[01:03] <bikcmp> I want to make an application run on startup
[01:03] <bikcmp> however, I want it to run as a user.
[01:11] <patdk-lap> bikcmp, add it to contrab
[01:11] <patdk-lap> crontab
[01:11] <bikcmp> patdk-lap: i've already gotten it from a kind man in #ubuntu
[01:12] <bikcmp> patdk-lap: i ended up going the /etc/rc.local route
[01:12] <bikcmp> about to test it.
[03:22] <Xirre> If I wanted to make a Shell Server where people can upload only a specific file type, (.dmb and .rsc) how would I go about doing so? They can also run, 'DreamDaemon "world.dmb" <port>' which will be a command to be ran. It basically runs the .dmb file and it opens the world so others can join. Is there any way I can do this?
[03:30] <qman__> for something that specific, your best option is writing up a web interface to do it
[03:30] <qman__> restricting shell access to something so narrow will be very difficult
[03:59] <Xirre> Agh. I already tried writing up one. I got stuck. I know basic php, html, a bit of shell, and some programming languages.
[04:00] <Xirre> What if I didn't have to restrict the file type? I can worry about that part later.
[04:02] <Xirre> I am back.
[05:38] <drag0nius> what is (squid-1) process? including brackets
[05:39] <drag0nius> i've removed squid3 package but this one was still running and responding on 3128
[07:59] <jamespage> Daviey, I'll pickup that iscsitarget SRU if you like; I'll do it after beta-1 as we really need to merge with Debian in quantal first
[08:01] <Daviey> jamespage: rocking!
[08:01] <jamespage> Daviey, actually thats not seeded anymore - I'll do it today
[09:10] <spidernik84> Hi everyone. Is the vm-builder project dead on 12.04?
[09:25] <eagles0513875_> hey guys is there a place where i can get the list of repositories for ubuntu server 12.04
[09:26] <ikonia> what do you mean a list of repositories ?
[09:26] <eagles0513875_> a list of default repositories
[09:26] <ikonia> what do you mean default repositories ?
[09:26] <eagles0513875_> my server is using the DC's local mirrors im trying out an experiment
[09:26] <eagles0513875_> in terms of creating my own repo
[09:26] <ikonia> so you want the mirror list
[09:26] <ikonia> not repositories
[09:27] <eagles0513875_> when i say repositories i mean like main universe etc
[09:28] <ikonia> you have those in your install
[09:28] <ikonia> sorry, I'm not understanding what you actually want
[09:28] <eagles0513875_> sorry for the confusion i just figured it out :(
[09:30] <ikonia> did you mean something like this http://ubuntu.virginmedia.com/archive/dists/precise/
[09:30] <ikonia> ?
[09:35] <spidernik84> or this? http://repogen.simplylinux.ch/
[09:36] <spidernik84> is there anyone doing automated virtual machines creation?
[09:37] <eagles0513875_> ikonia: ya thats it
[09:38] <thierry> Hi everyone, i'm working on an ubuntu-server image on my Pandaboard - :) - and i'm trying to connect a GPS / RFID  / GPRS to my system; the problem is that in a normal ubuntu environment theses devices are attached on /dev/ttyUSB* ( for GPS and GPRS) and /dev/ttyACM* ( for RFID) but on my system they are plugged on /dev/usbdev*.**
[09:38] <thierry> normally it's a usbserial driver issues
[09:38] <thierry> anyone has an idea?
[09:40] <thierry> i found some tutos that show how to enable usbserial for devices using their idProduct / Vendor : but i need this to be natively done by the system , regardless the GPS i'm connecting
[09:41] <melmoth> thierry, sounds like something udev can help you with
[09:41] <melmoth> (if it s just a matter of giving the device the name you want)
[09:42] <thierry> melmoth:  it's not just the name
[09:43] <thierry> with cutecom when the device is plugged on /dev/ttyISB i can open it , but not if it is plugged on /dev/usbdevx.y
[09:43] <ogra_> sounds more like something for #ubuntu-kernel
[09:44] <Kingsy> can someone in here talk to me for a moment about how to make sense of the results sar produces?
[09:53] <thierry> :) ok 'ill try there
[10:03] <melmoth> Kingsy, apt-get install ksar :-)
[10:04] <Kingsy> melmoth: I have it installed, and I am looking at the results of the past 24 hours now. The thing I am asking is how to make sense of the results
[10:05] <melmoth> not sure i understand the problem
[10:05] <Kingsy> melmoth: for example, in "processes" there is a graph plotted with time / proc/s   <-- what exactly is proc/s ? it cant be processes I have a peak here at 87.5 proc/s
[10:05] <Kingsy> that cant mean 87 and a half processes
[10:05] <melmoth> sa record some metrics along time, then you can see what those metrics were (i/o , rm used, number of process) later on.
[10:06] <melmoth> ksar just show it with picture so it s easier to feel how things chnaged according to time
[10:06] <melmoth> number of process created by second i think
[10:07] <Kingsy> so at a point in the day it created 87.5 processes per second? isnt that impossible? the server should have crashed.. and according to the load average it didnt go above like 2
[10:07] <melmoth> http://linux.die.net/man/1/sar
[10:07] <Kingsy> melmoth: I guess what I am trying to say is, the results just don't make any sense
[10:08] <melmoth> you can create 87 new process per second.. sound slike a fork bomb.
[10:08] <Kingsy> what is a fork bomb?
[10:09] <melmoth> http://en.wikipedia.org/wiki/Fork_bomb
[10:09] <Kingsy> melmoth: and your right yeah, ot does mean processes created per second, so for arguments sake, I wanted to try and find out WHY it did that..
[10:09] <melmoth> either a malicious fork bomb, or, more likely, some daemon init script that failed to run the daemon, and try again, and again, and again
[10:10] <Kingsy> melmoth: so, say you were looking at sar, and you spotted that spike, what would you check next in order to find out more about why that happened?
[10:10] <melmoth> what does /var/log/syslog says for the date and time this happened ?
[10:11] <Kingsy> melmoth: that file doesnt exist on our server
[10:12] <jamespage> Daviey, really enjoying the iscsitarget rabbit hole - dkms module does not build with 3.5 kernel....
[10:12] <melmoth> you have probably change the syslog configuration so it logs things elsewhere..
[10:12] <jamespage> smb, are you around to pester for help with iscsitarget/linux 3.5 build issues?
[10:12] <Kingsy> yeah I guess the hosting company did it.. man
[10:13] <smb> jamespage, I would be again in a bit. You guys seem to sense whenever I am about to leave for lunch
[10:14] <smb> jamespage, ;)
[10:14] <jamespage> smb, lol - give me a ping when you are done then - thanks!
[10:14] <smb> jamespage, Sure, will do :)
[10:21] <Daviey> jamespage: I am jealous!  If you want to swap with something cooler, it's all yours.
[10:50] <spidernik84> Hi all. Is there any alternative to ubuntu-vm-builder?
[10:53] <xnox> spidernik84: python-vm-builder, debootstrap, ubuntu core, cloud images, lxc containers....
[10:54] <jamespage> spidernik84, the cloud image builds switched a while back - lemme see if I can tell you what they use now - I think its livebuilder
[11:00] <spidernik84> thanks guys :)
[11:30] <Sachin__> Is there any alternative for ACL?
[11:32] <Sachin__> Is there any alternative for Access Control Lists ?
[11:43] <Daviey> jamespage: does ceph need xml2?
[11:45] <jamespage> Daviey, rest-bench uses it - but that should remain in universe
[11:45] <Daviey> jamespage: can you make sure it's handled? ta
[11:45] <Daviey> (showing in c-m)
[11:46] <jamespage> Daviey, how do I make that happen?  radosgw was as well but someone promoted it to main - I was expecting it to remain in universe
[11:47] <Daviey> jamespage: Why was it promoted ?
[11:49] <jamespage> Daviey, no idea
[11:49] <jamespage> Daviey, can I tell who promoted it?
[11:49] <Daviey> jamespage: no easily.
[11:49] <Daviey> not*
[11:50] <jamespage> Daviey, MIR has been requested but I think there are to many blockers for quantal IMHO - bug 1017978
[11:53] <Daviey> jamespage: ask in -release, see if anyone knows.
[11:53] <smb> jamespage, So whats up with scsitarget? :)
[11:55] <jamespage> smb, the dkms module is not so happy with the 3.5 kernel
[11:55] <jamespage> smb, I've picked a patch from upstream but I had to rework some stuff myself - would you be OK to review?
[11:57] <smb> jamespage, Yeah that has happened a few times. Yes, I could do. Where would I find your things?
[11:58] <smb> jamespage, Which place did you pick things from? Just out of interest. The debian package or the svn repo?
[11:58] <jamespage> smb, both - I merge in the latest packaging from debian, picked a few patches from upstream to resolve this and other bugs and added one myself!
[11:58] <jamespage> I've tested it locally and it appears to work - but specifically this patch needs a review:
[11:59] <jamespage> http://paste.ubuntu.com/1185604/
[11:59] <jamespage> ioc_task_link changed in 3.5 to be a void function - so I had to rework its use
[11:59] <jamespage> I think I have it right...
[12:06] <smb> jamespage, Hm, not sure. Now you would only conditionally call ioc_taslk_link while before it was always done... Give me a min to think about that
[12:07] <jamespage> smb, if you look at that function 3.4 and 3.5 you will see the change in behaviour
[12:07] <smb> jamespage, right, I was about to do that
[12:07] <jamespage> it won't work without a valid ioc context - whereas before the function would just return NULL
[12:20] <smb> jamespage, Yes, it seems the null check moved away. It seems it used to have another fail point when the refcount was 0. So mostly I guess your version is ok (maybe one could make it more readable by having an else case and do the assignment in the true case). And we hope that current->io_context is only set with a ref-count and always NULL otherwise...
[12:21] <jamespage> smb, I guess the "another fail point when the refcount was 0" was the bit that worried me a little
[12:24] <koolhead16> my internet phewwwwwwwwww
[12:24] <smb> jamespage, It seems that may be a guarantee... At least from the commit that removed the check: "ioc_task_link() is used to share %current's ioc on clone.  If
[12:24] <smb>     %current->io_context is set, %current is guaranteed to have refcount
[12:24] <smb>     on the ioc and, thus, ioc_task_link() can't fail."
[12:25] <jamespage> smb, \o/
[12:25] <jamespage> all good then...
[12:28] <smb> jamespage, Yeah I think that there is hope. Well at least from their point of view. They talk of current in the fork context and iscsitarget code uses something else.
[12:32] <smb> jamespage, But I guess whenever a task has a pointer to an ioc set there should be a reference (or there is something badly wrong with the usage of refcounting or the pointer)
[12:32] <jamespage> smb, I think so
[12:52] <dassouki> is there a way from terminal, I can list the 10 largest folders on my system, or folders that have more than 5gb in them includnig sub folders?
[13:01] <zul> dassouki: yeah look at the man page for find
[13:18] <dhanasekaran> HI, Guys i want go live session with ubuntu-server cd
[13:18] <dhanasekaran> please guide me guys
[13:21] <RoyK> dhanasekaran: there's no ubuntu server live cd
[13:21] <RoyK> there's the desktop cd, and the server cd, the desktop can go live, the server cd can operate in two modes, install or recovery
[13:26] <TheLordOfTime> dhanasekaran:  if you want a "live" server environment, install "desktop" onto a USB thumbdrive, add server packages
[13:26] <TheLordOfTime> (I have a system here that is the GUI Desktop but has additional server packages, to act as a dev system
[13:26] <TheLordOfTime> )
[13:27] <dhanasekaran> TheLordOfTime, i want find mac address for my server, But My server have ubuntu-server cd How to find out mac address
[13:27] <dhanasekaran> TheLordOfTime, it's new machine no os
[13:27] <dhanasekaran> only have ubuntu-server cd
[13:28] <patdk-lap> ifconfig
[13:28] <dhanasekaran> i want deploy os remotely using cobbler tool using mac address
[13:28] <dhanasekaran> please guide me.
[13:28] <patdk-lap> boot the system, the pxe bios on the nic will tell you
[13:29] <patdk-lap> or, tcpdump, and watch for dhcp requests from it when you turn it on
[13:29] <TheLordOfTime> dhanasekaran:  you install ubuntu server and then find  the MAC address for the NIC
[13:30] <TheLordOfTime> dhanasekaran:  or, you take apart the server and look at the NIC
[13:30] <TheLordOfTime> it may be written right on it what its MAC is
[13:30] <TheLordOfTime> (otherwise, you boot to a Desktop live session, run ifconfig -a in the terminal, and find the MAC)
[13:30] <dhanasekaran> TheLordOfTime, server it's located remote place
[13:32] <dhanasekaran> it's any way find mac address using ubuntu-server cd
[13:33] <dhanasekaran> i can login into busybox shell but not getting mac address information
[13:33] <TheLordOfTime> dhanasekaran:  no because the ubuntu server CD is not a live cd
[13:33] <TheLordOfTime> if its a remote server, it should get an OS before you can truly get any real information
[13:36] <dhanasekaran> it's a newly bot machine, there is no remote support hand,only machine power on state, inside ubuntu-server DVD
[13:38] <cerberos> If http://pastie.org/4662177 is the output of cat /proc/cpuinfo should I be installing _i386.deb or _amd64.deb packages?
[13:38] <_ruben> dhanasekaran: do you have remote kvm access?
[13:38] <dhanasekaran> _ruben, yes
[13:38] <dhanasekaran> _ruben, How to d it? any help
[13:39] <Gargoyle> Hello.
[13:39] <_ruben> then just follow patdk-lap's remark .. just hit the key to have it boot from pxe, and it'll show you the mac address it's using for that, unless it has a really broken pxe rom
[13:39] <Gargoyle> Is there a quick way to get the installed list of packages from one server, into a format that can be used to install all the same packages on another server?
[13:40] <_ruben> Gargoyle: there is, but i don't recall it from top of my head. should easy to find using google
[13:40] <yeats> !clone
[13:40] <_ruben> right, that one :)
[13:41] <Gargoyle> ahh. I was looking into dpkg options! :)
[14:03] <tmartiro>  I have problem changing network cards' affinity . After changing the smp_affinity , the changes are not applied. There is also bonding interface configured on host. Ubuntu 12.04 LTS. kernel  3.2.0-29-generic
[14:28] <dhanasekaran> _ruben, thanks man. it's showing mac address
[14:28] <dhanasekaran> !automate
[14:29] <dhanasekaran> !cloning
[14:54] <xnox> interesting
[14:55] <xnox> there is a similar bug of 12.10 guest cirrus driver not playing along with qemu-kvm on 12.10 host.
[14:55] <xnox> and that was thought to be kernel bug in 12.10, because 12.04 guest works fine on qemu-kvm 12.10 host.
[15:00] <jamespage> hallyn, around? I'm seeing a weird issue with my iscsi testing setup with libvirt/kvm
[15:01] <hallyn> jamespage: i know 0 about iscsi, but shoot
[15:01] <jamespage> my test VM's instantly go to paused state when I try to start them
[15:01] <hallyn> jamespage: nohting in logs?  how about setting debug level to 1 in /etc/libvirt/libvirtd.conf?
[15:01] <jamespage> hallyn, well not quite instantly - I get a little output
[15:03] <soren> hallyn: Did you see my question here: https://bugs.launchpad.net/ubuntu/+source/libvirt/+bug/997978/comments/86
[15:04]  * soren notes the time and runs off
[15:05] <hallyn> soren: looking
[15:07] <jamespage> hallyn, OK - here is a bit more - http://paste.ubuntu.com/1185918/
[15:08] <jamespage> hallyn, looks like this bug 1045027
[15:08] <jamespage> hallyn, love it when I managed to answer my own question
[15:09] <jamespage> anyway that blocks me from testing iscsi root for beta-1
[15:10] <hallyn> soren: great, thanks (commentin gin bug)
[15:11] <hallyn> jamespage: ah, ok :)  i waspuzzling over Dave's bug submission a few mins ago :)
[15:13] <jamespage> hallyn, I can reproduce that at will if that helps
[15:13] <jamespage> I reverted my kernel and I'm good for testing
[15:15] <hallyn> jamespage: marked it critical prio.  might comment there that you're ready to test.  i'll leave it to kernel team unless they say otherwise
[15:41] <tash> strange, am in group www-data and /var/www is www-data:www-data with 777 and I can't overwrite files via FTP
[15:41] <tash> perms denied
[15:41] <tash> any ideas?
[15:42] <Gargoyle> tash: Don't use ftp!
[15:43] <tash> well, it's winscp, port 22
[15:43] <tash> so, sftp
[15:43] <tash> still, odd that I can't overwrite files isn't it?
[15:43] <Gargoyle> tash: if it really is 777, then it wouldn't matter what group you were in.
[15:43] <Gargoyle> Anything in the logs?
[15:44] <Gargoyle> What error msg do you get?
[15:44] <tash> agreed on the 777, checking logs
[15:44] <tash> this was just dumped on me
[15:44] <tash> still investigating
[15:45] <tash> nothing in /var/log/messages or /var/log/syslog, not sure which log to look at
[15:45] <shauno> are the files you're trying to overwrite also owned by www-data ?
[15:46] <Gargoyle> If it's a windows to linux transfer, could be worth checking that if its a case sensitivity issue (eg, the file is uploaded as Index.HTM, but the webserver is serving index.htm)
[15:46] <tash> shauno: yes
[15:47] <tash> Gargoyle: it is Windows -> Ubuntu Server, but never had issues before and use WinSCP all the time.
[15:47] <tash> could it be trying to modify perms after the upload and maybe that is causing the issue?
[15:48] <Gargoyle> could be. You actually need to be the Owner of the file to change it's permissions.
[15:48] <Gargoyle> so if they are owned by www-data, you'll need to make them owned by you:www-data I think.
[15:51] <tash> yep, that was it
[15:51] <Gargoyle> annoying as hell! :)
[15:52] <jamespage> lynxman, OK - I give up - how do I get ipxe to tell me exactly why it won't boot off my iscsi target
[15:52] <jamespage> ubuntu installed to it just fine....
[15:54] <lynxman> jamespage: hm... :)
[15:54] <tash> so, even though perms are 777 on something you can't change perms unless you own the file? That is the lesson learned here right?
[15:54] <lynxman> jamespage: afaik it should relate to the network card you have on the server
[15:55] <lynxman> jamespage: there's quite a bunch of small bugs about this, we found one with TREllis 6 months ago
[15:55] <jamespage> lynxman, I get an Input/Output error message only
[15:56] <jamespage> it worked at alpha-3 frustratingly
[15:56] <lynxman> jamespage: anywhere I can see a screenshot of those?
[15:56] <jamespage> lynxman, not right this moment but it pointed me to http://ipxe.org/err/1d704439
[15:58] <lynxman> jamespage: I'd say the driver got a reply it didn't expect
[15:58] <lynxman> jamespage: although ipxe is mostly black magic, I'd contact the mailing list and hope for the best :)
[16:01] <jamespage> lynxman, someone beat me to it...
[16:01] <jamespage> http://ipxe.org/err/1d704439
[16:01] <jamespage> http://lists.ipxe.org/pipermail/ipxe-devel/2012-September/001796.html
[16:01] <jamespage> even
[16:01] <jamespage> :-)
[16:02] <lynxman> jamespage: so that's a high chance, I don't think our last ipxe update is from before May 16th
[16:02] <jamespage> lynxman, feb
[16:03] <lynxman> jamespage: yeah, so that'd make sense, although *crossing fingers* the package should be easy to update, and if not I'll be willing *steps forward* to update it
[16:04]  * jamespage tries it
[16:37] <roaksoax> Daviey: are we too late to upload maas to quantal?
[16:43] <Daviey> roaksoax: no, i'd rather it was in there.
[16:43] <Daviey> roaksoax: can i see the diff?
[16:43] <roaksoax> Daviey: ok, I'll upload later today
[16:44] <roaksoax> Daviey: ok, it will be a veeeeeeeeeeeeery long one :)
[16:46] <roaksoax> Daviey: http://paste.ubuntu.com/1186109/
[16:56] <Daviey> roaksoax: best hold off, actually.
[16:57] <roaksoax> Daviey: ok
[16:57] <roaksoax> Daviey: it will keep growing though
[16:57] <roaksoax> Daviey: right now it has been tested and has everything working but internal DNS/DHCP
[17:11] <Daviey> roaksoax: can you comment on what the diff does?
[17:11] <Daviey> how does it improve the beta?
[17:11] <Daviey> ?
[17:20] <roaksoax> Daviey: will do
[17:20] <roaksoax> Daviey: i didn't upload anything on friday nor yesterday
[17:20] <roaksoax> as some issue needed to be fixed
[17:20] <roaksoax> ii'm just updating to latest branch and testing eeeeeverything again
[18:00] <nandersson> Hi, I am setting up Ubuntu Core in a disk image. Got it to work, but now I would like to have my keyboard working. I installed the package keyboard-configuration, choose my keyboard and thought that would do it, but no. Still english keyboard. I have the right parameters in /etc/default/keyboard. Do I need to set some vars in .bashrc /etc/.profile or someplace?
[18:26] <nandersson> Ok, I solved part of the problem by installing console-setup and configure it for UTF-8 and Latin1. Still don't get åäö though :-(
[18:36] <SuperLag> How can you find out what the dependencies are for $PACKAGE? is there an apt option that will do that?
[18:40] <Nephelo> Hello, I've got a problem with my Ubuntu 10.04 Serversystem. I'm using mdadm (raid 5) and smbd for network shares. After a time the server freezes, only a reset helps. RAM is ok, and FS seems to. I tried to create an archive and this works. Nothing in syslog. Can someone help me?
[18:48] <RoyK> Nephelo: cat /proc/mdstat
[18:48] <RoyK> Nephelo: and use smartctl -H on all devs in that raid
[18:49] <RoyK> Nephelo: and perahsp smartctl -t short for them
[18:49] <Nephelo> RoyK The raid seems ok. But after a reset the raid starts with resync
[18:49] <Nephelo> Ok, I'll try smartctl
[18:49] <RoyK> Nephelo: if a drive hangs, it may hang the whole system
[18:50] <Nephelo> RoyK: But I tried to create a big archiv and this works
[18:50] <Nephelo> (using the comand line)
[18:56] <Daviey> roaksoax: Hey, how are you doing with using the squashfs with MAAS?
[18:56] <roaksoax> Daviey: well haven't really looked at it again yet as I was resolving issues with MAAS itself
[18:57] <roaksoax>  Daviey going to start looking at it again once I finish this package testing
[18:58] <Nephelo> RoyK: smartctl -H says for all devices that they passed the test
[18:58] <adam_g> zul: do you know whats up with the python-quantumclient MIR? hasn't moved in a while but looks like the concerns have been addressed?
[18:59] <zul> adam_g: no idea....lemme have a look
[18:59] <Daviey> roaksoax: ok
[19:01] <adam_g> zul: nosexcover is gone and the test suite is enabled. its now a dependency of nova and is blocking nova uploads to Q, which is why i ask.
[19:01] <zul> yeah lemme bug mterry
[19:02] <hallyn> pmatulis: hi, are there freeze dates relating to the server guide for quantal?
[19:02] <roaksoax> Daviey: do yo want me to attach a diff?
[19:02] <roaksoax> Daviey: to the FFe bug report?
[19:02] <roaksoax> Daviey: cause I think I'm ready to upload
[19:03] <pmatulis> hallyn: hi, i'm lagging behind this cycle for the guide.  i need to look into it
[19:05] <Daviey> roaksoax: jfdi please
[19:06] <Daviey> roaksoax: but please be around if it regresses to resolve
[19:06] <roaksoax> Daviey: k cool :)
[19:08] <hallyn> pmatulis: ok - there are several things in the lxc guide i need to update, but i'm hoping i can put that off a bit further as i chase some elusive bugs elsewhere :)
[19:10] <roaksoax> Daviey: ok cool. The only "regression" is that dhcp seems to not be writing the config just yet, which leaves it unsuable for now, but that should be addressed very soon
[19:10] <roaksoax> Daviey: due to the running our own maas dhcp upstart job
[19:12] <Daviey> roaksoax: Remember that maas doesn't always run it's own dhcp
[19:12] <Daviey> Many scenarios it will not.
[19:12] <roaksoax> Daviey: yeah I'm much rather concerned on having maas fully supported now with correctly running external dns/dhcp rather than increasing the diff later in the release
[19:14] <skrite> hey all
[19:20] <roaksoax> Daviey: uploaded \o/
[19:29] <Troy^> System restart required how often do those types of updates happen in server?
[19:31] <roaksoax> Daviey: let me know when you approve maas please
[19:40] <Daviey> roaksoax: done
[19:42] <roaksoax> Daviey: thanks
[19:54] <protoCall7> Hi all, I am attempting to install a Xen domU of Ubuntu Server 12.04 following the directions on the xen wiki @ http://wiki.xen.org/wiki/Ubuntu_10.04_domU, but I can not seem to get a boot loader to install.  Does anyone have any troubleshooting tips for tracking the problem down?
[20:12] <protoCall7> here are the relevant syslog entries regarding the grub failure: http://paste.ubuntu.com/1186449/
[20:16] <smoser> hallyn, https://code.launchpad.net/~smoser/ubuntu/quantal/lxc/lp-1045955/+merge/122749
[20:30] <hallyn> smoser: all the whitespace changes mask the meat of it
[20:31] <hallyn> you're adding check for auth_key (not mentioned in changelog).  but i have no objection to what it claims to change.  I assume utlemming is happy with it too?
[20:31] <hallyn> smoser: I'm adding some compile fixes right now too.  I'll copy in your .patch file if that's ok
[20:31] <hallyn> (subject to beta freeze of course)
[20:33] <arrrghhh> hello.  i'd like to run logrotate more frequently than 'daily' - at least on one specific set of logs.
[20:33] <arrrghhh> can i just put */10 * * * * logrotate in my root's crontab, or is there more to it?
[20:34] <smoser> hallyn, the comment about auth_key is correctly in commit messages
[20:34] <smoser> and i really suggest taking the wuite space changes too
[20:34] <smoser> as that file is a mess.
[20:35] <hallyn> smoser: pushed to ubuntu:lxc
[20:35] <hallyn> smoser: yeah, that file is a mess, cause ppl insist on fighting 'et sw=8 ts=8' common-sense setting :)
[20:35] <smoser> by just copying rather than merging you lose my per-commit commit messages.
[20:35] <hallyn> (j/k on whitespace)
[20:36] <smoser> hallyn, if i made invalid or inconsistent white space changes then you should reject it.
[20:36] <smoser> but you really should set some '# vi:'
[20:36] <hallyn> does all vi actually honor that?
[20:36] <smoser> as that makes the one true editor "just work"
[20:37] <hallyn> ISTM it generally ignores it
[20:37] <smoser> set modeline modelines=3
[20:37] <smoser> that searches top 3 and bottom the lines
[20:38] <hallyn> so it only works when people want it to - in which case they can correctly set it by hand anyway
[20:38] <hallyn> maybe those should be default in system-wide vimrc
[20:38] <hallyn> but anyway
[20:39] <hallyn> bzr merging with multiple quilt patches is not to be trusted.  sorry.  your commit msgs were lost.
[20:39] <roaksoax> Daviey: ok so I tried to do the squashfs thing and a few thoughts:
[20:39] <roaksoax> 1. Installer is being killed due to an apparent lack of memory
[20:39] <hallyn> I always keep upstream-directed msgs in DEP-5 tags, and distro-directed in changelog
[20:40] <roaksoax> 2. squid-deb-proxy doesn't allow to download the image, so I had to tweak squid0deb-proxy. For our case, I think it might be important to publish it in archive.u.c and download it and find a workaround
[20:40] <smoser> hallyn, but if you're a developer, you want it to. and so you should set that.
[20:40] <hallyn> in fact i think i'll copy yours into header
[20:40] <smoser> and then go on with life.
[20:41] <smoser> the reason its not default is
[20:41] <smoser> http://www.vim.org/scripts/script.php?script_id=1876
[20:41] <smoser> there are been issues at points in the past.
[20:41] <smoser> i think it might be saner now.
[20:42] <hallyn> smoser: are you forwarding that patch to lxc-devel or should i?
[20:42] <smoser> http://lwn.net/Articles/20249/
[20:42] <smoser> umm... you should :)
[20:42] <hallyn> ok
[20:42] <smoser> oh. and sorry i didn't do good patch headers there.
[20:43] <smoser> it would hvae been ok for you to NAk and say that.
[20:51] <smoser> hallyn,
[20:51] <smoser> have you seen this
[20:51] <smoser> with brfs
[20:51] <smoser> $ sudo lxc-destroy --name precise-amd64
[20:51] <smoser> rm: skipping `/var/lib/lxc/precise-amd64/rootfs', since it's on a different device
[20:51] <smoser> result is tha ti can't kill that root
[20:51] <smoser> i have to run now. will check in later.
[20:51] <smoser> i thought i had set up for btrfs
[20:52] <hallyn> smoser: hm.
[20:52] <hallyn> smoser: someone changed the rm -rf to                 rm -rf --one-file-system --preserve-root $rootdev
[20:53] <hallyn> git suggests it was me
[20:53] <hallyn> bastard
[20:54] <hallyn> we want that safety though, so i think the right answer is to first separately rm -rf --one-filesystem the $rootfs, then the $containerdir
[20:54] <hallyn> i'll push a fix
[20:55] <esuave> what is the difference between the /etc/rc.d/init.d/ directory and the /etc/init.d/?
[20:56] <arrrghhh> esuave, well as i understand it the rc.d stuff is by runlevel
[20:56] <arrrghhh> init.d is all the scripts for startup, regardless of runlevel
[20:56] <esuave> ahh got ya! thanks!!
[20:57] <hallyn> smoser: oh.  no.  that's already fixed in quantal.  but not in precise
[20:57] <hallyn> if you feel that should be SRU'd, please do open a bug
[20:57] <arrrghhh> esuave, that wasn't a great explanation.  this one is much better :)
[20:57] <arrrghhh> http://askubuntu.com/questions/5039/what-is-the-difference-between-etc-init-and-etc-init-d
[20:59] <arrrghhh> as my explanation really applies more to the rcX.d stuff
[20:59] <esuave> ah yes this helps
[20:59] <esuave> thank you
[20:59] <arrrghhh> np
[21:15] <Tellmarch> Hi... is there a way to allow a user to take ownership of folders/files owned by www-data?
[21:15] <arrrghhh> Tellmarch, if you can change the ownership, chown
[21:15] <arrrghhh> if you have to change the permissions without ownership, chmod
[21:16] <Tellmarch> my problem is that there is a user, let's say user1, with a website in his home folder
[21:16] <Tellmarch> however, php scripts create files there, owned by www-data
[21:16] <Tellmarch> user1 is not an administrative account, so normally he can't do the chown (no rights to do so)
[21:16] <arrrghhh> Tellmarch, chmod the files then
[21:17] <Tellmarch> but i'd like to allow him to, without giving him full root access to the server
[21:17] <arrrghhh> so more users/groups can access them
[21:17] <arrrghhh> just chmod the files he needs to access
[21:17] <Tellmarch> files might be created by php in the future, i want him to be able to do it
[21:17] <Tellmarch> or can i change the default permissions for files created by apache?
[21:17] <arrrghhh> cron job?
[21:17] <arrrghhh> hrm, you might be able to
[21:18] <arrrghhh> above my pay grade... i guess it really depends on how the files are created and for what purpose.... but again, i am pretty green with apache.
[21:19] <Tellmarch> basically he has a (heavy) php web application, which can install new modules directly from the web administration page etc.
[21:19] <Tellmarch> which means new files created, owned by www-data instead of user1...
[21:20] <arrrghhh> ah
[21:20] <arrrghhh> yea, i'd say you'll have to ensure the web app is creating the files with the appropriate permissions
[21:20] <arrrghhh> or create a cron job that chmod's the files - not exactly pretty.
[21:21] <Tellmarch> yeah that would be easy enough to do... but if for some security reason he wants some strange permissions on few files, might be dangerous
[21:22] <Tellmarch> though doing chown -R on the root of the website should be fine I guess
[21:22] <Tellmarch> ok, thanks for the hints
[21:22] <arrrghhh> np
[21:22] <arrrghhh> yea, if you limit the chmod to his website directory, it shouldn't be too dangerous
[21:23] <arrrghhh> just dangerous for his site potentially ;)
[21:27] <Tellmarch> one other idea I had was to allow user1 to run chown as www-data, in the sudoers file... But i don't know much about sudoers, so i'm not even sure if it's possible
[21:27] <Tellmarch> I think if I can do that it would be safe, nothing else should be owned by www-data in the server...
[21:28] <arrrghhh> Tellmarch, why not just add his user to the www-data group?
[21:28] <Tellmarch> because the default permissions are 755 :(
[21:28] <Tellmarch> it was my first idea too ^^
[21:28] <arrrghhh> ah
[21:28] <arrrghhh> heh
[21:28] <arrrghhh> sounds so simple :P
[21:59] <kyle__> Where do you set the system-wide defaults ala gsettings (previously gconf) in 12.04?
[21:59] <arrrghhh> kyle__, doesn't really apply to the -server platform - does it?
[22:00] <kyle__> arrrghhh: No, but mostly it only applies to people running multiple workstations at work, and they tend to hang out more in server.
[22:00] <arrrghhh> kyle__, ok...
[22:00] <arrrghhh> just checking.  my server has no gnome ;)
[22:01] <kyle__> arrrghhh: Technically, _technically_ I'm dealing with server, as I installed the minimal, then apt-get installed unity-desktop ;)
[22:01] <arrrghhh> i'm sorry.
[22:02] <kyle__> arrrghhh: No better way of handling cloud-desktops run off of 'buntu.  Using vncserver is _much_ more efficient than running vino or whatever with X running on a virtual video card provided by KVM.
[22:02] <arrrghhh> i just ssh, but i'm only managing a couple of servers.
[22:02] <kyle__> arrrghhh: I've got one for nearly every student in the department.
[22:02] <arrrghhh> heh
[22:03] <arrrghhh> my ubuntu usage is a bit more limited than that...
[22:04] <kyle__> arrrghhh: Most of them will use the VM once, and forget about it.  But it will be required for some of the classes.  And I'm hoping those students go away feeling like this is a great system.  Hoping.
[22:04] <kyle__> s/will be/is/
[22:04] <Gargoyle> How much overhead is involved if I install gdb on a server? I've tried cloning the exact same config onto another machine, but I can't get apache to segfault on the test machine.
[22:04] <arrrghhh> hah
[22:05] <Gargoyle> Or can anyone suggest any other method for tracking down a segfault without installing all the gdb stuff on a live server?
[22:07] <Tellmarch> well just installing gbd doesn't change much i think, but you'd have to run apache in gdb i guess?
[22:08] <Tellmarch> ... haven't used gdb since my C programming days, about 10 years ago...
[22:09] <Gargoyle> Tellmarch: I've never used it. But from googling up some docs, I was wondering if I needed to install dev libraries that contain all the debug symbols?
[22:09] <Tellmarch> well I don't think debugging apache on your live server is really what you want to do
[22:10] <Tellmarch> yes, you'd need apache with the debugging symbols too...
[22:10] <Tellmarch> apache logs don't tell you anything of value?
[22:10] <Gargoyle> [notice] child pid 12698 exit signal Segmentation fault (11)
[22:11] <Gargoyle> That's all!
[22:11] <Gargoyle> Sometimes apache carries on, othertimes it locks up and the load balancer deletes it from the pool!
[22:11] <Tellmarch> not very informative :(
[22:12] <Gargoyle> Can I get apache to be a bit more informative? or dump a core or something that might be of use?
[22:13] <Tellmarch> I'm not sure...
[22:13] <Tellmarch> but running apache in gdb means compiling it with the debugging symbols
[22:13] <Tellmarch> quite a bit of work
[22:14] <Tellmarch> CoreDumpDirectory /tmp/apache
[22:14] <Tellmarch> there seems to be an option to get the core dump
[22:15] <Gargoyle> Yeah, at which point the apache you are debugging is no longer the same as the running one!
[22:16] <Gargoyle> I'll see if I can get get something from a core dump. Or perhaps I can get the PID output into access.log too, so at least I can try and track down which part of the web app is crashing it?
[22:16] <Tellmarch> is it a php web app?
[22:17] <Gargoyle> yup
[22:17] <Tellmarch> some people recommend xdebug...
[22:18] <Gargoyle> Ah. Maybe I can get something from that.
[22:19] <Tellmarch> though it's better to use it on a dev environment, not production...
[22:20] <Tellmarch> and if you can't reproduce, that won't help '^^
[22:20] <soren> adam_g: Do you have any idea why all of glance's dependencies on its own binaries are so strictly versioned?
[22:20] <soren> adam_g: It has well and thoroughly painted itself into a corner on my laptop due to this.
[22:32] <arrrghhh> so anyone manage disk space using a web interface?
[22:32] <arrrghhh> i was looking at dfweb, but i can't seem to get it working.
[22:38] <adam_g> soren: looking thru history for the what/when/why now, but i can certainly see issues with attempting to use, say, the wrong python-glance with the server(s)
[22:40] <adam_g> looks like its been that way since the beginning?  what are you running into exactly?
[22:41] <adam_g> soren: ^
[23:51] <zul> adam_g: did you get quantumclient seeded?