manually, or have a udev script that will do it for you when/if ever needed
[14:55] <zetheroo> we have 3 running servers and one redundant server - all 4 are identical hardware and software wise ...
[14:56] <zetheroo> the idea is that in case a server dies we can simply remove the disks and place them into the redundant server and power it up without a hitch ...
[14:58] <TJ-> Well it doesn't take much to manually edit the files, but if you want it automatic, you'll need to add a script to do it
[14:59] <zetheroo> oh ok ... so just changing that files permissions to read-only would not do the trick ...
[15:02] <TJ-> As I said, once it has been written to once you'll have to edit anyhow, so either you never let it be written to, or you must edit one way or another
[15:06] <zetheroo> ok, so I removed the disk from a server and placed it into the redundant server ... booted up ... no eth0 interface is there ... if I comment out the SUBSYSTEM line in /etc/udev/rules.d/70-persistent-net.rules and then change it to read-only, then reboot the system, will that then work?
[15:10] <Linuxx> @zetheroo: Do you see the interface with dmesg | grep eth?
[15:21] <zul> jamespage/roaksoax: https://code.launchpad.net/~zulcss/neutron/usage-audit/+merge/181592
[15:22] <jamespage> smb, \o/ I'm not exploding my test servers any more
[15:22] <jamespage> needed one extra commit
[15:22] <smb> Yay!
[15:24] <TJ-> If a process being 'controlled' by system dbus-daemon fails, will that process restart with the same ownership as the dbus-daemon?
[15:30] <zetheroo> ok, I tested the above and it worked great :)
[15:45] <tgm4883> Does ubumirror actually work? I'm using almost a stock configuration and all I get is "Unexpected remote arg: archive.ubuntu.com::ubuntu/". This is on 12.04
[15:50] <hallyn_> zul: saucy connecting to saucy using xen+ssh://serge@ip/ works with virsh, hangs with virt-manager  <shrug>  giving up on virt-manager for now
[15:50] <zul> weird
[15:50] <hallyn_> yeah
[15:50] <smb> hallyn_, Does virsh connection hang with dumpxml 0
[15:51] <hallyn_> smb: yeah, even locally
[15:51] <hallyn_> virsh -c xen:/// empxml 0 or virsh -c xen:/// edit ubuntu both hang
[15:51] <smb> hallyn_, then you must have a non-fixed version of libvirt
[15:51] <hallyn_> presumably that's the cause of the virt-manager hang, makes sense
[15:51] <hallyn_> smb: non-fixed?
[15:51] <hallyn_> i've got 1.1.1-0ubuntu3
[15:52] <smb> hallyn_, This should have been fixed in previous
[15:52] <smb> versions
[15:52] <zul> jamespage:  http://people.canonical.com/~chucks/ca/ (that should get us up to date with whats in the cloud archvie)
[15:52] <smb> Hrm... And I thought the patch went into upstream at some point
[15:53] <smb> zul, did the pull of 1.1.1 contain d7a45bf22368161869963b92a0a1d5599590fdf5 from upstream?
[15:53] <smb> xen: Use internal interfaces in xenDomainUsedCpus
[15:53] <zul> smb:  which one is that?
[15:54] <zul> smb:  gitweb link?
[15:54] <smb> zul, That would be the sha1 in upstream... :-P
[15:54] <zul> smb:  lemme check
[15:54]  * hallyn_ is d/ling the pkg source to check
[15:55] <hallyn_> smb: no it's not in the pkg
[15:56] <smb> hallyn_, Yes, seems this did come too late into upstream...
[15:56] <hallyn_> smb: are there other patches I"ll need to push?
[15:57] <smb> hallyn_, It is in upstream master, so just needs picking
[15:57] <hallyn_> smb: yes, doing that.  but are there any ohters i'll need
[15:58] <hallyn_> ok, building and testing :)
[15:58] <hallyn_> thx smb
[15:58] <smb> hallyn_, no, just had to drop ubuntu-xen-fix-api-deadlocks.patch in favour of the other but I suppose dropping happened
[15:59] <smb> There are two patches I did not push upstream as they seemed rather depending on the way we have modified some things
[16:00] <hallyn_> i don't see anything like 'fix api deadlocks' in git history
[16:01] <smb> hallyn_, That was only the version we had in the package
[16:01] <hallyn_> oh - in our package, gotcha
[16:02] <smb> Right, and it had been evolving for the better anyway, so its kind of good it is gone but kinda bad the replacement did not make it into 1.1.1 upstream
[16:02] <hallyn_> patch is still in the directory, but not listed in series
[16:02] <hallyn_> ok, build is proceeding.  lessee
[16:02] <smb> hallyn_, Oh, so you can remove it while you are at it
[16:02] <hallyn_> yeah
[16:03] <hallyn_> i can also look at the -maint tree which i think has a patch for us to push
[16:04] <smb> hallyn_, feel free. I am outta here :)
[16:05] <hallyn_> smb: \o
[16:18] <hallyn_> smb: that fixed it, thanks!
[16:19] <hallyn_> BUT!  I now want to test with all the v1.1.1-maint patches pushed.
[16:19] <hallyn_> waiting for zul to get back so i can ask him to doublecheck :)
[16:36] <FunnyLookinHat> Anyone here know why SSH claims it can't resolve a domain but host -a returns the correct A record for it?  I just created it on my DNS host a minute ago, but I'd assume that seeing it with host -a would mean it's queryable
[16:38] <sgran> does it have an underscore in the name?
[16:39] <sgran> host -a queries for an A record
[16:39] <sgran> ping/ssh etc use a libc function that checks for validity of a hostname
[16:39] <sgran> these often overlap, but don't need to
[17:07] <hallyn_> smb: ppa:smb/xen seems to be working fine with libvirt
[18:37] <explodes> When I log in, it shows the temperature in Celsius, how do I show it in Fahrenheit?
[18:41] <sarnold> explodes: grep for 'sensors' in your /etc/update-motd.d/ -- if you find it, replace it with sensors -f
[18:41] <explodes> sarnold: Nice, thanks.
[18:44] <hallyn_> FunnyLookinHat: unlikely, but also conceivable that you have a .ssh/config entry for the domain which sets up a proxy, and the proxy doesn't actually know the host...
[18:46] <FunnyLookinHat> hallyn_, got it figured out - thx :D
[18:46] <hallyn_> ok
[19:31] <K4k> Hi, perhaps this is a better place to ask my question since it's more server specific. I'm trying to configure postfix smtp/sasl with dovecot and I keep getting the following error: "warning: SASL authentication failure: unable to canonify user and get auxprops" but I'm unsure why. All documentation I've found seems to indicate I've got it all configured properly.
[19:34] <sarnold> K4k: do you get better error messages if you look through log files?
[19:34] <K4k> mail.log is telling me what I posted, nothing more useful unfortunately
[19:35] <sarnold> oh, drat, I hoped that was a user-facing error message that glossed over details. :(
[19:36] <K4k> my client connects and presents my server certificate, then prompts for my password
[19:36] <K4k> Then tells me "SASL authentication failed"
[19:38] <K4k> on the server side, the chain of events looks like this: http://pastebin.com/9nPUfTFK
[19:39] <sarnold> K4k: your client presents the server cert? o_O
[19:39] <K4k> my server is using a mismatched hostname self signed cert at the moment so the client says "Hey, this looks funny, do you want to proceed anyway"
[19:40] <K4k> nothing weird
[19:40] <sarnold> oh, it presents the cert to the user :) I thoght it was trying to handshake iwth the server's certificate..
[19:41] <K4k> oh, haha, no
[19:43] <K4k> Looks like Ubuntu runs postfix in a chroot which might be an issue...
[19:54] <K4k> WOO! I think I fixed it
[19:54] <K4k> it was a client side configuration issue
[19:54] <K4k> I'm glad I'm not crazy
[19:55] <sarnold> K4k: woo, what was it?
[19:56] <K4k> my smtp_url line in my muttrc originally was user@example.com:587 and I saw in an example document that they didn't have the user@ portion so I took that out and it started working
[19:56] <K4k> it's a little unsettling though because it now doesn't ask for my password to send email...
[19:57] <K4k> so something is still not quite right
[21:14] <adam_g> Daviey, jamespage http://people.canonical.com/~agandelman/ca/grizzly/2013.1.3/
[21:14] <adam_g> ready for grizzly-proposed
[21:24] <Daviey> adam_g: visibly, they look good.  Are you also uploading to raring?
[21:37] <adam_g> Daviey, they're already in queue
[21:37] <adam_g> Daviey, gonna push to grizzly-proposed unless you have any objections