[00:08] <dougwt> I apologize in advance if this is the wrong place to ask this type of question, but I am trying to setup my first Apache server on Ubuntu and am a bit confused about what permissions I should be using. I've got multiple VirtualHosts setup. Should the document roots be chowned to the user and his group, or do I need to chown them to www-data?
[00:51] <qman__> dougwt, www-data is what apache runs as, so in order to serve pages, www-data needs read access
[00:51] <qman__> www-data should not have write access unless specifically required by your application, and when it does, it should be carefully limited
[01:05] <dougwt> qman__: Thanks. So it'd be fine to keep the document root and it's filed chowned to the user with 755/644 permissions respectively?
[01:05] <dougwt> its files*
[01:19] <qman__> yes
[01:25] <dougwt> great
[01:25] <dougwt> thanks again
[03:44] <redact3d> hi anyone guy the latest source of xymon compiled and working?
[03:48] <twb> Never heard of it
[03:56] <redact3d> network monitoring system - think nagios but easier to configure
[08:59] <jamespage> lynxman, do you use racoon?
[09:00] <lynxman> jamespage: hmm nope :)
[09:00] <jamespage> lynxman, bah - was hoping to palm something off onto you :-)
[09:00] <lynxman> jamespage: lol, I used to though, so you can still handle it to me
[09:01] <jamespage> lynxman, ever see this bug 972786
[09:04] <lynxman> jamespage: will give it a go today :)
[09:04] <jamespage> ta
[09:05] <jamespage> lynxman, actually that was not hard to test
[09:05] <jamespage> works for me in precise
[09:05] <jamespage> i.e. bind to new interfaces.
[09:16] <lynxman> jamespage: heh :) it's one of its basic operations, you want to take it then?
[09:17] <jamespage> lynxman, I don't actuall think its a bug - I could not reproduce with the stock configuration
[09:17] <jamespage> lynxman, don't worry about looking at it.
[09:22] <lynxman> jamespage: I'll tackle it after lunch so just let me know :) need to be copywriting now
[09:23] <jamespage> lynxman, thanks for the offer - I'm sure I'll find something more important for you todo during triage today
[09:23] <jamespage> enjoy copywriting
[09:23] <lynxman> jamespage: :)
[09:24] <lynxman> jamespage: this whitepaper wont update on its own :D
[09:24] <Barbo91> Hi guys, first of all, sorry for my english, i'm italian xD Now i explain my problem. First of all i'm on a Ubuntu 10.04 server with openssh,ssl and openvpn. I have to join all the clients that connect on the VPN in the same network. I can do it with bridging, am i right?
[09:28] <Barbo91> someone that can help me? I need help to configuring the bridge :S
[09:33] <jamespage> Barbo91, I think you either want to look at https://help.ubuntu.com/10.04/serverguide/C/network-configuration.html#bridging
[09:33] <jamespage> or you might actually want to look at setting your server up as a router
[09:35] <jamespage> Barbo91, in which case this https://help.ubuntu.com/community/Router might be useful
[09:35] <Barbo91> hi jamespage thanks for help, going to read that doc! by the way, the thing that i need is to create a "virtual" eth1 to bridge it, i think
[09:35] <jamespage> Barbo91, really depends on how your network is setup
[09:36] <jamespage> I do something similar with IPv6 routing
[09:36] <jamespage> Barbo91, do you just have a single network interface? or multiple?
[09:37] <Barbo91> jamespage, i got a single network interface. with a static pubblic IP.
[09:38] <jamespage> Barbo91, have you read https://help.ubuntu.com/11.10/serverguide/C/openvpn.html ?
[09:38] <Barbo91> jamespage,  my network is that i got a network with all servers on, and i don't want that clients that connect my VPN see those server. I want to create a "dedicated" private network client-to-client...
[09:39] <jamespage> Barbo91, You might be able todo that with openvpn plus bridges - but ufw might help there as well
[09:40] <Barbo91> i tried to create a eth 0:0 and bridge it, when i launched the script to create bridge, the interface 0:0 was deleted and the bridge was created on the eth0 and i closed my ssh connection -.- closed out of my own server xD
[09:41] <Barbo91> jamespage, already got a "hardware" firewall on my net
[09:41] <uksysadmin> hey all - can anybody here help with maas? I'm getting  a "Bad archive mirror" message but the pxe booted machine seems to be able to get to the right places...
[09:41] <uksysadmin> (on ubuntu 12.04 b2)
[09:42] <jamespage> Barbo91, yeah - but the VPN penetrates that so you openvpn server can act as a soft firewall between your clients and your server network
[09:42] <jamespage> uksysadmin, is that really early during the install process?
[09:42] <Barbo91> jamespage, yeah ur right! didn't thinked about it O.O thanks!
[09:44] <jamespage> lamont, would a sync of bing9 1:9.8.1.dfsg.P1-3 be a good idea today?
[09:45] <Barbo91> jamespage, thanks for all your help, now i'm going to read the docs u linked me then i will write back soon if i need help!! Newbie network administrator here :P
[09:46] <jamespage> Barbo91, np - you are welcome!
[09:56] <uksysadmin> hi jamespage - yes its just after the kernel messages and about to commence the install
[09:57] <jamespage> uksysadmin, wget calls?
[09:57] <uksysadmin> I've got a console up and can perform the wget it wants
[09:57] <uksysadmin> yes
[09:57] <jamespage> uksysadmin, OK - I think I've seen this to - but can't reproduce at the moment
[09:57] <jamespage> I believe that the wget calls made by the installed route through the squid-deb-proxy on the MAAS server
[09:58] <jamespage> and its got stuffed in some way I don't understand
[09:58] <uksysadmin> ok - I'll take a look around
[10:01] <uksysadmin> jamespage, yes I'm seeing 403 Forbidden messages in syslog...
[10:02] <akoma1s_> hello, how can I do something like 'if lsacpi | grep VBOX; then set default=3; fi' in grub.cfg?
[10:02] <uksysadmin> and given my test network falls outside a "localnet" default address range could be the reason
[10:02] <jamespage> uksysadmin, odd -lemme look at my setup here
[10:02] <jamespage> uksysadmin, have you got the exact URL that is failing?
[10:03] <uksysadmin> http://archive.ubuntu.com/ubuntu
[10:04] <uksysadmin> I can see the wgets going this URI /ubuntu/dists/precise
[10:07] <jamespage> uksysadmin, does it fail in choose-mirror?  I think that is where I saw the issue
[10:08] <uksysadmin> yes
[10:08] <jamespage> uksysadmin, any more info in those 403 messages?
[10:09] <uksysadmin> just looking through syslog now
[10:09] <jamespage> uksysadmin, /var/log/squid-deb-proxy might have a bit more in it
[10:11] <uksysadmin> ah, think I've found *my* issue
[10:12] <uksysadmin> I've a test setup on my desk on 172.15.0.0/16 network
[10:12] <uksysadmin> and the allowed ranges start from the expected non-internet ranges
[10:12] <uksysadmin> technically that range is internet
[10:12] <uksysadmin> I'll re-ip and redo
[10:13] <uksysadmin> (I'll add this esoteric range though as a test)
[10:15] <jamespage> uksysadmin, its quite possible I saw an actual network issue
[10:15] <uksysadmin> jamespage, I can attest to that though.
[10:16] <uksysadmin> I tried this on a machine in our datacentre and after a couple of re-tries it worked
[10:16] <uksysadmin> which is on a valid range
[10:16] <uksysadmin> (btw - after adding in my new IP range, this is installing)
[10:16] <uksysadmin> cheers for pointing me to the squid-deb-proxy setup though
[10:20] <uksysadmin> in fact I was doing this on a couple of machines next to me to see what was up with my datacentre set up
[10:33] <jamespage> uksysadmin, np
[10:58] <lamont> jamespage: I would support that
[11:02] <jamespage> lamont, ack - I'll do the sync now then
[11:10] <koolhead11> why we dont`t have memcache as dependency with openstack-dashboard?
[11:15] <koolhead11> i need a workaround of this https://bugs.launchpad.net/ubuntu/+source/openstack-dashboard/+bug/960070
[11:15] <koolhead11> if we are not getting memcache as dep for dash/horizon
[11:44] <zul> jamespage:  racoon and nis is so 19175
[11:45] <zul> er...1917
[11:45] <zul> koolhead11: open up an ubuntu task for it
[11:47] <koolhead11> zul, sorry 4 dumb question but where should i open ubuntu task? launchpad?
[11:48] <zul> koolhead11: "Also affects distrobution"
[12:08] <jamespage> zul, lol
[12:52] <zul> good morning
[13:01] <Caribou> pmatulis: around yet ?
[13:07] <alcuadrado> I have to start "mongod --journaling" as a service
[13:07] <alcuadrado> is there a way to pass arguments to the executable with the command service?
[13:11] <foo> I have user1 who logs into server and starts a screen. I am user2. as user2, I am root. I sudo su - user1. I do screen -x but I get this: Cannot open your terminal '/dev/pts/0' - please check - any ideas?
[13:26] <jamespage> roaksoax, around? having a few issues with maas/cobber-web not playing nicely
[13:29] <roaksoax> jamespage: here, what'ss wrong?
[13:29] <jamespage> roaksoax, OK - so I'm running maas from the PPA and I also have cobbler-web installed on the same server
[13:29] <jamespage> was working OK but now I get an error accessing cobbler-web
[13:29] <jamespage> "[Errno 13] Permission denied: '/var/lib/cobbler/webui_sessions/sessionid8ab2bb421d7efeb891e01e619b6b457e'"
[13:30] <jamespage> roaksoax, I think I know what it is - the maas wsgi stuff now runs as 'maas' user - but the content of that directory is only r/w able by www-data
[13:31] <roaksoax> jamespage: uhmmm right. Yeah that might be it
[13:31] <zul> ummm....why is keystone still in universe?
[13:31] <jamespage> roaksoax, not quite sure what the best fix is...
[13:31] <jamespage> zul: maybe no-one seeded it yet :-)
[13:31] <roaksoax> jamespage: yeah, let me give it a spin
[13:32] <jamespage> roaksoax, wanna bug for that one?
[13:32] <zul> jamespage:  partially it was
[13:33] <roaksoax> jamespage: well I don't know whether we should consider it a bug per se (in either side) because if we have MAAS we should not have cobbler-web isntalled, obviously we are not prevented from using it
[13:33] <roaksoax> jamespage: but pleas,e ifle one. I think it will go away once we ship cobbler inside maas
[13:33]  * jamespage shudders
[13:33] <jamespage> OK
[13:37] <roaksoax> jamespage: uhmmm so the weird thing here is that only the MAAS wsig should run as MAAS user right?
[13:39] <jamespage> roaksoax, so maas-http.conf sets the user/group config for the WSGI for the server - not just for maas
[13:39] <roaksoax> jamespage: exactly so that's the issue
[13:39] <jamespage> roaksoax, I think so
[13:40]  * jamespage is not a WSGI expert
[13:55] <roaksoax> jamespage: did you file the bug?
[13:55] <roaksoax> jamespage: can you provide me wiht the link please?
[13:55] <jamespage> roaksoax, doing it now - it was in the queue behind one for juju
[13:59] <roaksoax> jamespage: cool thanks
[14:02] <jamespage> roaksoax, https://bugs.launchpad.net/ubuntu/+source/cobbler/+bug/979899
[14:03] <roaksoax> jamespage: cool thnkas
[14:12] <hallyn> stgraber: bug 974584, I'm going to just rmdir $rootfs/dev/shm at end of templates.  Do you have a better idea?
[14:12] <hallyn> (Note, /etc/init/mounte-dev.conf is technically wrong, for anyone who uses /dev on a live system as not a separate fs, but I assume we just say we don't support that non-standard config?)
[14:14] <zul> Daviey: http://paste.ubuntu.com/926472/
[14:14] <stgraber> hallyn: that's definitely the easiest and the safest change, go for it
[14:15] <hallyn> thanks, will do.  do you have any other changes to queue up?
[14:15] <stgraber> it won't fix existing containers, but we didn't get that many reports of the bug for it to be a concern
[14:15] <stgraber> nope, I've been doing installer stuff lately, no lxc related stuff
[14:15] <stgraber> make sure you upload before 21:00 UTC though (final freeze)
[14:17] <hallyn> today???
[14:17] <hallyn> i thought ff was ina  week
[14:17] <zul> ditto
[14:17] <hallyn> i'll leave the workaroudn noted in the bug for ppl who need it for existint containers
[14:25] <hallyn> zul: joke's on us https://wiki.ubuntu.com/PrecisePangolin/ReleaseSchedule
[14:26] <zul> hallyn: are you sure?
[14:26] <zul> FinalFreeze Apr 12
[14:30] <hallyn> zul: maybe i misunderstood you.  i thought you meant you had been wrongly thinking like me that it was next week
[14:30] <zul> hallyn: meh :)
[15:02] <MoleMan> I believe I am recieving a DOS attack in progress, what can I do?
[15:04] <Jeeves_> Look the other way!
[15:08] <MoleMan> doesn't help... vnstat is showing 50Mb/s in and 5Mb/s out compared to my other box which is showing  3.3Mb/s in 8 Mb/s out...
[15:09] <MoleMan> and I don't want to go ask the host for advice because I know that they have a strict 3 strike policy on DoS attacks, and we've had more than that already...
[15:10] <zul> contact your isp and get them to cut it off
[15:11] <Jeeves_> MoleMan: What do you do to receive those attacks?
[15:14] <MoleMan> god only knows... It's gameservers so I would presume a script-kiddy thats upset at getting banned...
[15:14] <Jeeves_> Ah
[15:15] <hallyn> stgraber: lucid containers don't have /dev/shm mounted either.  any objections to my adding it back to /lib/init/fstab.lxc?
[15:16] <stgraber> hallyn: for lucid => natty it should be safe, we did the /run change in oneiric IIRC
[15:16] <stgraber> hallyn: though I can't see it breaking on oneiric if added to fstab.lxc there too
[15:16] <hallyn> fooi
[15:16] <hallyn> lemme try
[15:19] <hallyn> oddly, in oneiric /dev/shm is correctly created as a symlink
[15:19] <hallyn> (but /run/shm isnot mounted)
[15:19] <hallyn> anyway i guess that's SRU material anyway.  i wasn't thinking right
[15:33] <hallyn> utlemming: do you have a minute?  lxc-create -t ubuntu-cloud -n cloudo1 -- -r oneiric (and -r lucid) are failing for me?
[15:33] <hallyn> is that a transient ami error?
[15:33] <utlemming> hallyn: otp...do you have a paste of the failure?
[15:34] <hallyn> utlemming: http://paste.ubuntu.com/926586/
[15:35] <utlemming> hallyn: can you kick me a bug and I'll get this hashed out?
[15:36] <hallyn> utlemming: ok.  it looks ike the problem is just that the template looks for ubuntu-11.10-server-cloudimg-amd64.img:, but
[15:36] <hallyn> the tarball has oneiric-server-cloudimg-amd64.img
[15:37] <utlemming> hallyn: that is what my thought was...the naming between dailies and releases changes
[15:37] <hallyn> feh!
[15:37] <hallyn> ok, 1 min for bug #
[15:39] <hallyn> utlemming: bug 979996
[15:40] <hallyn> utlemming: I could try it myself, but with final freeze today am afraid i'll mess things up worse than they are.  at least precise ones work :)
[15:40] <hallyn> utlemming: thanks
[15:45] <hallyn> stgraber: http://people.canonical.com/~serge/lxc-shm.debdiff
[15:46] <stgraber> hallyn: you realize this means we won't be able to create a Q container on precise and have /dev/shm working?
[15:47] <hallyn> stgraber: well not without a patch...  ok I should use the release # and >= 12.04?  how can iget that in bash?
[15:48] <hallyn> substr($release,0,1) <= o? :)
[15:48] <hallyn> stgraber: well, my hope would be that we could fix initscripts by q :)
[15:48] <hallyn> feasible?
[15:49] <stgraber> hallyn: good point, fixing the logic of ischroot/initscripts in Q sounds like a plan :)
[15:49] <stgraber> hallyn: can you make sure we have a bug (probably against initscripts) for that so we don't forget about it?
[15:49] <hallyn> stgraber: shall i mark the current bug as affecting initscripts, and leave it open?
[15:49] <stgraber> I know we have some shell functions to before_release(a, b) but I don't think we ship that by default or even have a standard implementation of it
[15:50] <stgraber> probably something to get into distro-info
[15:50] <stgraber> hallyn: yeah, opening an initscripts task would be good
[15:50] <hallyn> yeah, standard implementation of that woudl be good
[15:52] <hallyn> hm, manpage of lsb_release says see also lsb(8), which doesn't exist :(
[16:05] <stgraber> hallyn: won't your change always print an error message? either because it's not a directory or because it doesn't exist (after the rmdir)?
[16:05] <stgraber> hallyn: the || true avoids the failure but not the error
[16:05] <hallyn> stgraber: should I shut it up?
[16:05] <stgraber> hallyn: or is it running early enough in the bootstrap process that the user won't see it anyway?
[16:06] <hallyn> no it's at the very end of creation
[16:06] <hallyn> *I* didn't notice it, but some people do like to file bugs about every msg...
[16:06] <hallyn> lemme check again
[16:06] <stgraber> hmm, 2>/dev/null would probably be a good idea then or even: [ -d /dev/shm ] && rmdir /dev/shm || mv -f $rootfs/dev/shm $rootfs/dev/shm.bad
[16:07] <stgraber> (sorry, should have seen it earlier)
[16:07] <hallyn> I don't want to go chaining like that and have it be as fragile as mounted-dev.conf
[16:07] <hallyn> (when i do it slightly wrong)
[16:07] <hallyn> mind you what you have there *does* look right
[16:08] <hallyn> stgraber: I do get mv: cannot stat `/var/lib/lxc/p3/rootfs/dev/shm': No such file or directory
[16:08] <stgraber> right, maybe in two then: [ -d "$rootfs/dev/shm" ] && rmdir $rootfs/dev/shm
[16:08] <stgraber> and then [ -e "$rootfs/dev/shm" ] && mv $rootfs/dev/shm $rootfs/dev/shm.bak
[16:08] <stgraber> before the ln call
[16:08] <hallyn> do you strongly prefer that to 2>/dev/null?
[16:08] <stgraber> I guess I do, in the sense that if mkdir then fails we won't silence the error
[16:09] <hallyn> btw, have you kicked my package from the archive? :)  can i re-use the versoin #?
[16:09] <stgraber> yeah, you can reuse the version number, I'll kick it out of the queue
[16:10] <hallyn> stgraber: thanks.  I'll wait till you look at the new debdiff after dinner to re-push?
[16:11] <stgraber> hallyn: upload directly to the queue, I'll review it there
[16:11] <hallyn> kthx
[16:16] <ruben23> hi guys any apps can check for HDD bad sector abd blocks..?
[16:17] <rbasak> ruben23: badblocks. Or smartctl to see what the disk thinks.
[17:00] <roaksoax> jamespage: new ppa upload fixes ur issure
[17:04] <roaksoax> utlemming: still not around?
[17:04] <utlemming> roaksoax: here
[17:04] <roaksoax> utlemming: so I wasn't able to get the image even booting (or so it seems)
[17:05] <roaksoax> utlemming: were you able to look into it further?
[17:05] <utlemming> roaksoax: okay, I'll see if I can to it. I'm pretty buried right now
[17:07] <roaksoax> utlemming: alright, not sure whether Daviey needs this for EOD or EOW
[17:08] <utlemming> roaksoax: EOD isn't going to happen, I have another higher priority issue going on right now
[17:08] <roaksoax> utlemming: no worries, just let me know when you have something I can test. Thanks! :)
[17:09] <utlemming> roaksoax: sure, if I get blocked, I'll work the usb stick
[17:22] <zul> adam_g: patch refreshed?
[17:26] <adam_g> zul: which?
[17:27] <zul> glance
[17:27] <adam_g> zul: yeah, i cherry picked the paches directly from upstream and applied them to -proposed
[17:27] <zul> adam_g: excelente
[17:33] <hallyn> utlemming: are you going to be able to look at the lxc-ubuntu-cloud bug in the next few hours, or is that not in the cards?
[17:44] <hallyn> SpamapS: Uh, combine a cable line unearthed by the builders at the neighbor, and a wildman with a lawnmower in back, and I may end up not online in a few mins.  fyi
[17:49] <hallyn> the install is going very slowly, though I can't really tell, it migh tjust be my apt-cache-ng mirror or gateway (both of which are my netbook :) slinging data slowly
[17:55] <SpamapS> hallyn: DOH
[19:21] <robo_> can someone who's running openjdk on ubuntu 10lts run 'jps' for me and see if it returns the bootstrap PID?
[19:22] <kees> robo_: does it work if you run jps as root?
[19:22] <robo_> kees, nope
[19:22] <robo_> it just returns the pid of the jps program itself
[19:23] <kees> weird
[19:33] <zul> kees: no one should use java anyways
[19:34] <kees> lol
[19:35] <henkjan> SpamapS: enjoying your time at perconalive?
[19:43] <blair> svn 1.6.18 was released today with some repository corruption issues, how do i request an update from 1.6.17 to 1.6.18.  i've opened https://bugs.launchpad.net/ubuntu/+source/subversion/+bug/980087
[19:53] <ewingate> how can i tell why my filesystem is in read-only mode
[19:53] <ewingate> and how can i tell which part of my filesystem is in read-only mode?
[19:54] <patdk-wk> cat /proc/mounts
[19:54] <RoyK> ewingate: ls -l will tell you permissions for files/dirs
[19:54] <ewingate> patdk-wk: thanks thats what i was lookin for!
[19:55] <patdk-wk> royk, inferrence :)
[19:55] <ewingate> patdk-wk: any idea how i can tell why it switched to read-only? or how do i get it out of read only? reboot?
[19:56] <patdk-wk> locally? dmesg
[19:56] <patdk-wk> only way I know to get it back out, is a reboot
[19:56] <RoyK> ewingate: usually it remounts to read only if something goes BAD
[19:56] <patdk-wk> normally goes into readonly mode if there is issues reading/writing to the disk
[19:56] <patdk-wk> big issues, like, there is no more disk :)
[19:56] <patdk-wk> or disk didn't respond after 90 seconds
[19:57] <RoyK> because it's safe to remount to readonly and wait for fsck to sort things out later
[19:57] <ewingate> ah ok.. so rebooting the server would get me back in business?
[19:57] <RoyK> hopefully, yes ;)
[19:57] <ewingate> or it will at least check the disk for errors?
[19:57] <ewingate> ok thanks
[19:58] <RoyK> if you don't have console access
[19:58] <RoyK> it might render the server useless until you get to the console
[20:11] <Daviey> zul: hey, thoughts on https://bugs.launchpad.net/ubuntu/+source/quantum/+bug/979192
[20:12] <zul> Daviey: i think in the long run that will help us for q so do it now
[20:14] <Daviey> zul: happy with the diff?
[20:14] <zul> Daviey: yeah
[20:14] <Daviey> zul: Would ypou be so kind to sponsor ir?
[20:14] <Daviey> it?
[20:15] <zul> uh?
[20:15] <zul> the patch didnt have a changelog
[20:16] <Daviey> zul: yes, but can you incorporate it? :)
[20:16] <zul> *sigh*
[20:16] <zul> yeah i can
[20:16] <adam_g> http://paste.ubuntu.com/926983/
[20:17] <adam_g> Daviey: jamespage: ^ that look reasonable for preserving squid conf? waiting on a build to do some last minute dist-upgrade tests
[20:17] <zul> adam_g:  holy crap dude
[20:17] <adam_g> zul: eh?
[20:18] <zul> largish
[20:18] <adam_g> well, yeah
[20:18] <zul> dont mind me
[20:20] <jamespage> adam_g, in the preinst for the squid package?
[20:20] <zul> Daviey: done
[20:21] <Daviey> adam_g: i think it's a reasoable	solution.
[20:21] <adam_g> jamespage: thats in postinst
[20:21] <adam_g> jamespage: AFAICS, /etc/squid3/ wont be there in preinst
[20:22] <jamespage> adam_g, yeah - that was what I was wondering
[20:22] <adam_g> jamespage: but original /etc/squid/ shouldn't have been purged when 'postinst upgrade' is run, will it?
[20:23] <jamespage> adam_g, maybe - I'd want to test it to be sure :-)
[20:23] <jamespage> adam_g, we have a transitional squid package right that is built out of squid3 source ?
[20:24] <adam_g> jamespage: yeah, about to do that.
[20:24] <adam_g> jamespage: yea
[20:27] <jamespage> adam_g: if that tests OK I think it looks good.
[20:28] <adam_g> jamespage: hmm, well it looks like that postinst wasnt called
[20:30] <adam_g> jamespage: though, the original /etc/squid/ isn't purged either by default. im wondering if we can not even worry about moving the file, since its going to stay there regardless
[20:31] <jamespage> adam_g, hmm - maybe
[20:31] <jamespage> actually that would explain a diff bug that I triaged today with regards to the logrotate from squid sticking around post upgrade
[20:34] <SpamapS> henkjan: it was a lot of fun, but I was just there for 1 day (yesterday)
[20:35] <adam_g> jamespage: hm
[20:35] <jamespage> adam_g, is this related to the fact that the package is not tracking them as configuration files?
[20:36] <adam_g> jamespage: what? the fact that the directory is not removed?
[20:37] <jamespage> adam_g, sorry brain a little fuddled
[20:37] <henkjan> SpamapS: ah. i saw al the buzz on twitter in #perconalive
[20:38] <henkjan> i've been to percona live londen last year
[20:38] <henkjan> the mysql community is awesome
[20:39] <adam_g> jamespage: thats okay. since the old squid.conf is never actually purged, we could probably skip maintaining a list of versions+md5's and just put a note at /etc/squid3/README explaingin the transition?
[20:40] <jamespage> adam_g, if it remains intact then yes that is OK
[20:41] <jamespage> although I believe that should go into NEWS for the package
[20:42] <adam_g> jamespage: i can do that
[20:42] <jamespage> adam_g, http://www.debian.org/doc/manuals/developers-reference/best-pkging-practices.html#bpp-news-debian
[20:44] <adam_g> jamespage: thanks
[20:44] <jamespage> I think dch will generate an entry for you :-)
[20:45] <hallyn> SpamapS: hm, this could be spurious as i've only done one test each, but a 'time find / -type f > o 2>&1' in qemu with rbd took 5.7 seconds vs 4.7 seconds in qemu without rbd
[20:45] <hallyn> trying again
[20:46] <hallyn> hoping it has to do with pagecache in host being hot
[20:47] <SpamapS> hallyn: that would be really odd
[20:47] <hallyn> yes it would
[20:48] <SpamapS> hallyn: I was looking at how the different backends are used.. and it seems like they are completely independent of one another
[20:49] <hallyn> it's somewhat consistent though
[20:52] <resno> what kinda of server monitoring tools do people use? i was looking at monitorix
[20:55] <SpamapS> resno: nagios/icinga are good (the latter being a fork of the former), collectd is cool.. statsd+graphite is pretty cool.
[20:56] <resno> any thoughts on monitorix?
[21:08] <RoyK> !bug 978458
[21:43] <T3CHKOMMIE> hey guys, i got a bit of a dumb request. I am trying to make a 10 second video of a hole bunch of printlns in terminal. I was wondering if anyone know of a good cat /proc/* command that would spit out a bunch of data for about 10 secnds. any ideas?
[21:50] <aceat64> Anyone know why the "guided partitioning with lvm" option creates a primary partition (on this server it's /dev/sda1) that looks like it should have been /boot but doesn't appear to be used
[21:50] <aceat64> https://gist.github.com/2371194
[21:51] <aceat64> instead /boot is on my lv for /
[21:59] <SpamapS> hallyn: so, did we break kvm performance or was it an aberration?
[22:00] <hallyn> SpamapS: i think it ws an aberration.  overall the times with rbd were higher, but i did get one 5 second time.  (only did 4 runs each)
[22:00] <hallyn> i'm still shocked at how slow the guest install was...
[22:00] <hallyn> but anyway, if there is a perf regression i'll have to find and fix it :)
[22:00] <hallyn> next week i'm hoping to do my perf runs
[22:01] <SpamapS> hallyn: precise-proposed awaits you :)
[22:01] <hallyn> lol
[22:17] <jkyle> so we're not supposed to edit resolv.conf by hand...what do we edit? looking at the resolvconf man and it seems things have gotten quite a bit more complex
[22:18] <stgraber> jkyle: http://www.stgraber.org/2012/02/24/dns-in-ubuntu-12-04/
[22:18] <stgraber> jkyle: short answer is add dns-nameservers and dns-search for your interface in /etc/network/interfaces
[22:19] <jkyle> I can't use the /etc/network/interfaces file...I'm having to set up my interfaces manually via the /sys/class/net
[22:20] <jkyle> bugs with bonding in the userland utilities for interfaces
[22:22] <jkyle> ah ok, so I _can_ edit the resolv.conf.d/head though the file itself has a no edit warning
[22:22] <jkyle> that'll work
[22:22] <stgraber> right, because the no edit warning is what becomes the no edit warning of /etc/resolv.conf
[22:22] <jkyle> ok, so in my original file I have two lines, a search and nameserver declaration
[22:23] <jkyle> but these are not appearing in my resolv.conf
[22:23] <jkyle> which is empty cept for ehead
[22:23] <stgraber> right, original is just there as a backup. resolvconf will only use it if tail is a symlink pointing to original
[22:24] <stgraber> (which we do automatically on upgrade for setups where we "guess" that dns won't work without it)
[22:25] <jkyle> cool networkign is up
[22:28] <jkyle> stgraber: what was the incentive for moving to resolvconf?
[22:28] <Daviey> zul: we are now tracking stable tree for nova?
[22:30] <stgraber> jkyle: having a single tool modifying /etc/resolv.conf instead of the 5-6 we had in the past that didn't talk to each other and would end up overwriting the file every time
[22:31] <jcastro> I suspect our use of "maas effect" will get old after about a month.
[22:31] <stgraber> jkyle: another incentive was to have it in /run so that you can boot your system in read-only mode
[22:57] <kklimonda> hum, anyone is using precise on rackspace vps? I'm seeing a weird problem where root partition is mounted r/o and because of some journal issues and I get some funny errors in dmesg
[23:11] <utlemming> hallyn: you still around?
[23:17] <hallyn> utlemming: not really
[23:17] <utlemming> hallyn: lol, I'm probably way too late, but I'm submitting a fix for .img bug
[23:17] <hallyn> utlemming: yeah, final freeze has passed I'm afraid
[23:18] <hallyn> but we can queue it up for sru in a few weeks
[23:18] <hallyn> utlemming: of course I'd say there's an argument to be made that the different naming for contents of release vs devel images is a bug in itself
[23:19] <hallyn> but I assume there is plumbing out there depending on the current naming
[23:19] <utlemming> hallyn: UDS topic, I believe
[23:19] <hallyn> ok
[23:19] <utlemming> hallyn: arosales wanted me to get that consistent and I told him that plumbing was a no go this cycle
[23:19] <utlemming> there are too many moving parts and people are relying on the current convention
[23:19] <hallyn> ok (too bad :)
[23:19] <hallyn> yeah
[23:20] <hallyn> all right - have a good night
[23:20] <utlemming> you too
[23:21] <arosales> be a good thing to keep on our radar for future work :-)