[00:38] <rnigam> Hello Everyone please look at this thread and help me out: http://ubuntuforums.org/showthread.php?p=10542465#post10542465
[00:52] <hallyn> kim0: packages built in ppa
[00:53]  * hallyn out
[00:54] <thesheff17_> rnigam: I just ran your vmbuilder command on my machine...it worked with no problems....it looks like it is isolated to maverick.
[00:55] <thesheff17_> I would suggest putting the debug level really high for libvirt or using lucid.
[00:55] <thesheff17_> I can't find anything on the web that has that same error.
[01:00] <rnigam> tthanks thesheff17
[01:03] <rnigam> thesheff17: could you do a virsh --version on your machine and tell me which version of virsh and libvirtd is running
[01:03] <thesheff17_> rnigam: I built maverick w/ lucid as the underlying os....so strange that doesn't work for you on 10.10  my version is: 0.7.5
[01:06] <rnigam> thsheff17: I see. how can i change the debug level? and how will that help?
[01:07] <thesheff17_> rnigam: first check out /var/log/libvirt/qemu/vmName.log and see if there is anything else in there that will give you more info about virsh start not working
[01:08] <thesheff17_> rnigam: /etc/libvirt/libvirtd.conf contains the log level stuff.
[01:09] <thesheff17_> rnigam: I'm just shocked that no one else has that error...usually google gives me the answer pretty quickly...maybe the debug level will give you/us more insight into what is wrong.
[01:14] <rnigam> thesheff17: qemu: could not open disk image /root/vmimages/mav-vm2/ubuntu-kvm/tmpBYvUD4.qcow2: Permission denied
[01:15] <rnigam> this is from the logs
[01:15] <thesheff17_> rnigam: can you paste your xml
[01:15] <thesheff17_> pastebin the xml from /etc/libvirt/qemu/vmName.xml
[01:20] <thesheff17_> rnigam: are you using an SELinux kernel?...just looking on google for some fixes.. what does ls -la   /root/vmimages/mav-vm2/ubuntu-kvm/tmpBYvUD4.qcow2 say...is that file even there?
[01:21] <rnigam> thesheff17: heres the xml :  http://paste.ubuntu.com/578124/
[01:23] <thesheff17_> rnigam: looks good.... you are using virsh start as root correct?
[01:23] <rnigam> yup
[01:24] <rnigam> and yes thatmfile is present there.
[01:25] <rnigam> actually that file gets created when i run the vmbuilder command
[01:25] <rnigam> the qcow2 disk file
[01:25] <thesheff17_> rnigam: right I just wanted to make sure it is getting created correctly...my permissions on my files is -rw-r--r-- 1 root root 1238958080 2011-03-05 13:30 /mnt/vm/vm021/ubuntu-kvm/tmpqyXQXh.qcow2
[01:26] <thesheff17_> does yours match that?
[01:28] <rnigam> -rw-r--r-- 1 root root 627376128 2011-03-09 14:20 tmpBYvUD4.qcow2
[01:29] <thesheff17_> /etc/libvirt/qemu.conf contains username to run the service as...can you check that this is root.
[01:31] <rnigam> you mean the userid variable?
[01:32] <thesheff17_> I have user = "root" and group = "root"
[01:32] <rnigam> everything in that file is commented out and hence default values are being used. I am not sure what defaults are. LEt me make those changes in mine and try again ! it might take some time to get back.
[01:33] <rnigam> are there any other flags in that file turned on for you?
[01:33] <thesheff17_> rnigam: yea yours might be a little...since mine in lucid.
[01:33] <thesheff17_> different
[01:33] <rnigam> other than user and group?
[01:34] <thesheff17_> rnigam: that is the only thing that isn't commented out in that file...you aren't running SELinux kernel are you?
[01:35] <rnigam> nope
[01:35] <rnigam> no se linux
[01:37] <thesheff17_> rnigam: ok...not sure what it could be...permission denied should not happen if you running it as root.
[01:37] <thesheff17_> rnigam: to just test try just doing chown 777 filename and see if it will start.
[01:38] <rnigam> ya i am recreating a vm again to check if anything changes after changing those options !
[01:43] <rnigam> thesheff17: In case i forget, thanks so much for all the help.
[01:44] <thesheff17_> rnigam: np....I know how hard it is to find help for virsh/kvm sometimes.
[01:44] <thesheff17_> rnigam: not that I have helped much :)
[02:00] <rnigam> thesheff17- Great news.. you helped me enough
[02:00] <rnigam> it works
[02:01] <rnigam> turns out those two variables needed to be set in /etc/libvirt/qemu.conf
[02:01] <thesheff17_> rnigam: excellent
[02:02] <thesheff17_> rnigam: I'm on vacation next week but if that is the default problem you/I should file a bug for it.
[02:02] <rnigam> man... I thought this was the easiest part. Yes we must file a bug
[02:03] <rnigam> theres another bug that we need to file... or rather a request for a patch
[02:04] <rnigam> when we undefine an image... it doesnt remove the qcow2 files and doesnt free other resources.
[02:04] <rnigam> it would be great if we can achieve that with undefine rather than doing manually !
[02:05] <rnigam> thesheff17: what do you think about the request? please comment !
[02:12] <thesheff17_> rnigam: I would love that as well
[02:12] <thesheff17_> rnigam: my python script does it for me...basically just have a huge wrapper around vmbuilder & virsh.
[02:13] <rnigam> ya
[02:27] <zul> smoser hallyn: the lxcguest scripts work perfectly on the uec images
[03:44] <tc2k11> could anybody help me a with postfix issue?
[03:50] <tc2k11> i'm having issues send email to a particular domain where it always times outs when there is an attachment
[03:57] <Roxyhart0> somebody know how to upgrade samba in ubuntu? i tried with apt-get upgrade samba but is doesn work
[04:01] <EvilPhoenix> Roxyhart0:  define "upgrade" because the repositories are not always cutting-edge up to date
[04:02] <Roxyhart0> i have samba 3.4.5 and I want to upgrade to 3.5.x
[04:02] <Roxyhart0> as I am having problems with windows 7 clients and aparently samba 3.5.x solve it
[04:03] <EvilPhoenix> which version of ubuntu server?
[04:03] <EvilPhoenix> 10.04? 10.10?
[04:03] <EvilPhoenix> OLDER?
[04:03] <Roxyhart0> not 9.10
[04:04] <EvilPhoenix> didnt say 9.10
[04:04] <EvilPhoenix> i said 10.10
[04:04] <Roxyhart0> but i read the version 10.04 still come with samba 3.4.5
[04:04] <EvilPhoenix> 10.04 has samba 3.4.7 on it from what I can tell
[04:04] <EvilPhoenix> lemme check my 10.10 system
[04:05] <EvilPhoenix> 10.10 has samba 3.5.4 on it
[04:05] <EvilPhoenix> is upgrading your distribution an option?
[04:06] <Roxyhart0> maybe, but is could be risky as I am running file server here...
[04:07] <Roxyhart0> also I am not sure if 10.10 has so much bugs?
[04:08] <EvilPhoenix> I cant vouch for it as i stick with LTS releases (and never use samba)
[04:08] <EvilPhoenix> although FYI, you should always keep backups even when you arent upgrading your system
[04:08] <patdk-lap> I guess samba isn't in backports?
[04:09] <EvilPhoenix> patdk-lap:  didnt see it in there
[04:10] <EvilPhoenix> and out of curiosity... patdk-lap, do you support the use of backports updates in a production system?
[04:10] <EvilPhoenix> s/system/environment/
[04:10] <patdk-lap> roxyart0, you do know you have like only a month left before 9.10 is dead?
[04:10] <EvilPhoenix> oh that too
[04:10] <patdk-lap> EvilPhoenix, depends
[04:10] <patdk-lap> some things yes, most things no
[04:11] <patdk-lap> but for something like that, I would just download samba from backports and install
[04:11] <EvilPhoenix> Roxyhart0:  you should probably consider upgrading anyways, what're you using right now, 9.10?  you should definitely upgrade the distro (MAKE BACKUPS FIRST) because its going to die very soon (end of life)
[04:12] <Roxyhart0> yes i need to do...but what is the version supported, is not 10.04?
[04:12] <patdk-lap> 10.04 will be supported for 4more years
[04:12] <EvilPhoenix> the version that has 3.5.x samba is 10.10
[04:12] <patdk-lap> 10.10 will be for another year
[04:12] <EvilPhoenix> its also supported (and recent, as well as stable from what I can tell)
[04:13] <patdk-lap> most of my samba servers are 10.04
[04:13] <Roxyhart0> well I will try to update at least to 10.04
[04:13] <patdk-lap> one is still 8.04
[04:13] <EvilPhoenix> Roxyhart0:  FYI, 3.4.7 is whats in the 10.04 repositories
[04:13] <Roxyhart0> but i know there are someway to install samba 3.5.x in 10.04 ...but i dont know the procedure
[04:13] <EvilPhoenix> Roxyhart0:  to get 3.5.x, you'll need 10.10, as I dont see 3.5.x in the repos (including backports)
[04:13] <patdk-lap> what about 3.4 doesn't support win7?
[04:13] <EvilPhoenix> Roxyhart0:  possibly compiling from SOURCE
[04:13] <EvilPhoenix> which is a pita
[04:13] <patdk-lap> I haven't had any issues with windows 7 and my samba's
[04:14] <Roxyhart0> i didn time ago samba 3.4.5 for ubuntu 9.04
[04:14] <EvilPhoenix> Roxyhart0:  but i'll betcha patdk-lap knows more about samba than i do, you should talk to him
[04:14] <Roxyhart0> but i dont remeber how
[04:15] <Roxyhart0> patdk, what version of samba and ubuntu do you have?
[04:15] <patdk-lap> Roxyhart0, ask something I didn't just answer
[04:19] <Roxyhart0> you said about your ubuntu verison, but what is the match with your samba versions and are all of them working with windows clients for printing? as my problems is sharing printers
[04:19] <patdk-lap> my samba version is whatever is in ubuntu 10.04
[04:19] <patdk-lap> and I haven't had an issue printing
[04:20] <patdk-lap> the wife and me both used a printer via samba in win7
[04:21] <patdk-lap> at work, I use a win2003 machine for a print server
[05:15] <Roxyhart0> somebody know how to install a patch file?
[07:57] <_Neytiri_> i am having a issue settign up a iscsi target on my ubuntu server, i dont know where i went wrong but i cant get my esx host to find the server
[07:57] <_Neytiri_> http://www.howtoforge.com/using-iscsi-on-ubuntu-9.04-initiator-and-target i used that tutorial to setup the target
[08:57] <jibel> jamespage, hi
[08:57] <jamespage> jibel: good morning
[08:58] <jibel> jamespage, i'll push an update of usit for desktop testing. I set the version number to ppa15 in the changelog ?
[08:58] <jibel> actually, 1.0-1~ppa15
[08:59] <jamespage> jibel: thats good
[08:59] <jibel> I'm asking this because last time you updated it but I don't know why
[09:11] <jamespage> I updated because for some reason that I can't remember your merge proposal version conflicted with the latest release in the PPA
[09:11] <jamespage> ~ppa15 is OK
[09:11] <jamespage> (~ppa14 is currently published)
[10:00] <lnx-usr> Any kvm virtualisation guru's in here?
[10:01] <Jeeves_> Guru is a big word :)
[10:02] <lnx-usr> only 4 letters.....
[10:02] <lnx-usr> Anyway, I've been wanting to give spice a try
[10:02] <lnx-usr> Is there an easy way of doing this on ubuntu
[10:03] <twb> I'm not sure simulation of electrical circuitry is what people normally mean by "virtualization"
[10:04] <lnx-usr> @twb SPICE is a "new" high end remote protocol designed to work with kvm virtual geust systems
[10:04] <lnx-usr> 2d acceleration and stuff like that
[10:05] <lnx-usr> It's really good, i've tested it about 2 years ago and it was working great then.
[10:05] <lnx-usr> so I'm not talking about spice-ng
[10:05] <lnx-usr> ;-)
[10:05] <twb> "2d acceleration"?  As in, a GUI?
[10:05] <lnx-usr> http://spice-space.org/
[10:06] <lnx-usr> Hardware assisted 2d playback
[10:06] <twb> Lame
[10:06] <twb> VMs shouldn't have a GUI any more than the dom0 should
[10:06] <lnx-usr> twb are you really going to asume you know every use case?
[10:07] <twb> Only the important ones
[10:07] <lnx-usr> SPICE is mostly designed to host virtual desktops
[10:07] <twb> desktops are lame, too
[10:07] <lnx-usr> High performance 2d graphics using a remote protocol
[10:07] <soren> *chuckle*
[10:08] <twb> Dunno what's wrong with NX-compressed X11
[10:08] <lnx-usr> A lot
[10:08] <lnx-usr> It's an "in guest" protocol
[10:08] <lnx-usr> that means it's a service running inside the guest
[10:09] <twb> So what you're saying is that your VMs are not only GUIs, but they're not unix GUIs
[10:09] <pnunn> lnx-usr: sounds intersting.. we use KVM here for servers, but are looking at pushing into desktops..
[10:09] <lnx-usr> No thats not what I'm saying, allthough it's true... ;-)
[10:10] <twb> Is bind9 still best practice for hosting ones domains?
[10:10] <twb> I'm going through the debtags list and stuff like maradns sounds interesting
[10:10] <lnx-usr> Anyway, I'm implementing Desktop virtualisation systems based on vmware view a lot
[10:11] <lnx-usr> I'd like to be able to give my customers the choice of using ubuntu, kvm with spice
[10:11] <twb> vmware can FOAD.  I wasted the best part of a week trying to make SCO4 work with it.
[10:11] <lnx-usr> I can't right now
[10:11] <lnx-usr> What vmware system?
[10:12] <pnunn> lnx-usr: be interested to hear how you go with this. Doesn't seem to be much in the way of software on the link you gave.
[10:12] <twb> Unforunately, it was vmware-server.  We weren't confindent esxi could be tricked into working with our existing backup infrastructure.
[10:12] <twb> s/backup/DR/
[10:12] <lnx-usr> Why would you use SCO  anyway
[10:13] <twb> lnx-usr: because the shitty ERP system was written for it twenty years ago
[10:13] <lnx-usr> VMware server is not the way to go
[10:13] <twb> It's like saying "why are banks still using AIX"
[10:13] <lnx-usr> Kind of like saying "why use desktops..."
[10:14] <twb> Yeah
[10:14] <lnx-usr> lol
[10:14] <twb> Customers should FOAD, too
[10:14] <lnx-usr> Anyway, what does your backup system look like
[10:14] <lnx-usr> end users should FOAD.....
[10:15] <twb> It's basically cron, ssh, lvcreate --snapshot and rsync
[10:15] <lnx-usr> Ah
[10:15] <lnx-usr> That wont work on vmware esxi
[10:15] <twb> Not without buggerizing around with undocumented vmware calls
[10:16] <lnx-usr> Where are the vmdk's stored?
[10:16] <lnx-usr> iscsi?
[10:16] <twb> in the dom0
[10:16] <twb> esxi is just a broken linux dom0
[10:16] <lnx-usr> Nope, esxi is not linux
[10:17] <lnx-usr> esxi has it's own kernel
[10:17] <twb> "Up through the current ESX version 4.1, a Linux kernel is started first,^[5] and is used to load a variety of specialized virtualization components, including VMware's 'vmkernel' component."
[10:17] <joschi> twb: esx != esxi
[10:17] <lnx-usr> Almost no linux userspace anymore
[10:17] <lnx-usr> As I understand you are not using central storage
[10:18] <twb> They only have ten users
[10:18] <twb> The server is the centralized storage
[10:18] <joschi> twb: esx is (was) a rhel based hypervisor. esxi doesn't run the linux kernel anymore
[10:18] <lnx-usr> It's like saying Netware is DOS because dos is booted first
[10:18] <twb> joschi: hum, I thought the only difference was esxi's dom0 didn't have a shitty GUI
[10:19] <lnx-usr> Only the console system was based on Linux
[10:19] <lnx-usr> This made it easier for vmware to implement userspace stuff
[10:19] <joschi> twb: no, they actually strapped the linux parts away
[10:19] <lnx-usr> Now most things moved to vmkernel space
[10:21] <lnx-usr> My advice would be this: Use vmware esxi, use SAN storage (even if it is something like Nexenta, ubuntu with iscsi target or freenas), mount the vmware vmfs filesystem via iscsi on your backup server and backup from there
[10:22] <twb> That sounds way too "enterprisey" to me
[10:22] <lnx-usr> ok
[10:23] <lnx-usr> then use esxi with the vma
[10:23] <twb> Where "enterprisey" means "expensive and crap"
[10:23] <lnx-usr> Or move to KVM, headless Virtualbox or Xen. But please don't use vmware server
[10:23] <twb> vmware-server doesn't work on 10.04 anyway
[10:23] <twb> I can't remember why, I think it was a problem with the NIC going spastic
[10:23] <lnx-usr> VMware server is the crappiest product vmware ever made
[10:24] <twb> I'd have used kvm but it makes sco koops
[10:24] <lnx-usr> I use ESXi, vsphere and workstation on a daily basis. And I mostly like them.
[10:24] <twb> Yeha but you also like GUIs
[10:25] <lnx-usr> Yeah, I also use KVM quite  a bit, never liked xen tough
[10:26] <twb> I like kvm and lxc best
[10:26] <twb> By 12.04 LXC will be kickass
[10:26] <lnx-usr> lol
[10:26] <twb> Maybe LCR will even be ready
[10:27] <lnx-usr> LCR?
[10:27] <lnx-usr> new to me
[10:27] <twb> Maybe it's CXR
[10:27] <lnx-usr> sorry, they must not have a gui yet
[10:27] <twb> Linux Checkpoint/Restart
[10:27] <twb> Allows you to suspend and resume arbitrary processes/process groups, maybe move them between hosts
[10:28] <lnx-usr> Must be a stupid project, I don't see screenshots..... ;-)
[10:29] <lnx-usr> Kind of like "live migration for applications"
[10:29] <twb> yeah
[10:29] <lnx-usr> Sounds awesomely cloudy googly like
[10:29] <twb> Wikipedia actually directed me to http://en.wikipedia.org/wiki/Application_checkpointing
[10:33] <twb> But for LXC, it would mean the ability to migrate domUs between hosts (and suspend them prior to restarting the dom0) without needing to emulate a full computer (a la kvm)
[10:33] <lnx-usr> less overhead is always better.
[10:34] <twb> LXC is about as efficient as a chroot
[10:34] <twb> But it's a lot leakier than kvm at the moment
[10:34] <twb> Needs more work in the kernel
[10:35] <lnx-usr> btw TWB, if I where you I'd look into using ESXi with the "Vmware vSphere 4.1 command line interface"
[10:35] <lnx-usr> It's a commandline interface you can install on a linux system and use it to script esxi
[10:36] <twb> Meh
[10:36] <twb> I'm not really interested in vmware until they start using copyleft licenses
[10:36] <lnx-usr> Ok
[10:37] <lnx-usr> Well I guest I'm going to have to build SPICE myself
[10:37] <twb> ckpt.wiki.kernel.org is what I was thinking of
[10:38] <lnx-usr> Yeah I found that to
[10:38] <lnx-usr> I'm going to eat now
[10:38] <lnx-usr> Nice talking to you
[10:51] <twb> LXC and Linux-CR are lightweight enough that I expect e.g. a smartphone to put each "app" in its own container, and to suspend/resume them individually.
[11:43] <sabgenton> llutz: do you think I need an equivalent to wireless-mode    managed
[11:43] <sabgenton> the guide doesn't mention it
[11:44] <sabgenton> opps wrong channel
[12:42] <kim0> hallyn: just tested the new ppa code. Here are the findings
[12:42] <kim0> - win7 seems to work well with vnc
[12:43] <kim0> - once installed qxl drivers, windows seemed to have become a bit unstable (crashed a couple of times)!
[12:43] <kim0> - I cannot connect using spicec at all "Warning: Invalid mode"
[12:44] <kim0> - I just found out about http://www.spice-space.org/download/gtk/ .. a gtk client, unsure how it compares to spicec!
[12:44] <kim0> hallyn: that's all, if that client is the official or better one, perhaps if you can build it, I'll try connecting with it as well
[12:55] <\sh> hmm...how do we attack the problem that authbind (which is used by tomcat6) still doesn't support ipv6? eventually this patch could help us here: http://toroid.org/ams/etc/authbind-ipv6-support
[13:01] <kim0> hallyn: someone on #kvm mentioned to check this patch http://www.mail-archive.com/kvm@vger.kernel.org/msg50244.html
[13:17] <hXm> hello
[13:18] <zul> hallyn: new lxc uploaded
[13:21] <hXm> isn't possible to execute files in a webdav server?
[13:27] <hallyn> zul: cool, thx.  have you tested it?  (IIUC it will break current configs)
[13:27] <zul> no i havent had a chance to
[13:28] <hallyn> kim0: we already have that patch.  without that, kvm locks up when using qxl
[13:31] <kim0> hallyn: ok good .. what about that new client thing
[13:31] <hallyn> dunno
[13:32] <hallyn> is that what people said to use?
[13:32] <kim0> well people are not too responsive
[13:32] <hallyn> ok.  i *was* going to build spice-qemu as a separate package today.  Do you think testing the new client is more urgent?
[13:33] <kim0> hallyn: I think so yeah, since currently spicec is not connecting at all
[13:33] <hallyn> ok
[13:33] <kim0> at least in my case
[13:33] <hallyn> got some unrelated testing to do, but I'll go retest my current packages right now, and then try out that new client.  bbl
[14:12] <kim0> hallyn: spice people confirm spice-gtk is the thing to use, spicec is legacy they say
[14:12] <eagles0513875> hey guys i wanna create a new user cuz the place that is hosting my vps seems to have enabled root :(
[14:12] <eagles0513875> how exactly should i add the user too the sudoers
[14:13] <eagles0513875> do i add it like roots setup User ALL=(ALL)  ALL ?
[14:20] <RoAkSoAx> morning all
[14:20] <zul> arrgh stop snowing
[14:20] <soren> zul: I'm not.
[14:20] <RoAkSoAx> Daviey: howdy! hey I resubmitted the eucalyptus-branch against the one in ubunut-virt
[14:20] <zul> soren: yes you are
[14:21] <eagles0513875> zinser: qhey guys question
[14:21] <eagles0513875> whoops
[14:21] <eagles0513875> guys how do i add a user to the sudoers file
[14:21] <Daviey> RoAkSoAx, rockin'
[14:21] <eagles0513875> do i add it like roots setup User ALL=(ALL)  ALL ?
[14:21] <eagles0513875> or just add the user to the admin group
[14:21] <Daviey> RoAkSoAx, We'll get that uploaded today.
[14:22] <RoAkSoAx> Daviey: awesome. thanks
[14:26] <smoser> RoAkSoAx, thank you for fixing that.
[14:27] <RoAkSoAx> smoser: glad to help ;)
[14:32] <hallyn> kim0: ok.  sorry, i got a bit side-tracked by email backlog.  off to create teh package soon.  (hopefully before i'm pulled away by calls)
[14:49] <kim0> hallyn: I tried building it locally ( but gave up) I just had to sed 's/celt051/celt/' in the configure script, if that's helpful to you
[14:59] <hallyn> kim0: that's actually troublesome.
[15:00] <hallyn> probably means there will be more changes to do, mirroring the ones made in spicec to handle higher version of celt
[15:03] <progre55> hi guys. is it possible to specify a system-wide alias for a hostname? just like you'd map an IP address in /etc/hosts, but I need to map a long domain name to a shorter alias, and afaik, I cannot do that in /etc/hosts
[15:04] <jkg> you probably want a 'search' line in /etc/resolv.conf
[15:05] <progre55> jkg: oh, let me try that, thanks
[15:08] <RoAkSoAx> zul: howdy!! when you have the time, tell me what you think about bug #731616
[15:08] <zul> RoAkSoAx: yeah i saw that last night ill merge it
[15:09] <RoAkSoAx> zul: ok awesome. I'm also looking at the yum thingy and we also need python-ethtool (though apparently someone's already working to get it in debian: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=549323)
[15:10] <evilsushi> anyone seen this error before ? http://pastie.org/private/igmpcadl4txerf2kpequa
[15:11] <RoAkSoAx> evilsushi: try sudo service apache2 restart
[15:11] <zul> RoAkSoAx: if you want to take care of it ill upload it for you
[15:12] <evilsushi> err i was in a deleted directory from a cron job
[15:13] <RoAkSoAx> zul: yeah I'm gonna look at it today
[15:13] <zul> RoAkSoAx: k thanks
[15:29] <smoser> hallyn, ping
[15:29] <smoser> your comment in bug 723361 doesn't make sense
[15:30] <smoser> you said "Information was provided as new bug 723361 (which I've marked as a dup of this)."
[15:30] <smoser> ah. never mind. the duplicate info is correct though. the comment is just wrong
[16:05] <jibel> SpamapS, good morning, is bug 732314 on your radar  ?
[16:06] <SpamapS> jibel: I saw it last night before passing out.. reading now
[16:07] <SpamapS> oh darn it
[16:07] <SpamapS> thats because squid doesn't use dh_installinit
[16:07] <SpamapS> jibel: fix forthcoming.. thanks for pointing it out
[16:07] <zul> squid is evil
[16:08] <SpamapS> lifeless: cover your ears! ;)
[16:09] <zul> well that packaging is evil
[16:10] <zul> kees: ping who do i nag about the lxc MIR?
[16:14] <jibel> SpamapS, thanks for fixing it
[16:18] <SpamapS> zul: yeah, it seems to try to maintain compatibility a little too hard.
[16:19] <jdstrand> hallyn: fyi, I marked 723361 as a dupe of 545795. It seems that the reporter doesn't understand how the driver works, but that the main problem is that the USB device doesn't work. 545795 is fixed in maverick but the reporter is using lucid. 545795 has a workaround for the usb issue
[16:35] <hallyn> jdstrand: cool, thanks for the info
[16:35] <hallyn> jdstrand: smoser: honestly I don't know what happened with that bug, or why I ended up marking it a dup of itself.  I swear there was nother bug# involved where the info actually sat :)
[16:36] <smoser> hallyn, there is.
[16:36] <smoser> and it is marked as a duplicate
[16:36] <smoser> (look at the duplicates)
[16:36] <hallyn> smoser: waht i mean is i don't know why i did that wrong the first time
[16:37] <smoser> ah
[16:37] <smoser> k
[16:44] <kees> zul: what's the bug # for the lxc MIR?
[16:44] <zul> kees: one sec
[16:44] <zul> kees: 727200
[16:45]  * kees checks
[16:47] <kees> zul: can you target it to natty? then it should show up in the MIR queue for mterry to assign.
[16:47] <zul> kees: sure
[16:55] <RFleming> Greetings and salutations.
[16:56] <RFleming> With processes (in ps auwx) showing as (SERVER), is there any way to find out what the command-line switches were?
[17:27] <RoAkSoAx> ScottK: howdy!! I have a python packaging question. I'm packaging a python-ethtool. During the build it shows that it is creating <BUILDROOT>/python-ethtool/usr/local/lib and during install W: dh_python2:95: Python 2.7 should install files in /usr/lib/python2.7/dist-packages/. Did you forget "--install-layout=deb"? http://pastebin.ubuntu.com/578435/ . Howver, the resulting binary obviosly has the correct paths. http://pastebin.ubuntu.com/578436/
[17:46] <Slyboots> Right..
[17:46] <Slyboots> Is there a easy way to see what services my server is actually running
[17:48] <Dans326> can anyone here tell me how to switch php from an apache module to cgi mode
[17:57] <patdk-wk> Dans326, you do know cgi will be dog slow?
[17:57] <patdk-wk> better go with fastcgi
[17:57] <smoser> jamespage, do you know how i can see the java console in firefox ?
[17:57] <smoser> or anyone else
[17:59] <Dans326> yea I know its slow but trying to make a dev server match a live server I have no control over
[18:02] <patdk-wk> apt-get install php5-cgi
[18:02] <patdk-wk> a2dismod php5
[18:06] <Ryan_Lane> I'm trying to make an apparmor policy for memcached, and memcached fails to start because setrlimit is restricted. I can't seem to find a way to allow the process to use setrlimit unrestricted. is this possible?
[18:06] <Ryan_Lane> it looks like I can enforce a stricter rlimit only
[18:10] <Ryan_Lane> specifically, apparmor is denying setrlimit(RLIMIT_NOFILE, {rlim_cur=20000, rlim_max=20000})
[18:14] <jdstrand> jjohansen: ^
[18:15] <jjohansen> Ryan_Lane: if the profile has a rule setting the rlimit, then it becomes controlled, and the task can only change the rlimit to a value <= what is specified in the profile
[18:15] <zul> SpamapS: need a little help?
[18:17] <jjohansen> Ryan_Lane: if you want to allow the task to change its rlimit, you can raise the value in the profile, and or remove the "set rlimit" rule, and reload the profile
[18:17] <Ryan_Lane> the profile doesn't have it set
[18:17] <Ryan_Lane> maybe base or nameservice does...
[18:17] <SpamapS> zul: w/ what?
[18:18] <SpamapS> https://code.launchpad.net/~clint-fewbar/ubuntu/natty/squid/fix-init.d-symlink/+merge/52898
[18:18] <SpamapS> needs some testing and an upload
[18:18] <zul> SpamapS: yep
[18:18]  * SpamapS is headed to the dentist.. bbiab
[18:18] <RoAkSoAx> SpamapS: good louck
[18:18] <RoAkSoAx> luck*
[18:19] <jjohansen> Ryan_Lane: hrmm, you shouldn't be getting a reject message from apparmor unless the profile has it set some where.  It doesn't mediate rlimits unless they are specifically set (or there is a bug)
[18:20] <Ryan_Lane> memcached fails to set the ulimit when its profile is enabled, and doesn't fail when it is disabled
[18:20] <jjohansen> Ryan_Lane: can you see if apparmor_parser -p </path/to/profile> dumps output?
[18:21] <jjohansen> Ryan_Lane: -p is used to dump the profile after include processing etc, so you get a flattened profile
[18:21] <Ryan_Lane> jjohansen: apparmor_parser: invalid option -- 'p'
[18:22] <jjohansen> :(
[18:22] <jjohansen> Ryan_Lane: well I can give you an older or newer parser with it enabled, unfortunately it got disabled for a release or two
[18:22] <RoAkSoAx> zul: ok so I just package python-ethtool bug I don't yet know the status of debian bug #549323
[18:23] <RoAkSoAx> zul: what do you think we should do?
[18:23] <zul> RoAkSoAx: upload it and then we can sync it from debian if it gets done
[18:23] <Ryan_Lane> I'm just going to ignore apparmor for memcached right now
[18:24] <RoAkSoAx> zul: alrgiht. gonna ask FFe then first
[18:24] <jjohansen> Ryan_Lane: which OS release?
[18:24] <Ryan_Lane> lucid
[18:24] <Ryan_Lane> it was working properly in hardy
[18:24] <zul> RoAkSoAx: k
[18:24] <jjohansen> Ryan_Lane: okay, if you want you can send me the profile and I'll see if I can replicate
[18:24] <Ryan_Lane> jjohansen: ok. sec
[18:25] <Ryan_Lane> jjohansen: http://pastebin.com/2c3YH2L5
[18:26] <jjohansen> okay thanks
[18:27] <RoAkSoAx> zul: could you review it first though?
[18:27] <zul> RoAkSoAx: sure
[18:27] <RoAkSoAx> zul: do you want me to upload it to chinstrap or revu?
[18:27] <zul> chinstrap please
[18:29] <_Neytiri_> is anyone here familiar with using Ubuntu as a iSCSI target. i cant seem to get my initiators to connect
[18:29] <patdk-wk> sure
[18:31] <_Neytiri_> http://www.howtoforge.com/using-iscsi-on-ubuntu-10.04-initiator-and-target i used that tutorial to set up my target and iam useign vmware esxi as my initatior
[18:32] <_Neytiri_> i have 2 2tb drives /dev/dsb and /dev/sdd
[18:33] <RoAkSoAx> zul: done! Please, also note this warning "W: dh_python2:95: Python 2.7 should install files in /usr/lib/python2.7/dist-packages/. Did you forget "--install-layout=deb"?", which I can't figure out why, but if you do let me know. More info here:http://pastebin.ubuntu.com/578435/
[18:36] <zul> SpamapS: done
[18:38] <_Neytiri_> Target iqn.2011-03.net..pandpra.eywa:iscsi.1 OutgoingUser       Lun 0 Path=/dev/sdb,Type=fileio       Lun 1 Path=/dev/sdd,Type=fileio
[18:38] <_Neytiri_> that is my ietd.conf file
[18:39] <patdk-wk> comment out outgoing user
[18:39] <patdk-wk> and you shouldn't use fileio for a block device
[18:39] <_Neytiri_> ok so it should be a blockio ?
[18:41] <patdk-wk> if using a block device yep, should perform better, but doubt it's the cause of any *issues*
[18:42] <_Neytiri_> ok thats done
[18:43] <_Neytiri_> ok
[18:43] <_Neytiri_> i am still got getting the drives to show up on my initiator
[18:49] <_Neytiri_> patdk-wk, any idea what could be causin git?
[18:49] <patdk-wk> causing what?
[18:49] <patdk-wk> you never even said you had a problem
[18:49] <_Neytiri_> my initiators to not be able to connect
[18:49] <patdk-wk> let along what the problem is
[18:50] <patdk-wk> error message?
[18:50] <_Neytiri_> no messages what so ever
[18:50] <patdk-wk> I distinctly remember vmware throwing up messages if it can't connect
[18:51] <_Neytiri_> it might be connecting but the drives are not showing up
[18:51] <patdk-wk> did you format them?
[18:51] <_Neytiri_> no
[18:52] <patdk-wk> maybe you need to go back and read the esx manual
[18:52] <patdk-wk> it won't *auto* showup
[18:52] <patdk-wk> not in esx
[18:52] <hggdh> smoser, do we have an EC2 image for the karmic proposed?
[18:52] <smoser> yes
[18:52] <_Neytiri_> it doesent even show up under he storage adaptar
[18:53] <patdk-wk> under storage adaptors, select the iscsi adaptor you configured
[18:53] <patdk-wk> it should show each iscsi target
[18:54] <_Neytiri_> there not showing up there
[18:54] <_Neytiri_> targets 0
[18:54] <patdk-wk> click properties
[18:54] <patdk-wk> configure, enable
[18:55] <patdk-wk> then on the dynamic discovery page type in the ip of your iscsi device (or dns name)
[18:55] <smoser> hggdh, use http://uec-images.ubuntu.com/server/karmic/20110308/
[18:55] <patdk-wk> it's all in the esx manual, really simple
[18:55] <smoser> the 20110310 has the non-proposed
[18:55] <_Neytiri_> i did
[18:55] <smoser> you can ssee such things by looking at the 'kernel-info' files there
[18:55] <patdk-wk> if it can't find anything, then iscsi isn't runnin on your host
[18:55] <_Neytiri_> rescan shows nothing
[18:55] <patdk-wk> did you restart the iscsi daemon after you edited that config file?
[18:55] <_Neytiri_> yes
[18:56] <patdk-wk> mine is very simple config
[18:56] <_Neytiri_> it seemed simple whats why i dont know what i did wrong
[18:57] <patdk-wk> Target ign:1998-02.com.example:esx
[18:57] <patdk-wk>   Lun 0 Path=/dev/vg/iscsi,type=blockio
[18:57] <patdk-wk> and is all
[18:58] <patdk-wk> and yes, I just added a new target to my esx install to test it
[18:58] <_Neytiri_> well tiem for lunck this has to weait
[19:05] <hggdh> smoser, thank you
[19:39] <RoAkSoAx> zul: is kickstarting working in cobbler? http://me.roaksoax.com/Screenshot.png
[19:39] <RoAkSoAx> i get that error
[20:03] <cavefish> hey
[20:05] <cavefish> i've got a question: I want to virtualize an Windows XP client on a Ubuntu Server 10.10 ... is there a way to do this without installing the X environment.
[20:05] <cavefish> ?
[20:06] <elb0w`> So funny thing, apt-get install vsftpd. Then a /etc/init.d/vsftp start yields  * MySQL is not running
[20:06] <ivoks> hm
[20:06] <ivoks> cavefish: yes
[20:07] <ivoks> cavefish: kvm and libvirt are your friends
[20:07] <cavefish> Oke thank you mate!
[20:20] <ScottK> RoAkSoAx: There is (IIRC) a fixup that moves it out of /usr/local, but it's better you add install-layout=deb so it doesn't have to save you.
[20:23] <RoAkSoAx> ScottK: I have --install-layout=deb and it still shows the warning
[20:24] <ScottK> RoAkSoAx: Can you pastebin your debian/rules?
[20:24] <RoAkSoAx> ScottK: http://paste.ubuntu.com/578505/
[20:27] <ScottK> RoAkSoAx: What happens if you make line 14 dh_install?
[20:28]  * RoAkSoAx checking
[20:29] <RoAkSoAx> ScottK: ah!! now there's no warning. Thanks for that...w ouldn't have guessed it xD
[20:30]  * RoAkSoAx wonders why he didn't think about that
[20:30] <ScottK> RoAkSoAx: dh_auto_install was trying to build that package again.
[20:31] <RoAkSoAx> ScottK: I see. So, when overriding dh_auto_install should we always use dh_install rather than dh_auto_install againa, or this is just a particular case?
[20:32] <ScottK> RoAkSoAx: Dunno about always, but usually.
[20:32] <ScottK> I can't think of a case you wouldn't though.
[20:33] <RoAkSoAx> Alright. I'll keep that in mind for the future. Thanks :)
[20:52] <RoAkSoAx> hallyn: howdy! Do you know by any change if there's any reason whye PXE Booting a KVM/virt-manager instance would result in not being able to find a root filesystem if the disk is using virtIO instead of IDE?
[20:52] <hallyn> RoAkSoAx: that sounds familiar.  It might be a known bug
[20:52] <hallyn> which i think soren filed
[20:52]  * RoAkSoAx looks
[20:54] <hallyn> for some reason it's not in my bugs.otl file
[20:55] <RoAkSoAx> yeha can't find any
[20:55] <RoAkSoAx> either
[20:58] <hallyn> RoAkSoAx: oh, pxe.  His was different I think.  you say it works when you use ide?
[20:58] <hallyn> RoAkSoAx: what is your full cmdline?
[20:59] <RoAkSoAx> hallyn: /usr/bin/kvm -S -M pc-0.14 -enable-kvm -m 512 -smp 2,sockets=2,cores=1,threads=1 -name natty-pxe -uuid 0e9133db-924d-4b6c-efda-e07910ea2369 -nodefconfig -nodefaults -chardev socket,id=charmonitor,path=/var/lib/libvirt/qemu/natty-pxe.monitor,server,nowait -mon chardev=charmonitor,id=monitor,mode=readline -rtc base=utc -no-reboot -boot nc -drive file=/media/vm1/natty5.img,if=none,id=drive-ide0-0-0,format=raw -device ide-drive,bus=ide.0,uni
[21:01] <hallyn> RoAkSoAx: what if you add 'boot=on' (and maybe 'index=0') to the file=/media/vm1/natty5.img spec?
[21:01] <RoAkSoAx> hallyn: with the command above it's installing normally. let me get you the command with virtIO
[21:06] <RoAkSoAx> hallyn: ok, this this it fails: /usr/bin/kvm -S -M pc-0.14 -enable-kvm -m 512 -smp 2,sockets=2,cores=1,threads=1 -name natty-pxe2 -uuid 36bc2185-3dad-246c-2506-7ab226bdab47 -nodefconfig -nodefaults -chardev socket,id=charmonitor,path=/var/lib/libvirt/qemu/natty-pxe2.monitor,server,nowait -mon chardev=charmonitor,id=monitor,mode=readline -rtc base=utc -no-reboot -boot nc -drive file=/media/vm1/natty6.img,if=none,id=drive-virtio-disk0,boot=on,form
[21:07] <oneseventeen> I'm trying to set up a MySQL server to be connected to remotely, I've set it to bind-to the correct IP but still am not allowed to connect remotely.
[21:07] <oneseventeen> I'm assuming this is a user permissions thing, but wanted to double check if there was something else that was common
[21:10] <thesheff17_> oneseventeen: you want to make sure your comment out #bind-address           = 127.0.0.1
[21:11] <thesheff17_> oneseventeen: then you have to create a remote user GRANT ALL PRIVILEGES ON *.* TO 'monty'@'%' IDENTIFIED BY 'some_pass' WITH GRANT OPTION; % means the user can connect from anywhere.
[21:11] <hallyn> RoAkSoAx: mind you it is possible that the bios simply works with ide as backup and not with virtio.  But your cmdline still seems weird - can you try with if=virtio,index=0 in the last bit?
[21:12] <oneseventeen> thesheff17_: thanks!
[21:12] <oneseventeen> that worked perfectly, I can now connect remotely.
[21:13] <Guest98139> good evening
[21:13] <GeekyAdam> hi all. getting a padlock_sha error on boot of my server and it halts. gonna paste to pastebin.
[21:13] <GeekyAdam> http://paste.ubuntu.com/578523
[21:13] <Guest98139> i have a question about x-forwarding via ssh and using screen the same time
[21:14] <Guest98139> is it possible to x-forward a program which is running on a server and detach an later reattach it?
[21:14] <GeekyAdam> Guest98139: this might help: http://www.vanemery.com/Linux/XoverSSH/X-over-SSH2.html
[21:15] <Guest98139> thanks geekyadam, i will try i right away
[21:18] <guntbert> Guest98139: and you want to look into gnome screen  (for detaching,...)
[21:19] <Guest98139> gnome screen? i do not have a windowmanager on my server
[21:20] <guntbert> Guest98139: its text only :-), and it GNU screen (silly me)
[21:20] <Guest98139> isn't it possible to run the application without that bloat?
[21:21] <Guest98139> i mean, of course i can run the application with ssh user@server -X
[21:21] <Guest98139> but i would like to de- an reattach later
[21:21] <Guest98139> and this does not work with plain screen and ssh -X
[21:21] <Patrickdk> if you want that, you will need like xvnc
[21:21] <guntbert> Guest98139: I was really talking about a text only solution for detaching/reattaching
[21:22] <Guest98139> thanks guntbert, i know that a lot of people use irsi like that
[21:22] <guntbert> Guest98139: :)
[21:27] <dravekx> i have a phpbb forum. When an admin adds a module from inside phpbb, the ownership is set to user:user instead of user:www-data. how do i fix that?
[21:31] <ivoks> that sounds like a correct behavior
[21:32] <ivoks> and you can't make user to change permissions to a group he's not part of
[21:32] <Guest98139> or is there a command line client for 'direct connect' aka dc++?
[21:32] <ivoks> and you really don't want www-data to own any .php file on your server
[21:34] <dravekx> ivoks, /var/www is user:www-data. When I copy files and folders via SFTP it also defaults them to user:user. Is that normal?
[21:35] <ivoks> dravekx: of course
[21:35] <ivoks> www-data shouldn't be owner of any page
[21:36] <ivoks> if you give rw permissions to www-data user on, let's say, joomla, you can be sure that your page will get defaced
[21:36] <ivoks> www-data runs apache service
[21:36] <ivoks> that user should only have read permissions on files in your web page
[21:36] <ivoks> except for the files that people upload
[21:37] <ivoks> smart CMS systems don't upload files over HTTP; they usually provide FTP/SFTP option for uploading
[21:37] <ivoks> great example is wordpress or joomla
[21:37] <ivoks> even drupal supports upload over FTP
[21:37] <dravekx> ivoks, ahhh. :) thx.
[21:38] <ivoks> directorires that www-data usually has rw permissions are something like cache and temporary
[21:39] <ivoks> don't ever give www-data rw permissions on your page
[21:39] <ivoks> i can't stress that enough :)
[21:40] <ivoks> and if you have multiple clients on your server, always chroot them
[21:40] <dravekx> ivoks, so secure-wise, it should be root:root or user:user almost always.
[21:40] <ivoks> yes
[21:40] <ivoks> with read permission for other users
[21:41] <ivoks> that way a security hole in your web application won't allow attacker to overwrite php files
[21:42] <ivoks> downside is that your configuration file, that containes database connection configuration, will be readable by anyone
[21:42] <ivoks> that's why you should chroot your users :)
[21:43] <ivoks> no, making that file owned by www-data is not good idea; i've seen joomla websites defaced cause the attacker was able to write into that file :)
[21:44] <ivoks> best approach would be fcgid
[21:44] <ivoks> ok, that's it... good night ;)
[21:58] <GeekyAdam> hi all. getting a padlock_sha error on boot of my server and it halts. http://paste.ubuntu.com/578523
[21:59] <Jinxed-> Has anyone found advantages of using Net-SNMP vs OpenNMS ?
[22:00] <jmedina> I never used opennms, what is for?
[22:01] <Jinxed-> SNMP
[22:01] <Jinxed-> Monitoring
[22:01] <Jinxed-> it is supose to be "enterprise grade network management"
[22:01] <Jinxed-> 100% open source
[22:02] <jmedina> ohhh, well I know nagios for that which supports snmp
[22:02] <jmedina> zennos,
[22:29] <RoAkSoAx> zul: is cobbler-web disabled but default or something?
[22:29] <RoAkSoAx> i mean, not working out-of-the box
[22:31] <hallyn> kim0: will that was more painfal than it should have been, but spice-gtk finally compiled locally, now I'm pushing it to ppa.  (it'll likely be awhile before it gets built there)
[22:31]  * kim0 feels guilty :)
[22:31] <kim0> why does it have to be that hard
[22:31]  * kim0 hugs hallyn 
[22:32] <kim0> hallyn: can u throw the deb somewhere I can easily grab
[22:32] <hallyn> yeah i'll put it on people.canonical, one sec
[22:33] <hallyn> kim0: http://people.canonical.com/~serge/spice-gtk-debs.tgz
[22:34] <kim0> thnx grabbing
[22:34] <hallyn> note i said it compiled.  i didn't run it :)
[22:35] <hallyn> i based on devzero's pkg so will send the debdiff to him
[22:35] <hallyn> it's be nice if we can sync all with him
[22:35] <hallyn> but i he was happy to apckage a separate older celt-0.5.1, i wasn't, which is why my pkgs diverge
[22:35] <hallyn> anyway, off to clear my head and think about how to package spice-qemu
[22:36] <kim0> hallyn: I can't spot any binary in there
[22:36] <hallyn> zul: lxc update seems to be working for me.  old containers don't work, as expected.  i should probably see if i can come up with a fix for that
[22:36] <kim0> only /usr/lib/stuff
[22:37] <hallyn> kim0: yeah, hm
[22:37] <hallyn> oh.  haha
[22:37] <kim0> hallyn: docs mention "spicy"
[22:37] <kim0> a cli tool
[22:38] <hallyn> i had to rename the client package (bc it conflicts the one from spicec)
[22:38] <hallyn> but i didn't rename the debian/spice-client.install
[22:38] <hallyn> 'oops'
[22:38] <hallyn> biab
[22:38] <kim0> okie
[22:42] <hallyn> kim0: gimme 5 more mins
[22:43] <kim0> sure thingie
[22:47] <hallyn> kim0: http://people.canonical.com/~serge/spice-gtk_0.5-1ubuntu2-debs.tgz
[22:48] <hallyn> this time there are 'spicy' and 'snappy' binaries
[22:48] <hallyn> whatever those are
[22:48] <hallyn> ok, out for a walk, back in awhile
[22:49] <oneseventeen> is there a best practice for setting up a website in Ubuntu-Server?
[22:49] <oneseventeen> do I just put my default website at /var/www/ or should I create a new site and disable the default?
[22:52] <kim0> hallyn: initial testing looks good :)
[23:00] <_Neytiri_> patdk-wk,  you still here, i managed to get the drives to shw up now in vm-ware but i get the error: unable to read partition information from this disk
[23:10] <nertil> anyone knows about mysql on ubuntu?
[23:11] <air^> nertil: what about it?
[23:11] <air^> https://help.ubuntu.com/10.10/serverguide/C/mysql.html
[23:11] <nertil> i dont know my mysql username and password
[23:11] <nertil> how can i reconfigure mysql ?
[23:12] <oneseventeen> I'm setting up a webserver and can't remember how to set permissions so all members of a specific group have read/write access to certain folders.
[23:12] <oneseventeen> I know it has to do with chmod, but there was a flag I remember seeing at one point to say "apply this to new folders as well"
[23:12] <jmedina> man chmod
[23:12] <sabgenton> Hi, how do I enable the power off button do do a shutdown
[23:13] <air^> nertil: http://ubuntu.flowconsult.at/en/mysql-set-change-reset-root-password/
[23:13] <air^> first hit at google.
[23:17] <sabgenton> only use that method if you have nothing open to the public internet
[23:18] <sabgenton> it opens your system for  a second
[23:18] <sabgenton> well untill you flush privs
[23:31] <Slyboots> How.. "safe" would it be to run some web-services behind Apache with Basic Digest auth over SSL
[23:39] <guampa> is there a problem with seahorse not importing a ssh private key? it says "invalid file format". the key is a valid one though
[23:39] <guampa> it only isn't named "Id_rsa", i use more than one key
[23:46] <hallyn> kim0: excellent
[23:48] <guampa> nm found it. had to install seahorse-plugins
[23:49] <kim0> hallyn: yeah, it's working great so far :) with pv qxl drivers, all good. Still cant get sound out, but that's another story, I'll dig into it later. You're awesome ;)
[23:50]  * hallyn dances a little jig
[23:50] <hallyn> kim0: what's the sound issue again?  Is it windows + 64-bit qemu-compatible sound drivers?
[23:51] <kim0> hallyn: I'm running 32bit win7, with ac97 card, windows update could find and install drivers for it
[23:51] <kim0> but I'm just still not hearing anything
[23:51] <hallyn> with libvirt?
[23:52] <kim0> started via virt-manager yes
[23:53] <hallyn> i don't even know how /dev/audio perms get handled nowadays
[23:54] <kim0> hallyn: it's pulseaudio!
[23:55] <kim0> hallyn: another annoying thing, while guest cpu is idle (0%), host has kvm process eating 25% with fans spinning
[23:55] <kim0> I wonder if that's acpi/apic...etc related
[23:56] <hallyn> (just to make sure) you don't get that without spice?
[23:58] <kim0> hmm
[23:58] <kim0> let me try
[23:59] <kim0> hallyn: which vga should I test with ? cyrrus ?