3:43 a.m.
Hello:
I need to know if I can use the restore operation (virsh o the equivalent in
libvirt) to recover a previous state of a guest, but recovered previously in
another host.
I did a test, but I got an error:
The exactly sequence using virsh I testes is:
On [HOST SOURCE]: Using virsh
1) save [domain] [file]
2) restore file
3) destroy [domain]
On [HOST SOURCE] using ubuntu sh
4) cp [guest.img] [guest.xml] [file] to HOST2
On [HOST TARGET] using virsh
5) define [guest.xml] (using image on destination in HOST2)
6) restore [file]
The restore troughs the following message:
*virsh # restore sv-chubut-2011-04-01-09:58
error: Failed to restore domain from sv-chubut-2011-04-01-09:58
error: monitor socket did not show up.: Connection refused*
Thank you very much.
Marcela.
5:20 a.m.
On Sun, Apr 03, 2011 at 10:43:45AM +0200, Marcela Castro León wrote:
Hello:
I need to know if I can use the restore operation (virsh o the equivalent in
libvirt) to recover a previous state of a guest, but recovered previously in
another host.
I did a test, but I got an error:
The exactly sequence using virsh I testes is:
On [HOST SOURCE]: Using virsh
1) save [domain] [file]
2) restore file
3) destroy [domain]
On [HOST SOURCE] using ubuntu sh
4) cp [guest.img] [guest.xml] [file] to HOST2
On [HOST TARGET] using virsh
5) define [guest.xml] (using image on destination in HOST2)
6) restore [file]
As a general rule you should only ever 'restore' from a
file *once*. This is because after the first restore
operation, the guest may have made writes to its disk.
Restoring a second time the guest OS will likely have
an inconsistent view of the disk & will cause filesystem
corruption.
If you want to be able to restore from a saved image
multiple times, you need to also take a snapshot of
the disk image at the same time, and restore that
snapshot when restoring the memory image.
That aside, saving on one host & restoring on a
different host is fine. So if you leave out steps
2+3 in your example above, then your data would
still be safe.
The restore troughs the following message:
*virsh # restore sv-chubut-2011-04-01-09:58
error: Failed to restore domain from sv-chubut-2011-04-01-09:58
error: monitor socket did not show up.: Connection refused*
There is probably some configuration difference on your 2nd host
that prevented the VM from starting up. If you're lucky the file
/var/log/libvirt/qemu/$NAME.log will tell you more
Daniel
--
|: http://berrange.com -o- http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org -o- http://virt-manager.org :|
|: http://autobuild.org -o- http://search.cpan.org/~danberr/ :|
|: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :|
4:57 a.m.
Hello Daniel
Thank you for all your information, but I still didn't solve the problem. I
tried the option you mention, with two differents guest into two differents
host, but all the cases I've got:
*virsh # restore sv-chubut-2011-04-04-17:38*
*error: Failed to restore domain from sv-chubut-2011-04-04-17:38*
*error: monitor socket did not show up.: Connection refused***
I cannot get any useful information (at least form me) on the log you
mention.
I'd appreciate a lot a new suggestion.
Thanks
Marcela
2011/4/4 Daniel P. Berrange <berrange(a)redhat.com>
On Sun, Apr 03, 2011 at 10:43:45AM +0200, Marcela Castro León wrote:
> Hello:
> I need to know if I can use the restore operation (virsh o the equivalent
in
> libvirt) to recover a previous state of a guest, but recovered previously
in
> another host.
> I did a test, but I got an error:
>
> The exactly sequence using virsh I testes is:
> On [HOST SOURCE]: Using virsh
> 1) save [domain] [file]
> 2) restore file
> 3) destroy [domain]
>
> On [HOST SOURCE] using ubuntu sh
> 4) cp [guest.img] [guest.xml] [file] to HOST2
>
> On [HOST TARGET] using virsh
> 5) define [guest.xml] (using image on destination in HOST2)
> 6) restore [file]
As a general rule you should only ever 'restore' from a
file *once*. This is because after the first restore
operation, the guest may have made writes to its disk.
Restoring a second time the guest OS will likely have
an inconsistent view of the disk & will cause filesystem
corruption.
If you want to be able to restore from a saved image
multiple times, you need to also take a snapshot of
the disk image at the same time, and restore that
snapshot when restoring the memory image.
That aside, saving on one host & restoring on a
different host is fine. So if you leave out steps
2+3 in your example above, then your data would
still be safe.
> The restore troughs the following message:
> *virsh # restore sv-chubut-2011-04-01-09:58
> error: Failed to restore domain from sv-chubut-2011-04-01-09:58
> error: monitor socket did not show up.: Connection refused*
There is probably some configuration difference on your 2nd host
that prevented the VM from starting up. If you're lucky the file
/var/log/libvirt/qemu/$NAME.log will tell you more
Daniel
--
|: http://berrange.com -o- http://www.flickr.com/photos/dberrange/:|
|: http://libvirt.org -o- http://virt-manager.org:|
|: http://autobuild.org -o- http://search.cpan.org/~danberr/:|
|: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc:|
6:38 a.m.
Hi Marcela,
is any other guest on the host that cannot restore this VM working fine ?
You could also try running the:
*/# LIBVIRT_DEBUG=1 virsh restore sv-chubut-2011-04-04-17:38 2>
virsh-restore.log
/*command which would enable the libvirt logging and output the debug
log into the virsh-restore.log file. This file could be sent to the list
for analysis what's wrong.
Thanks,
Michal
On 04/05/2011 11:57 AM, Marcela Castro León wrote:
Hello Daniel
Thank you for all your information, but I still didn't solve the
problem. I tried the option you mention, with two differents guest
into two differents host, but all the cases I've got:
*/virsh # restore sv-chubut-2011-04-04-17:38/*
*/error: Failed to restore domain from sv-chubut-2011-04-04-17:38/*
*/error: monitor socket did not show up.: Connection refused/*
I cannot get any useful information (at least form me) on the log you
mention.
I'd appreciate a lot a new suggestion.
Thanks
Marcela
2011/4/4 Daniel P. Berrange <berrange(a)redhat.com
<mailto:berrange@redhat.com>>
On Sun, Apr 03, 2011 at 10:43:45AM +0200, Marcela Castro León wrote:
> Hello:
> I need to know if I can use the restore operation (virsh o the
equivalent in
> libvirt) to recover a previous state of a guest, but recovered
previously in
> another host.
> I did a test, but I got an error:
>
> The exactly sequence using virsh I testes is:
> On [HOST SOURCE]: Using virsh
> 1) save [domain] [file]
> 2) restore file
> 3) destroy [domain]
>
> On [HOST SOURCE] using ubuntu sh
> 4) cp [guest.img] [guest.xml] [file] to HOST2
>
> On [HOST TARGET] using virsh
> 5) define [guest.xml] (using image on destination in HOST2)
> 6) restore [file]
As a general rule you should only ever 'restore' from a
file *once*. This is because after the first restore
operation, the guest may have made writes to its disk.
Restoring a second time the guest OS will likely have
an inconsistent view of the disk & will cause filesystem
corruption.
If you want to be able to restore from a saved image
multiple times, you need to also take a snapshot of
the disk image at the same time, and restore that
snapshot when restoring the memory image.
That aside, saving on one host & restoring on a
different host is fine. So if you leave out steps
2+3 in your example above, then your data would
still be safe.
> The restore troughs the following message:
> *virsh # restore sv-chubut-2011-04-01-09:58
> error: Failed to restore domain from sv-chubut-2011-04-01-09:58
> error: monitor socket did not show up.: Connection refused*
There is probably some configuration difference on your 2nd host
that prevented the VM from starting up. If you're lucky the file
/var/log/libvirt/qemu/$NAME.log will tell you more
Daniel
--
|: http://berrange.com -o-
http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org -o-
http://virt-manager.org :|
|: http://autobuild.org -o-
http://search.cpan.org/~danberr/
<http://search.cpan.org/%7Edanberr/> :|
|: http://entangle-photo.org -o-
http://live.gnome.org/gtk-vnc :|
--
libvir-list mailing list
libvir-list(a)redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list
--
Michal Novotny <minovotn(a)redhat.com>, RHCE
Virtualization Team (xen userspace), Red Hat
9:54 a.m.
Hello
This is the log I got doing the restore. It's says that it coun't get the
image, but the image is ok, because I can startup the guest.
Neither I can migrate the guest, so I suppose I've a problem in my
configuration.
Thank you very much in advance.
Marcela.
2011/4/5 Michal Novotny <minovotn(a)redhat.com>
Hi Marcela,
is any other guest on the host that cannot restore this VM working fine ?
You could also try running the:
*/# LIBVIRT_DEBUG=1 virsh restore sv-chubut-2011-04-04-17:38 2>
virsh-restore.log
/*command which would enable the libvirt logging and output the debug
log into the virsh-restore.log file. This file could be sent to the list
for analysis what's wrong.
Thanks,
Michal
On 04/05/2011 11:57 AM, Marcela Castro León wrote:
> Hello Daniel
> Thank you for all your information, but I still didn't solve the
> problem. I tried the option you mention, with two differents guest
> into two differents host, but all the cases I've got:
>
> */virsh # restore sv-chubut-2011-04-04-17:38/*
> */error: Failed to restore domain from sv-chubut-2011-04-04-17:38/*
> */error: monitor socket did not show up.: Connection refused/*
>
> I cannot get any useful information (at least form me) on the log you
> mention.
> I'd appreciate a lot a new suggestion.
> Thanks
> Marcela
>
>
>
>
> 2011/4/4 Daniel P. Berrange <berrange(a)redhat.com
> <mailto:berrange@redhat.com>>
>
> On Sun, Apr 03, 2011 at 10:43:45AM +0200, Marcela Castro León wrote:
> > Hello:
> > I need to know if I can use the restore operation (virsh o the
> equivalent in
> > libvirt) to recover a previous state of a guest, but recovered
> previously in
> > another host.
> > I did a test, but I got an error:
> >
> > The exactly sequence using virsh I testes is:
> > On [HOST SOURCE]: Using virsh
> > 1) save [domain] [file]
> > 2) restore file
> > 3) destroy [domain]
> >
> > On [HOST SOURCE] using ubuntu sh
> > 4) cp [guest.img] [guest.xml] [file] to HOST2
> >
> > On [HOST TARGET] using virsh
> > 5) define [guest.xml] (using image on destination in HOST2)
> > 6) restore [file]
>
> As a general rule you should only ever 'restore' from a
> file *once*. This is because after the first restore
> operation, the guest may have made writes to its disk.
> Restoring a second time the guest OS will likely have
> an inconsistent view of the disk & will cause filesystem
> corruption.
>
> If you want to be able to restore from a saved image
> multiple times, you need to also take a snapshot of
> the disk image at the same time, and restore that
> snapshot when restoring the memory image.
>
>
> That aside, saving on one host & restoring on a
> different host is fine. So if you leave out steps
> 2+3 in your example above, then your data would
> still be safe.
>
> > The restore troughs the following message:
> > *virsh # restore sv-chubut-2011-04-01-09:58
> > error: Failed to restore domain from sv-chubut-2011-04-01-09:58
> > error: monitor socket did not show up.: Connection refused*
>
> There is probably some configuration difference on your 2nd host
> that prevented the VM from starting up. If you're lucky the file
> /var/log/libvirt/qemu/$NAME.log will tell you more
>
> Daniel
> --
> |: http://berrange.com -o-
> http://www.flickr.com/photos/dberrange/ :|
> |: http://libvirt.org -o-
> http://virt-manager.org :|
> |: http://autobuild.org -o-
> http://search.cpan.org/~danberr/
> <http://search.cpan.org/%7Edanberr/> :|
> |: http://entangle-photo.org -o-
> http://live.gnome.org/gtk-vnc :|
>
>
>
> --
> libvir-list mailing list
> libvir-list(a)redhat.com
> https://www.redhat.com/mailman/listinfo/libvir-list
--
Michal Novotny <minovotn(a)redhat.com>, RHCE
Virtualization Team (xen userspace), Red Hat
10:09 a.m.
Hi Marcela,
I was investigating the log file and it seems like the image file cannot
be opened on the remote host.
According to the lost you're doing the restore on the host named
rionegro so not the localhost. This seems like the saved guest image is
not accessible from the rionegro system. Could you please try to connect
to rionegro system using SSH and then connect to the default system
hypervisor using:
# virsh restore <image>
with no specification of remote system to connect to the default
hypervisor (default is qemu:///system under root account).
Also, what may be causing issues is the colon character (':') AFAIK so
try renaming the image from sv-chubut-2011-04-04-17:38 to some other
name without spaces and colon characters, e.g. to
sv-chubut-2011-04-04-17-38 and try to restore this way.
Since according to the code it's about opening file error I guess the
remote system is not having access to the file.
Michal
On 04/05/2011 04:54 PM, Marcela Castro León wrote:
Hello
This is the log I got doing the restore. It's says that it coun't get
the image, but the image is ok, because I can startup the guest.
Neither I can migrate the guest, so I suppose I've a problem in my
configuration.
Thank you very much in advance.
Marcela.
2011/4/5 Michal Novotny <minovotn(a)redhat.com <mailto:minovotn@redhat.com>>
Hi Marcela,
is any other guest on the host that cannot restore this VM working
fine ?
You could also try running the:
*/# LIBVIRT_DEBUG=1 virsh restore sv-chubut-2011-04-04-17:38 2>
virsh-restore.log
/*command which would enable the libvirt logging and output the debug
log into the virsh-restore.log file. This file could be sent to
the list
for analysis what's wrong.
Thanks,
Michal
On 04/05/2011 11:57 AM, Marcela Castro León wrote:
> Hello Daniel
> Thank you for all your information, but I still didn't solve the
> problem. I tried the option you mention, with two differents guest
> into two differents host, but all the cases I've got:
>
> */virsh # restore sv-chubut-2011-04-04-17:38/*
> */error: Failed to restore domain from sv-chubut-2011-04-04-17:38/*
> */error: monitor socket did not show up.: Connection refused/*
>
> I cannot get any useful information (at least form me) on the
log you
> mention.
> I'd appreciate a lot a new suggestion.
> Thanks
> Marcela
>
>
>
>
> 2011/4/4 Daniel P. Berrange <berrange(a)redhat.com
<mailto:berrange@redhat.com>
> <mailto:berrange@redhat.com <mailto:berrange@redhat.com>>>
>
> On Sun, Apr 03, 2011 at 10:43:45AM +0200, Marcela Castro
León wrote:
> > Hello:
> > I need to know if I can use the restore operation (virsh o the
> equivalent in
> > libvirt) to recover a previous state of a guest, but recovered
> previously in
> > another host.
> > I did a test, but I got an error:
> >
> > The exactly sequence using virsh I testes is:
> > On [HOST SOURCE]: Using virsh
> > 1) save [domain] [file]
> > 2) restore file
> > 3) destroy [domain]
> >
> > On [HOST SOURCE] using ubuntu sh
> > 4) cp [guest.img] [guest.xml] [file] to HOST2
> >
> > On [HOST TARGET] using virsh
> > 5) define [guest.xml] (using image on destination in HOST2)
> > 6) restore [file]
>
> As a general rule you should only ever 'restore' from a
> file *once*. This is because after the first restore
> operation, the guest may have made writes to its disk.
> Restoring a second time the guest OS will likely have
> an inconsistent view of the disk & will cause filesystem
> corruption.
>
> If you want to be able to restore from a saved image
> multiple times, you need to also take a snapshot of
> the disk image at the same time, and restore that
> snapshot when restoring the memory image.
>
>
> That aside, saving on one host & restoring on a
> different host is fine. So if you leave out steps
> 2+3 in your example above, then your data would
> still be safe.
>
> > The restore troughs the following message:
> > *virsh # restore sv-chubut-2011-04-01-09:58
> > error: Failed to restore domain from
sv-chubut-2011-04-01-09:58
> > error: monitor socket did not show up.: Connection refused*
>
> There is probably some configuration difference on your 2nd host
> that prevented the VM from starting up. If you're lucky the file
> /var/log/libvirt/qemu/$NAME.log will tell you more
>
> Daniel
> --
> |: http://berrange.com -o-
> http://www.flickr.com/photos/dberrange/ :|
> |: http://libvirt.org -o-
> http://virt-manager.org :|
> |: http://autobuild.org -o-
> http://search.cpan.org/~danberr/
<http://search.cpan.org/%7Edanberr/>
> <http://search.cpan.org/%7Edanberr/> :|
> |: http://entangle-photo.org -o-
> http://live.gnome.org/gtk-vnc :|
>
>
>
> --
> libvir-list mailing list
> libvir-list(a)redhat.com <mailto:libvir-list@redhat.com>
> https://www.redhat.com/mailman/listinfo/libvir-list
--
Michal Novotny <minovotn(a)redhat.com <mailto:minovotn@redhat.com>>,
RHCE
Virtualization Team (xen userspace), Red Hat
--
Michal Novotny <minovotn(a)redhat.com>, RHCE
Virtualization Team (xen userspace), Red Hat
10:15 a.m.
On 04/05/2011 09:09 AM, Michal Novotny wrote:
Hi Marcela,
I was investigating the log file and it seems like the image file cannot
be opened on the remote host.
Are you using SELinux, and is your shared storage on NFS? If so, did
you run 'setsebool -P virt_use_nfs on' to let the destination properly
access the shared storage from NFS?
--
Eric Blake eblake(a)redhat.com +1-801-349-2682
Libvirt virtualization library http://libvirt.org
10:17 a.m.
On 04/05/2011 05:15 PM, Eric Blake wrote:
On 04/05/2011 09:09 AM, Michal Novotny wrote:
> Hi Marcela,
> I was investigating the log file and it seems like the image file cannot
> be opened on the remote host.
Are you using SELinux, and is your shared storage on NFS? If so, did
you run 'setsebool -P virt_use_nfs on' to let the destination properly
access the shared storage from NFS?
Eric, I'm not too familiar with SELinux. Do you think the
virFileOpenAs() would fail if the SELinux context won't be set properly?
Michal
--
Michal Novotny <minovotn(a)redhat.com>, RHCE
Virtualization Team (xen userspace), Red Hat
10:42 a.m.
On 04/05/2011 09:17 AM, Michal Novotny wrote:
On 04/05/2011 05:15 PM, Eric Blake wrote:
> On 04/05/2011 09:09 AM, Michal Novotny wrote:
>> Hi Marcela,
>> I was investigating the log file and it seems like the image file cannot
>> be opened on the remote host.
> Are you using SELinux, and is your shared storage on NFS? If so, did
> you run 'setsebool -P virt_use_nfs on' to let the destination properly
> access the shared storage from NFS?
>
Eric, I'm not too familiar with SELinux. Do you think the
virFileOpenAs() would fail if the SELinux context won't be set properly?
Yes, if you are running with SELinux enforcing, then it is possible for
all attempts to open() to fail due to denied permissions, even with
virFileOpenAs() trying different uids, if you haven't told SELinux to
allow NFS images.
--
Eric Blake eblake(a)redhat.com +1-801-349-2682
Libvirt virtualization library http://libvirt.org
11:16 a.m.
Hellol
I've downloaded the policycoreutils, and then I execute on the nfs clients:
*radic@santacruz:~/discoguest/aplicaciones/virsh$ setsebool -P virt_use_nfs
on
setsebool: SELinux is disabled.
*
It's enough? I understand that it's not working, but I don't find good
information at internet.
Thank you.
Marcela
2011/4/5 Michal Novotny <minovotn(a)redhat.com>
On 04/05/2011 05:15 PM, Eric Blake wrote:
> On 04/05/2011 09:09 AM, Michal Novotny wrote:
>> Hi Marcela,
>> I was investigating the log file and it seems like the image file cannot
>> be opened on the remote host.
> Are you using SELinux, and is your shared storage on NFS? If so, did
> you run 'setsebool -P virt_use_nfs on' to let the destination properly
> access the shared storage from NFS?
>
Eric, I'm not too familiar with SELinux. Do you think the
virFileOpenAs() would fail if the SELinux context won't be set properly?
Michal
--
Michal Novotny <minovotn(a)redhat.com>, RHCE
Virtualization Team (xen userspace), Red Hat
11:19 a.m.
On 04/05/2011 06:16 PM, Marcela Castro León wrote:
Hellol
I've downloaded the policycoreutils, and then I execute on the nfs
clients:
*radic@santacruz:~/discoguest/aplicaciones/virsh$ setsebool -P
virt_use_nfs on
setsebool: SELinux is disabled.
*
It's enough? I understand that it's not working, but I don't find good
information at internet.
Thank you.
Marcela
That's fine since you have the SELinux disabled and therefore it can't
cause any issues.
Michal
--
Michal Novotny <minovotn(a)redhat.com>, RHCE
Virtualization Team (xen userspace), Red Hat
3:43 a.m.
Hello:
I repeated all the test again and reproduce a similar situation.
Apparently the restore is finished ok, and the guest is running but:
a) I cannot connect to the guest trough ssh.
b) I cannot enter to the guest using the virt-manager: It's say " Connecting
to graphical console for guest".
c) on the log /var/log/libvirt(qemu/chompi1.log say that an error happened
during the restore.
I'm sending the trace I did while restoring the file and the log. (both on
target rionegro
Thank you very much in advance.
Marcela.
2011/4/5 Michal Novotny <minovotn(a)redhat.com>
On 04/05/2011 06:16 PM, Marcela Castro León wrote:
> Hellol
> I've downloaded the policycoreutils, and then I execute on the nfs
> clients:
>
> *radic@santacruz:~/discoguest/aplicaciones/virsh$ setsebool -P
> virt_use_nfs on
> setsebool: SELinux is disabled.
> *
>
> It's enough? I understand that it's not working, but I don't find good
> information at internet.
>
> Thank you.
> Marcela
That's fine since you have the SELinux disabled and therefore it can't
cause any issues.
Michal
--
Michal Novotny <minovotn(a)redhat.com>, RHCE
Virtualization Team (xen userspace), Red Hat
5:09 a.m.
Hi,
the interesting thing is the "Permission denied" message in the log file:
LC_ALL=C
PATH=/usr/local/sbin:/usr/local/bin:/usr/bin:/usr/sbin:/sbin:/bin
QEMU_AUDIO_DRV=none /usr/bin/qemu-system-x86_64 -S -M pc-0.12
-enable-kvm -m 4096 -smp 4 -name chompi1 -uuid
22f013aa-25d4-b317-9eac-b232692db8c7 -chardev
socket,id=monitor,path=/var/lib/libvirt/qemu/chompi1.monitor,server,nowait
-monitor chardev:monitor -boot c -drive
file=/home/radic/mvdata/imagenes/chompi1.img,if=ide,index=0,boot=on,format=raw
-drive if=ide,media=cdrom,index=2,format=raw -net
nic,macaddr=52:54:00:82:ce:80,vlan=0,model=virtio,name=virtio.0 -net
tap,fd=34,vlan=0,name=tap.0 -chardev pty,id=serial0 -serial
chardev:serial0 -parallel none -usb -vnc 0.0.0.0:0 -vga cirrus -incoming
exec:cat
char device redirected to /dev/pts/2
cat: -: Permission denied <== THIS ONE
load of migration failed
I don't know much about the internal libvirt design since libvirt is not
my primary space so I hope my e-mails from yesterday helped a little but
I'll leave the rest for libvirt guys since I'm not having any other
ideas what could be wrong there.
Michal
On 04/06/2011 10:43 AM, Marcela Castro León wrote:
Hello:
I repeated all the test again and reproduce a similar situation.
Apparently the restore is finished ok, and the guest is running but:
a) I cannot connect to the guest trough ssh.
b) I cannot enter to the guest using the virt-manager: It's say "
Connecting to graphical console for guest".
c) on the log /var/log/libvirt(qemu/chompi1.log say that an error
happened during the restore.
I'm sending the trace I did while restoring the file and the log.
(both on target rionegro
Thank you very much in advance.
Marcela.
2011/4/5 Michal Novotny <minovotn(a)redhat.com <mailto:minovotn@redhat.com>>
On 04/05/2011 06:16 PM, Marcela Castro León wrote:
> Hellol
> I've downloaded the policycoreutils, and then I execute on the nfs
> clients:
>
> *radic@santacruz:~/discoguest/aplicaciones/virsh$ setsebool -P
> virt_use_nfs on
> setsebool: SELinux is disabled.
> *
>
> It's enough? I understand that it's not working, but I don't
find good
> information at internet.
>
> Thank you.
> Marcela
That's fine since you have the SELinux disabled and therefore it can't
cause any issues.
Michal
--
Michal Novotny <minovotn(a)redhat.com <mailto:minovotn@redhat.com>>,
RHCE
Virtualization Team (xen userspace), Red Hat
--
Michal Novotny <minovotn(a)redhat.com>, RHCE
Virtualization Team (xen userspace), Red Hat
11:32 a.m.
On 04/06/2011 04:09 AM, Michal Novotny wrote:
Hi,
the interesting thing is the "Permission denied" message in the log file:
LC_ALL=C
PATH=/usr/local/sbin:/usr/local/bin:/usr/bin:/usr/sbin:/sbin:/bin
QEMU_AUDIO_DRV=none /usr/bin/qemu-system-x86_64 -S -M pc-0.12
-enable-kvm -m 4096 -smp 4 -name chompi1 -uuid
22f013aa-25d4-b317-9eac-b232692db8c7 -chardev
socket,id=monitor,path=/var/lib/libvirt/qemu/chompi1.monitor,server,nowait
-monitor chardev:monitor -boot c -drive
file=/home/radic/mvdata/imagenes/chompi1.img,if=ide,index=0,boot=on,format=raw
-drive if=ide,media=cdrom,index=2,format=raw -net
nic,macaddr=52:54:00:82:ce:80,vlan=0,model=virtio,name=virtio.0 -net
tap,fd=34,vlan=0,name=tap.0 -chardev pty,id=serial0 -serial
chardev:serial0 -parallel none -usb -vnc 0.0.0.0:0 -vga cirrus -incoming
exec:cat
char device redirected to /dev/pts/2
cat: -: Permission denied <== THIS ONE
load of migration failed
I don't know much about the internal libvirt design since libvirt is not
my primary space so I hope my e-mails from yesterday helped a little but
I'll leave the rest for libvirt guys since I'm not having any other
ideas what could be wrong there.
That's a weird error - cat is complaining that it can't read stdin,
which is usually a symptom associated with SELinux permissions. But you
said earlier in this thread that you aren't using SELinux. Are you
using AppArmor?
At any rate, does upgrading to libvirt 0.9.0 improve the situation?
exec: migration has always been risky (there are various data races
involved in coordinating multiple processes, where qemu is using popen()
to manage the use or even opening the file), and so newer libvirt uses
fd: migration (where the fd is already open, so we don't even have to
worry about permissions from qemu's point of view).
--
Eric Blake eblake(a)redhat.com +1-801-349-2682
Libvirt virtualization library http://libvirt.org
3:18 a.m.
Hello
I'm thinking to upgrade the libvirt to solve the problem as you said in your
last mail.
But I've installed the package using "apt-get install" and I can''t
get
higher version of all the packages involves doing an upgrade in ubuntu (I'm
using 10.04)
Can you suggest-me "an easy" way to upgrate the libvirt?
Thank you very much.
Marcela
2011/4/6 Eric Blake <eblake(a)redhat.com>
On 04/06/2011 04:09 AM, Michal Novotny wrote:
> Hi,
> the interesting thing is the "Permission denied" message in the log file:
>
> LC_ALL=C
> PATH=/usr/local/sbin:/usr/local/bin:/usr/bin:/usr/sbin:/sbin:/bin
> QEMU_AUDIO_DRV=none /usr/bin/qemu-system-x86_64 -S -M pc-0.12
> -enable-kvm -m 4096 -smp 4 -name chompi1 -uuid
> 22f013aa-25d4-b317-9eac-b232692db8c7 -chardev
>
socket,id=monitor,path=/var/lib/libvirt/qemu/chompi1.monitor,server,nowait
> -monitor chardev:monitor -boot c -drive
>
file=/home/radic/mvdata/imagenes/chompi1.img,if=ide,index=0,boot=on,format=raw
> -drive if=ide,media=cdrom,index=2,format=raw -net
> nic,macaddr=52:54:00:82:ce:80,vlan=0,model=virtio,name=virtio.0 -net
> tap,fd=34,vlan=0,name=tap.0 -chardev pty,id=serial0 -serial
> chardev:serial0 -parallel none -usb -vnc 0.0.0.0:0 -vga cirrus -incoming
> exec:cat
> char device redirected to /dev/pts/2
> cat: -: Permission denied <== THIS ONE
> load of migration failed
>
> I don't know much about the internal libvirt design since libvirt is not
> my primary space so I hope my e-mails from yesterday helped a little but
> I'll leave the rest for libvirt guys since I'm not having any other
> ideas what could be wrong there.
That's a weird error - cat is complaining that it can't read stdin,
which is usually a symptom associated with SELinux permissions. But you
said earlier in this thread that you aren't using SELinux. Are you
using AppArmor?
At any rate, does upgrading to libvirt 0.9.0 improve the situation?
exec: migration has always been risky (there are various data races
involved in coordinating multiple processes, where qemu is using popen()
to manage the use or even opening the file), and so newer libvirt uses
fd: migration (where the fd is already open, so we don't even have to
worry about permissions from qemu's point of view).
--
Eric Blake eblake(a)redhat.com +1-801-349-2682
Libvirt virtualization library http://libvirt.org
3:22 a.m.
Hi,
I think the best thing would be to download it directly from libvirt.org
site and install it. It's not the easiest way but it's the best way IMHO.
You can find relevant information there:
http://libvirt.org/compiling.html
Michal
On 04/08/2011 10:18 AM, Marcela Castro León wrote:
Hello
I'm thinking to upgrade the libvirt to solve the problem as you said
in your last mail.
But I've installed the package using "apt-get install" and I
can''t
get higher version of all the packages involves doing an upgrade in
ubuntu (I'm using 10.04)
Can you suggest-me "an easy" way to upgrate the libvirt?
Thank you very much.
Marcela
2011/4/6 Eric Blake <eblake(a)redhat.com <mailto:eblake@redhat.com>>
On 04/06/2011 04:09 AM, Michal Novotny wrote:
> Hi,
> the interesting thing is the "Permission denied" message in the
log file:
>
> LC_ALL=C
> PATH=/usr/local/sbin:/usr/local/bin:/usr/bin:/usr/sbin:/sbin:/bin
> QEMU_AUDIO_DRV=none /usr/bin/qemu-system-x86_64 -S -M pc-0.12
> -enable-kvm -m 4096 -smp 4 -name chompi1 -uuid
> 22f013aa-25d4-b317-9eac-b232692db8c7 -chardev
>
socket,id=monitor,path=/var/lib/libvirt/qemu/chompi1.monitor,server,nowait
> -monitor chardev:monitor -boot c -drive
>
file=/home/radic/mvdata/imagenes/chompi1.img,if=ide,index=0,boot=on,format=raw
> -drive if=ide,media=cdrom,index=2,format=raw -net
> nic,macaddr=52:54:00:82:ce:80,vlan=0,model=virtio,name=virtio.0 -net
> tap,fd=34,vlan=0,name=tap.0 -chardev pty,id=serial0 -serial
> chardev:serial0 -parallel none -usb -vnc 0.0.0.0:0
<http://0.0.0.0:0> -vga cirrus -incoming
> exec:cat
> char device redirected to /dev/pts/2
> cat: -: Permission denied <== THIS ONE
> load of migration failed
>
> I don't know much about the internal libvirt design since
libvirt is not
> my primary space so I hope my e-mails from yesterday helped a
little but
> I'll leave the rest for libvirt guys since I'm not having any other
> ideas what could be wrong there.
That's a weird error - cat is complaining that it can't read stdin,
which is usually a symptom associated with SELinux permissions.
But you
said earlier in this thread that you aren't using SELinux. Are you
using AppArmor?
At any rate, does upgrading to libvirt 0.9.0 improve the situation?
exec: migration has always been risky (there are various data races
involved in coordinating multiple processes, where qemu is using
popen()
to manage the use or even opening the file), and so newer libvirt uses
fd: migration (where the fd is already open, so we don't even have to
worry about permissions from qemu's point of view).
--
Eric Blake eblake(a)redhat.com <mailto:eblake@redhat.com>
+1-801-349-2682
Libvirt virtualization library http://libvirt.org
--
Michal Novotny <minovotn(a)redhat.com>, RHCE
Virtualization Team (xen userspace), Red Hat
10:25 a.m.
Hello
I don't do this, I'd find out about it and set it. But, i don't know if this
avoid the error, but I'm trying to restore the image from the host that is
nfs server.
I will execute, try again and tell you if I succeed.
Thank you.
Marcela
2011/4/5 Eric Blake <eblake(a)redhat.com>
On 04/05/2011 09:09 AM, Michal Novotny wrote:
> Hi Marcela,
> I was investigating the log file and it seems like the image file cannot
> be opened on the remote host.
Are you using SELinux, and is your shared storage on NFS? If so, did
you run 'setsebool -P virt_use_nfs on' to let the destination properly
access the shared storage from NFS?
--
Eric Blake eblake(a)redhat.com +1-801-349-2682
Libvirt virtualization library http://libvirt.org
10:20 a.m.
Hello
OK, this is the new log. Now, the old error appeared again...
*error: Failed to restore domain from XX
error: monitor socket did not show up.: Connection refused
*
Regards.
2011/4/5 Michal Novotny <minovotn(a)redhat.com>
Hi Marcela,
I was investigating the log file and it seems like the image file cannot
be opened on the remote host.
According to the lost you're doing the restore on the host named
rionegro so not the localhost. This seems like the saved guest image is
not accessible from the rionegro system. Could you please try to connect
to rionegro system using SSH and then connect to the default system
hypervisor using:
# virsh restore <image>
with no specification of remote system to connect to the default
hypervisor (default is qemu:///system under root account).
Also, what may be causing issues is the colon character (':') AFAIK so
try renaming the image from sv-chubut-2011-04-04-17:38 to some other
name without spaces and colon characters, e.g. to
sv-chubut-2011-04-04-17-38 and try to restore this way.
Since according to the code it's about opening file error I guess the
remote system is not having access to the file.
Michal
On 04/05/2011 04:54 PM, Marcela Castro León wrote:
> Hello
> This is the log I got doing the restore. It's says that it coun't get
> the image, but the image is ok, because I can startup the guest.
> Neither I can migrate the guest, so I suppose I've a problem in my
> configuration.
> Thank you very much in advance.
> Marcela.
>
> 2011/4/5 Michal Novotny <minovotn(a)redhat.com <mailto:minovotn@redhat.com
>>
>
> Hi Marcela,
> is any other guest on the host that cannot restore this VM working
> fine ?
>
> You could also try running the:
>
> */# LIBVIRT_DEBUG=1 virsh restore sv-chubut-2011-04-04-17:38 2>
> virsh-restore.log
>
> /*command which would enable the libvirt logging and output the debug
> log into the virsh-restore.log file. This file could be sent to
> the list
> for analysis what's wrong.
>
> Thanks,
> Michal
>
> On 04/05/2011 11:57 AM, Marcela Castro León wrote:
> > Hello Daniel
> > Thank you for all your information, but I still didn't solve the
> > problem. I tried the option you mention, with two differents guest
> > into two differents host, but all the cases I've got:
> >
> > */virsh # restore sv-chubut-2011-04-04-17:38/*
> > */error: Failed to restore domain from sv-chubut-2011-04-04-17:38/*
> > */error: monitor socket did not show up.: Connection refused/*
> >
> > I cannot get any useful information (at least form me) on the
> log you
> > mention.
> > I'd appreciate a lot a new suggestion.
> > Thanks
> > Marcela
> >
> >
> >
> >
> > 2011/4/4 Daniel P. Berrange <berrange(a)redhat.com
> <mailto:berrange@redhat.com>
> > <mailto:berrange@redhat.com <mailto:berrange@redhat.com>>>
> >
> > On Sun, Apr 03, 2011 at 10:43:45AM +0200, Marcela Castro
> León wrote:
> > > Hello:
> > > I need to know if I can use the restore operation (virsh o
the
> > equivalent in
> > > libvirt) to recover a previous state of a guest, but
recovered
> > previously in
> > > another host.
> > > I did a test, but I got an error:
> > >
> > > The exactly sequence using virsh I testes is:
> > > On [HOST SOURCE]: Using virsh
> > > 1) save [domain] [file]
> > > 2) restore file
> > > 3) destroy [domain]
> > >
> > > On [HOST SOURCE] using ubuntu sh
> > > 4) cp [guest.img] [guest.xml] [file] to HOST2
> > >
> > > On [HOST TARGET] using virsh
> > > 5) define [guest.xml] (using image on destination in HOST2)
> > > 6) restore [file]
> >
> > As a general rule you should only ever 'restore' from a
> > file *once*. This is because after the first restore
> > operation, the guest may have made writes to its disk.
> > Restoring a second time the guest OS will likely have
> > an inconsistent view of the disk & will cause filesystem
> > corruption.
> >
> > If you want to be able to restore from a saved image
> > multiple times, you need to also take a snapshot of
> > the disk image at the same time, and restore that
> > snapshot when restoring the memory image.
> >
> >
> > That aside, saving on one host & restoring on a
> > different host is fine. So if you leave out steps
> > 2+3 in your example above, then your data would
> > still be safe.
> >
> > > The restore troughs the following message:
> > > *virsh # restore sv-chubut-2011-04-01-09:58
> > > error: Failed to restore domain from
> sv-chubut-2011-04-01-09:58
> > > error: monitor socket did not show up.: Connection refused*
> >
> > There is probably some configuration difference on your 2nd
host
> > that prevented the VM from starting up. If you're lucky the
file
> > /var/log/libvirt/qemu/$NAME.log will tell you more
> >
> > Daniel
> > --
> > |: http://berrange.com -o-
> > http://www.flickr.com/photos/dberrange/ :|
> > |: http://libvirt.org -o-
> > http://virt-manager.org :|
> > |: http://autobuild.org -o-
> > http://search.cpan.org/~danberr/
> <http://search.cpan.org/%7Edanberr/>
> > <http://search.cpan.org/%7Edanberr/> :|
> > |: http://entangle-photo.org -o-
> > http://live.gnome.org/gtk-vnc :|
> >
> >
> >
> > --
> > libvir-list mailing list
> > libvir-list(a)redhat.com <mailto:libvir-list@redhat.com>
> > https://www.redhat.com/mailman/listinfo/libvir-list
>
>
> --
> Michal Novotny <minovotn(a)redhat.com <mailto:minovotn@redhat.com>>,
> RHCE
> Virtualization Team (xen userspace), Red Hat
>
>
--
Michal Novotny <minovotn(a)redhat.com>, RHCE
Virtualization Team (xen userspace), Red Hat
10:28 a.m.
Hi,
what about /var/log/libvirt/qemu/$NAME.log file from the attempt? This
seems like the monitor socket cannot be connected which may be the
result of different version of QEMU on source and target system IMHO.
Could you please provide result of
# virsh version
command from both source and target machine and most likely install both
source and target machine?
Thanks,
Michal
On 04/05/2011 05:20 PM, Marcela Castro León wrote:
Hello
OK, this is the new log. Now, the old error appeared again...
/*error: Failed to restore domain from XX
error: monitor socket did not show up.: Connection refused
*/
Regards.
2011/4/5 Michal Novotny <minovotn(a)redhat.com <mailto:minovotn@redhat.com>>
Hi Marcela,
I was investigating the log file and it seems like the image file
cannot
be opened on the remote host.
According to the lost you're doing the restore on the host named
rionegro so not the localhost. This seems like the saved guest
image is
not accessible from the rionegro system. Could you please try to
connect
to rionegro system using SSH and then connect to the default system
hypervisor using:
# virsh restore <image>
with no specification of remote system to connect to the default
hypervisor (default is qemu:///system under root account).
Also, what may be causing issues is the colon character (':') AFAIK so
try renaming the image from sv-chubut-2011-04-04-17:38 to some other
name without spaces and colon characters, e.g. to
sv-chubut-2011-04-04-17-38 and try to restore this way.
Since according to the code it's about opening file error I guess the
remote system is not having access to the file.
Michal
On 04/05/2011 04:54 PM, Marcela Castro León wrote:
> Hello
> This is the log I got doing the restore. It's says that it
coun't get
> the image, but the image is ok, because I can startup the guest.
> Neither I can migrate the guest, so I suppose I've a problem in my
> configuration.
> Thank you very much in advance.
> Marcela.
>
> 2011/4/5 Michal Novotny <minovotn(a)redhat.com
<mailto:minovotn@redhat.com> <mailto:minovotn@redhat.com
<mailto:minovotn@redhat.com>>>
>
> Hi Marcela,
> is any other guest on the host that cannot restore this VM
working
> fine ?
>
> You could also try running the:
>
> */# LIBVIRT_DEBUG=1 virsh restore sv-chubut-2011-04-04-17:38 2>
> virsh-restore.log
>
> /*command which would enable the libvirt logging and output
the debug
> log into the virsh-restore.log file. This file could be sent to
> the list
> for analysis what's wrong.
>
> Thanks,
> Michal
>
> On 04/05/2011 11:57 AM, Marcela Castro León wrote:
> > Hello Daniel
> > Thank you for all your information, but I still didn't
solve the
> > problem. I tried the option you mention, with two
differents guest
> > into two differents host, but all the cases I've got:
> >
> > */virsh # restore sv-chubut-2011-04-04-17:38/*
> > */error: Failed to restore domain from
sv-chubut-2011-04-04-17:38/*
> > */error: monitor socket did not show up.: Connection refused/*
> >
> > I cannot get any useful information (at least form me) on the
> log you
> > mention.
> > I'd appreciate a lot a new suggestion.
> > Thanks
> > Marcela
> >
> >
> >
> >
> > 2011/4/4 Daniel P. Berrange <berrange(a)redhat.com
<mailto:berrange@redhat.com>
> <mailto:berrange@redhat.com <mailto:berrange@redhat.com>>
> > <mailto:berrange@redhat.com <mailto:berrange@redhat.com>
<mailto:berrange@redhat.com <mailto:berrange@redhat.com>>>>
> >
> > On Sun, Apr 03, 2011 at 10:43:45AM +0200, Marcela Castro
> León wrote:
> > > Hello:
> > > I need to know if I can use the restore operation
(virsh o the
> > equivalent in
> > > libvirt) to recover a previous state of a guest, but
recovered
> > previously in
> > > another host.
> > > I did a test, but I got an error:
> > >
> > > The exactly sequence using virsh I testes is:
> > > On [HOST SOURCE]: Using virsh
> > > 1) save [domain] [file]
> > > 2) restore file
> > > 3) destroy [domain]
> > >
> > > On [HOST SOURCE] using ubuntu sh
> > > 4) cp [guest.img] [guest.xml] [file] to HOST2
> > >
> > > On [HOST TARGET] using virsh
> > > 5) define [guest.xml] (using image on destination in
HOST2)
> > > 6) restore [file]
> >
> > As a general rule you should only ever 'restore' from a
> > file *once*. This is because after the first restore
> > operation, the guest may have made writes to its disk.
> > Restoring a second time the guest OS will likely have
> > an inconsistent view of the disk & will cause filesystem
> > corruption.
> >
> > If you want to be able to restore from a saved image
> > multiple times, you need to also take a snapshot of
> > the disk image at the same time, and restore that
> > snapshot when restoring the memory image.
> >
> >
> > That aside, saving on one host & restoring on a
> > different host is fine. So if you leave out steps
> > 2+3 in your example above, then your data would
> > still be safe.
> >
> > > The restore troughs the following message:
> > > *virsh # restore sv-chubut-2011-04-01-09:58
> > > error: Failed to restore domain from
> sv-chubut-2011-04-01-09:58
> > > error: monitor socket did not show up.: Connection
refused*
> >
> > There is probably some configuration difference on
your 2nd host
> > that prevented the VM from starting up. If you're
lucky the file
> > /var/log/libvirt/qemu/$NAME.log will tell you more
> >
> > Daniel
> > --
> > |: http://berrange.com -o-
> > http://www.flickr.com/photos/dberrange/ :|
> > |: http://libvirt.org -o-
> > http://virt-manager.org :|
> > |: http://autobuild.org -o-
> > http://search.cpan.org/~danberr/
<http://search.cpan.org/%7Edanberr/>
> <http://search.cpan.org/%7Edanberr/>
> > <http://search.cpan.org/%7Edanberr/> :|
> > |: http://entangle-photo.org -o-
> > http://live.gnome.org/gtk-vnc :|
> >
> >
> >
> > --
> > libvir-list mailing list
> > libvir-list(a)redhat.com <mailto:libvir-list@redhat.com>
<mailto:libvir-list@redhat.com <mailto:libvir-list@redhat.com>>
> > https://www.redhat.com/mailman/listinfo/libvir-list
>
>
> --
> Michal Novotny <minovotn(a)redhat.com
<mailto:minovotn@redhat.com> <mailto:minovotn@redhat.com
<mailto:minovotn@redhat.com>>>,
> RHCE
> Virtualization Team (xen userspace), Red Hat
>
>
--
Michal Novotny <minovotn(a)redhat.com <mailto:minovotn@redhat.com>>,
RHCE
Virtualization Team (xen userspace), Red Hat
--
Michal Novotny <minovotn(a)redhat.com>, RHCE
Virtualization Team (xen userspace), Red Hat
10:59 a.m.
Hello
The restore finished ok, I found at a log that was looking the image at an
"old" location. Anyway, the image don't start well, but I've to revise
all
the configuration again, to see what's exactly is happening now. Thank you a
lot, I learned a lot.
The version are the same
*source*
radic@chubut:/var/log/libvirt/qemu$ virsh
virsh # version
Compiled against library: libvir 0.7.5
Using library: libvir 0.7.5
Using API: QEMU 0.7.5
Running hypervisor: QEMU 0.12.3
*target
*radic@rionegro:~$ virsh
virsh # version
Compiled against library: libvir 0.7.5
Using library: libvir 0.7.5
Using API: QEMU 0.7.5
Running hypervisor: QEMU 0.12.3
Thank you very much!!
Regards.
2011/4/5 Michal Novotny <minovotn(a)redhat.com>
Hi,
what about /var/log/libvirt/qemu/$NAME.log file from the attempt? This
seems like the monitor socket cannot be connected which may be the
result of different version of QEMU on source and target system IMHO.
Could you please provide result of
# virsh version
command from both source and target machine and most likely install both
source and target machine?
Thanks,
Michal
On 04/05/2011 05:20 PM, Marcela Castro León wrote:
> Hello
> OK, this is the new log. Now, the old error appeared again...
> /*error: Failed to restore domain from XX
> error: monitor socket did not show up.: Connection refused
> */
>
> Regards.
>
> 2011/4/5 Michal Novotny <minovotn(a)redhat.com <mailto:minovotn@redhat.com
>>
>
> Hi Marcela,
> I was investigating the log file and it seems like the image file
> cannot
> be opened on the remote host.
>
> According to the lost you're doing the restore on the host named
> rionegro so not the localhost. This seems like the saved guest
> image is
> not accessible from the rionegro system. Could you please try to
> connect
> to rionegro system using SSH and then connect to the default system
> hypervisor using:
>
> # virsh restore <image>
>
> with no specification of remote system to connect to the default
> hypervisor (default is qemu:///system under root account).
>
> Also, what may be causing issues is the colon character (':') AFAIK
so
> try renaming the image from sv-chubut-2011-04-04-17:38 to some other
> name without spaces and colon characters, e.g. to
> sv-chubut-2011-04-04-17-38 and try to restore this way.
>
> Since according to the code it's about opening file error I guess the
> remote system is not having access to the file.
>
> Michal
>
> On 04/05/2011 04:54 PM, Marcela Castro León wrote:
> > Hello
> > This is the log I got doing the restore. It's says that it
> coun't get
> > the image, but the image is ok, because I can startup the guest.
> > Neither I can migrate the guest, so I suppose I've a problem in my
> > configuration.
> > Thank you very much in advance.
> > Marcela.
> >
> > 2011/4/5 Michal Novotny <minovotn(a)redhat.com
> <mailto:minovotn@redhat.com> <mailto:minovotn@redhat.com
> <mailto:minovotn@redhat.com>>>
> >
> > Hi Marcela,
> > is any other guest on the host that cannot restore this VM
> working
> > fine ?
> >
> > You could also try running the:
> >
> > */# LIBVIRT_DEBUG=1 virsh restore sv-chubut-2011-04-04-17:38 2>
> > virsh-restore.log
> >
> > /*command which would enable the libvirt logging and output
> the debug
> > log into the virsh-restore.log file. This file could be sent to
> > the list
> > for analysis what's wrong.
> >
> > Thanks,
> > Michal
> >
> > On 04/05/2011 11:57 AM, Marcela Castro León wrote:
> > > Hello Daniel
> > > Thank you for all your information, but I still didn't
> solve the
> > > problem. I tried the option you mention, with two
> differents guest
> > > into two differents host, but all the cases I've got:
> > >
> > > */virsh # restore sv-chubut-2011-04-04-17:38/*
> > > */error: Failed to restore domain from
> sv-chubut-2011-04-04-17:38/*
> > > */error: monitor socket did not show up.: Connection
refused/*
> > >
> > > I cannot get any useful information (at least form me) on the
> > log you
> > > mention.
> > > I'd appreciate a lot a new suggestion.
> > > Thanks
> > > Marcela
> > >
> > >
> > >
> > >
> > > 2011/4/4 Daniel P. Berrange <berrange(a)redhat.com
> <mailto:berrange@redhat.com>
> > <mailto:berrange@redhat.com <mailto:berrange@redhat.com>>
> > > <mailto:berrange@redhat.com <mailto:berrange@redhat.com>
> <mailto:berrange@redhat.com <mailto:berrange@redhat.com>>>>
> > >
> > > On Sun, Apr 03, 2011 at 10:43:45AM +0200, Marcela Castro
> > León wrote:
> > > > Hello:
> > > > I need to know if I can use the restore operation
> (virsh o the
> > > equivalent in
> > > > libvirt) to recover a previous state of a guest, but
> recovered
> > > previously in
> > > > another host.
> > > > I did a test, but I got an error:
> > > >
> > > > The exactly sequence using virsh I testes is:
> > > > On [HOST SOURCE]: Using virsh
> > > > 1) save [domain] [file]
> > > > 2) restore file
> > > > 3) destroy [domain]
> > > >
> > > > On [HOST SOURCE] using ubuntu sh
> > > > 4) cp [guest.img] [guest.xml] [file] to HOST2
> > > >
> > > > On [HOST TARGET] using virsh
> > > > 5) define [guest.xml] (using image on destination in
> HOST2)
> > > > 6) restore [file]
> > >
> > > As a general rule you should only ever 'restore' from
a
> > > file *once*. This is because after the first restore
> > > operation, the guest may have made writes to its disk.
> > > Restoring a second time the guest OS will likely have
> > > an inconsistent view of the disk & will cause filesystem
> > > corruption.
> > >
> > > If you want to be able to restore from a saved image
> > > multiple times, you need to also take a snapshot of
> > > the disk image at the same time, and restore that
> > > snapshot when restoring the memory image.
> > >
> > >
> > > That aside, saving on one host & restoring on a
> > > different host is fine. So if you leave out steps
> > > 2+3 in your example above, then your data would
> > > still be safe.
> > >
> > > > The restore troughs the following message:
> > > > *virsh # restore sv-chubut-2011-04-01-09:58
> > > > error: Failed to restore domain from
> > sv-chubut-2011-04-01-09:58
> > > > error: monitor socket did not show up.: Connection
> refused*
> > >
> > > There is probably some configuration difference on
> your 2nd host
> > > that prevented the VM from starting up. If you're
> lucky the file
> > > /var/log/libvirt/qemu/$NAME.log will tell you more
> > >
> > > Daniel
> > > --
> > > |: http://berrange.com -o-
> > > http://www.flickr.com/photos/dberrange/ :|
> > > |: http://libvirt.org -o-
> > > http://virt-manager.org :|
> > > |: http://autobuild.org -o-
> > > http://search.cpan.org/~danberr/
> <http://search.cpan.org/%7Edanberr/>
> > <http://search.cpan.org/%7Edanberr/>
> > > <http://search.cpan.org/%7Edanberr/> :|
> > > |: http://entangle-photo.org -o-
> > > http://live.gnome.org/gtk-vnc :|
> > >
> > >
> > >
> > > --
> > > libvir-list mailing list
> > > libvir-list(a)redhat.com <mailto:libvir-list@redhat.com>
> <mailto:libvir-list@redhat.com <mailto:libvir-list@redhat.com>>
> > > https://www.redhat.com/mailman/listinfo/libvir-list
> >
> >
> > --
> > Michal Novotny <minovotn(a)redhat.com
> <mailto:minovotn@redhat.com> <mailto:minovotn@redhat.com
> <mailto:minovotn@redhat.com>>>,
> > RHCE
> > Virtualization Team (xen userspace), Red Hat
> >
> >
>
>
> --
> Michal Novotny <minovotn(a)redhat.com <mailto:minovotn@redhat.com>>,
> RHCE
> Virtualization Team (xen userspace), Red Hat
>
>
--
Michal Novotny <minovotn(a)redhat.com>, RHCE
Virtualization Team (xen userspace), Red Hat
11:01 a.m.
On 04/05/2011 05:59 PM, Marcela Castro León wrote:
Hello
The restore finished ok, I found at a log that was looking the image
at an "old" location. Anyway, the image don't start well, but I've to
revise all the configuration again, to see what's exactly is happening
now. Thank you a lot, I learned a lot.
Hi,
that's fine it passed however what do you mean by "the image don't start
well" ?
Michal
--
Michal Novotny <minovotn(a)redhat.com>, RHCE
Virtualization Team (xen userspace), Red Hat
11:07 a.m.
Yes, its say that recover all right and the guest is running, but I cannot
connect doing ssh, and in the virt-manager say "connecting to graphical
console".
I've just found in the log:
*radic@rionegro:/var/log/libvirt/qemu$ sudo cat chompi1.log *
*LC_ALL=C PATH=/usr/local/sbin:/usr/local/bin:/usr/bin:/usr/sbin:/sbin:/bin
QEMU_AUDIO_DRV=none /usr/bin/qemu-system-x86_64 -S -M pc-0.12 -enable-kvm -m
4096 -smp 4 -name chompi1 -uuid 22f013aa-25d4-b317-9eac-b232692db8c7
-chardev
socket,id=monitor,path=/var/lib/libvirt/qemu/chompi1.monitor,server,nowait
-monitor chardev:monitor -boot c -drive
file=/home/radic/mv/discoguest/chompi1.img,if=ide,index=0,boot=on,format=raw
-drive if=ide,media=cdrom,index=2,format=raw -net
nic,macaddr=52:54:00:82:ce:80,vlan=0,model=virtio,name=virtio.0 -net
tap,fd=33,vlan=0,name=tap.0 -chardev pty,id=serial0 -serial chardev:serial0
-parallel none -usb -vnc 127.0.0.1:0 -vga cirrus -incoming exec:cat *
*char device redirected to /dev/pts/1*
*cat: -: Permission denied*
*load of migration failed*
2011/4/5 Michal Novotny <minovotn(a)redhat.com>
On 04/05/2011 05:59 PM, Marcela Castro León wrote:
> Hello
> The restore finished ok, I found at a log that was looking the image
> at an "old" location. Anyway, the image don't start well, but I've
to
> revise all the configuration again, to see what's exactly is happening
> now. Thank you a lot, I learned a lot.
Hi,
that's fine it passed however what do you mean by "the image don't start
well" ?
Michal
--
Michal Novotny <minovotn(a)redhat.com>, RHCE
Virtualization Team (xen userspace), Red Hat
11:09 a.m.
On 04/05/2011 06:07 PM, Marcela Castro León wrote:
Yes, its say that recover all right and the guest is running, but I
cannot connect doing ssh, and in the virt-manager say "connecting to
graphical console".
I've just found in the log:
*radic@rionegro:/var/log/libvirt/qemu$ sudo cat chompi1.log *
*LC_ALL=C
PATH=/usr/local/sbin:/usr/local/bin:/usr/bin:/usr/sbin:/sbin:/bin
QEMU_AUDIO_DRV=none /usr/bin/qemu-system-x86_64 -S -M pc-0.12
-enable-kvm -m 4096 -smp 4 -name chompi1 -uuid
22f013aa-25d4-b317-9eac-b232692db8c7 -chardev
socket,id=monitor,path=/var/lib/libvirt/qemu/chompi1.monitor,server,nowait
-monitor chardev:monitor -boot c -drive
file=/home/radic/mv/discoguest/chompi1.img,if=ide,index=0,boot=on,format=raw
-drive if=ide,media=cdrom,index=2,format=raw -net
nic,macaddr=52:54:00:82:ce:80,vlan=0,model=virtio,name=virtio.0 -net
tap,fd=33,vlan=0,name=tap.0 -chardev pty,id=serial0 -serial
chardev:serial0 -parallel none -usb -vnc 127.0.0.1:0
<http://127.0.0.1:0> *
Well, I guess this is the problem since you have the VNC server bound to
127.0.0.1 which is localhost-only access. You have to change the libvirt
configuration to put 0.0.0.0 there instead to bind to all network
interfaces and not just loopback interface.
Michal
--
Michal Novotny <minovotn(a)redhat.com>, RHCE
Virtualization Team (xen userspace), Red Hat
5247
days inactive
5252
days old
22 comments
4 participants
participants (4)
-
Daniel P. Berrange -
Eric Blake -
Marcela Castro León -
Michal Novotny