libvirt0 unable to operate due to missing essential files

I am not able to reproduce. Can you give the output of /var/run/ibvirt?

Thanks
chuck

Thanks for answering so quickly. The output should be

idella@karmic:~/Documents$ ls /var/run/libvirt
qemu

It needs to list the sockets as listed above.

libvir: Remote error : unable to connect to '/var/run/libvirt/libvirt-sock-ro': No such file or directory
libvir: Remote error : unable to connect to '/var/run/libvirt/libvirt-sock-ro': No such file or

and '/var/run/libvirt/libvirt-sock-rw.

I can only say I've installed it and unistalled it a number of times, trying to get a workable state.
I've determined these are key components to libvirt.
In a regular kernel booted environment, they are made. It is sane
That's what so ridiculous about this. It's occuring in a xen environment, NOT a regular kernel. How can xend or such interfere?

I can't understand why libvirt actually connects to xen at all! There should be set of sockets under ~/.libvirt too, but bo.

idella@karmic:~$ ls .libvirt
qemu storage

Not there. Their absence is undermining every other thing that libvirt interacts with re virtinst and virt-manager, as I described above. It doesn't surprise me that it doesn't reproduce. I think something has gone haywire on my setup, but they are all standard packages installed in a standard way. I've just relied on re-installing to straighten it out, but no. Maybe can't be reproduced which is unfortunate on my part, since it still needs to be debugged.

Key points, occurs in a xen environment, wrecks most everything else. Nothing radical in /etc/libvirt/libvirtd.conf

Chuck Short, I hope to hear from you again, I can see you are a key participant in dealing with these bugs.
It's Saturday now so I suppose I can't expect you to reply before Monday, and you are busy I'm sure.
From what I've gleaned, you're not going to reproduce. The sockets are so fundamental that they surely would have been reported as a bug by someone else by now. So I suppose it follows that is is low priority at your end. But something has gone awry. I would hope for you to grant it some input so as to finally achieve a sane setup on my karmic.

I've pursued libvirt quite a bit, on lenny for extended periods, and gentoo. http://bugs.gentoo.org/show_bug.cgi?id=305401
This flaw leaves my completely puzzled, short of being fluent in python programming, there is nowhere to go with it. From what I've read, creating the sockets and assigning permissions is one of the first things executed by libvirtd. How can this sytem do anything without them? i.e. Why does virsh connect to xen in their absence?

The only link to xen I can trace that might bear some influence I've already mentioned. In xend-config.sxp, one of the fundamental choices to select is the type of server. The choice of

(xend-http-server yes)
(xend-port 8000)

finally allowed connectivity. The promoted selection is (xend-unix-server yes), which I have just tried. It is absent as a selection in the xend-config.sxp, so I had to manually add it. It also works, but leaves me wondering,
why is this selection absent?
what is a recommended standard setup for karmic with its standard packages?
It's too intricate for me.

I decided to up the level of input. Please persevere.

idella@karmic:~$ sudo rm /var/run/libvirtd.pid
idella@karmic:~$ sudo gdb libvirtd
GNU gdb (GDB) 7.0-ubuntu
Copyright (C) 2009 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law. Type "show copying"
and "show warranty" for details.
This GDB was configured as "i486-linux-gnu".
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>...
Reading symbols from /usr/sbin/libvirtd...(no debugging symbols found)...done.
(gdb) run
Starting program: /usr/sbin/libvirtd
[Thread debugging using libthread_db enabled]
20:11:17.815: warning : qemudStartup:521 : Unable to create cgroup for driver: No such device or address

Program received signal SIGSEGV, Segmentation fault.
virDomainEventCallbackListFree (list=0x0) at domain_event.c:41
41 domain_event.c: No such file or directory.
 in domain_event.c
(gdb) bt
#0 virDomainEventCallbackListFree (list=0x0) at domain_event.c:41
#1 0x0808bece in ?? ()
#2 0x0808f992 in ?? ()
#3 0xb7e8a915 in virStateInitialize (privileged=1) at libvirt.c:786
#4 0x0805b5eb in ?? ()
#5 0xb7a1bb56 in __libc_start_main () from /lib/tls/i686/nosegneg/libc.so.6
#6 0x08053d11 in ?? ()

comment: I suspect this relates to attempting to qemu /kvm, which it can't connect to in this xen environ.

idella@karmic:~$ gdb virsh
GNU gdb (GDB) 7.0-ubuntu
Copyright (C) 2009 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law. Type "show copying"
and "show warranty" for details.
This GDB was configured as "i486-linux-gnu".
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>...
Reading symbols from /usr/bin/virsh...(no debugging symbols found)...done.
(gdb) run
Starting program: /usr/bin/virsh
[Thread debugging using libthread_db enabled]
Connecting to uri: qemu:///session
              (Here it hangs, I interrupt)
^Z
Program received signal SIGTSTP, Stopped (user).
0xb7fe2416 in __kernel_vsyscall ()
(gdb) run -c qemu:///system
The program being debugged has been started already.
Start it from the beginning? (y or n) y
Starting program: /usr/bin/virsh -c qemu:///system
[Thread debugging using libthread_db enabled]
Connecting to uri: qemu:///system
error: unable to connect to '/var/run/libvirt/libvirt-sock': No such file or directory
error: failed to connect to the hypervisor

Program exited with code 01.
(gdb) run -c xen
Starting program: /usr/bin/virsh -c xen
[Thread debugging using libthread_db enabled]
Connecting to uri: xen
error: unable to connect to '/var/run/libvirt/libvirt-sock': No such file or directory
error: failed to connect to the hypervisor

Program exited with code 01.

however

idella@karmic:~$ sudo gdb virsh
GNU gdb (GDB) 7.0-ubuntu
Copyright (C) 2009 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This...

ok, you're going to hate me for this, I know it. As if I've not input enough. I did do a separate install of karmic,
and I reproduced this state. I installed a 64 version.

I installed, prepared the kernel for karmic. Unless I missed it, Ubuntu didn't provide a xen kernel for karmic. I provided my own, and it's run lenny 64, karmic 32, now karmic 64, gentoo 32 and gentoo 64. It may not be perfect. Anyway, having booted into xen.

idella@karmic64:~$ sudo libvirtd
[sudo] password for idella:
22:58:12.769: error : qemudWritePidFile:457 : Failed to open pid file '/var/run/libvirtd.pid' : File exists
idella@karmic64:~$ ls /var/run/libvirt
qemu
idella@karmic64:~$ sudo rm /var/run/libvirtd.pid
idella@karmic64:~$ sudo libvirtd
libvir: Network Config error : cannot enable /proc/sys/net/ipv6/conf/virbr0/disable_ipv6: No such file or directory
23:00:47.610: warning : qemudStartup:521 : Unable to create cgroup for driver: No such device or address
Segmentation fault
idella@karmic64:~$ ls /var/run/libvirt
qemu
idella@karmic64:~$ uname -a
Linux karmic64 2.6.31-xen-r10-AMD-64 #19 SMP Sun Feb 28 21:28:33 WST 2010 x86_64 GNU/Linux

More of the same. Installed virt-manager, try connecting to xen by gui,

Unable to open connection to hypervisor URI 'xen:///':
<class 'libvirt.libvirtError'> unable to connect to '/var/run/libvirt/libvirt-sock': No such file or directory

I booted into the freshly installed server kernel, and

idella@karmic64:~$ sudo libvirtd
[sudo] password for idella:
Sorry, try again.
[sudo] password for idella:
22:39:18.849: error : qemudWritePidFile:457 : Failed to open pid file '/var/run/libvirtd.pid' : File exists
idella@karmic64:~$ sudo rm /var/run/libvirtd.pid
idella@karmic64:~$ sudo libvirtd
22:39:39.118: warning : qemudStartup:521 : Unable to create cgroup for driver: No such device or address
22:39:39.204: warning : lxcStartup:1460 : Unable to create cgroup for driver: No such device or address
WARNING: Unhandled message: interface=org.freedesktop.DBus.Introspectable, path=/, member=Introspect

^X^Z
[1]+ Stopped sudo libvirtd
idella@karmic64:~$ sudo libvirtd
22:40:21.411: error : qemudWritePidFile:457 : Failed to open pid file '/var/run/libvirtd.pid' : File exists
idella@karmic64:~$ ls /var/run/libvirt
libvirt-sock libvirt-sock-ro network qemu
idella@karmic64:~$ uname -a
Linux karmic64 2.6.31-19-server #56-Ubuntu SMP Thu Jan 28 03:40:48 UTC 2010 x86_64 GNU/Linux

Those elusive sockets libvirt-sock libvirt-sock-ro are there, in an environment absent of xen.

Yes well, I fear this may lead nowhere if you can't reproduce. What kernel would you use??

Are you running libvirt in a dom0 or domU?

libvirt is running in dom0.

idella@karmic64:~$ uname -a
 Linux karmic64 2.6.31-xen-r10-AMD-64 #19 SMP Sun Feb 28 21:28:33 WST 2010 x86_64 GNU/Linux

Host. libvirt is intended for a host environment. Wouldn't be inclined to install it into a guest.
It's all very well me confirming myself. Can someone confirm???

Can you try running libvirtd through strace?

bounces@canonical, according to soren from my other bug,

Ubuntu has no support for running as a Xen dom0. This is no secret. When
you insist on doing so anyway, don't expect us to be all over it trying
to work out your bugs. ......

So I proceed with uncertainty. I've not applied strace before. However, I have done as requested.
Oh my, it is large. The first, trace1.txt is run and aborts due to the presence of the pid file, /var/run/libvirtd.pid.
So I remove it, and re-run. I stopped at recording over 4,000 lines, trace2.txt It's obviously repetitive.
Anyway,

On Wed, Mar 03, 2010 at 11:29:36AM -0000, IAN DELANEY wrote:
> Ubuntu has no support for running as a Xen dom0. This is no secret.
> When you insist on doing so anyway, don't expect us to be all over it
> trying to work out your bugs. ......
>
> So I proceed with uncertainty. I've not applied strace before. However, I have done as requested.
> Oh my, it is large. The first, trace1.txt is run and aborts due to the presence of the pid file, /var/run/libvirtd.pid.
> So I remove it, and re-run. I stopped at recording over 4,000 lines, trace2.txt It's obviously repetitive.

It looks like libvirt is unhappy with your kernels lack of namespace
support:

clone(child_stack=0x9b70fe4, flags=CLONE_NEWNS|0x2c000000|SIGCHLD) = -1 EINVAL (Invalid argument)

--
Soren Hansen
Ubuntu Developer
http://www.ubuntu.com/

yes, ok, well thanks for that. Good comment. I'll take note and add it.

Soren, or bounces, well, despite your position that you didn't have to, you did it. You were indeed right. I'm quite happy for that too.
I did do so anyway (dom0 in xen). That makes this not a bug and resolved and the deficit in my setup, perhaps an invalid status.
I could not have figured it.

idella@karmic:~$ uname -a
Linux karmic 2.6.31-xen-r10 #6 SMP Wed Mar 3 21:31:58 WST 2010 i686 GNU/Linux
idella@karmic:~$ ls /var/run/libvirt
libvirt-sock libvirt-sock-ro network qemu

and my kernel more refined. That gives karmic a real lift. Thanks