Xen users - Aug 2009 - What can collapse all the domU????????''

Hi.

I have a dom0 running etch and 2 domU:  my webserver and mailserver.

In the lastest 3 months i was experimenting a complete collapse in all
the domU, forcing us to reboot the dom0 to restore the virtual
servers.  The latest issue was at the weekend, and the company was
without email and website for 2 days.


Now my boss have some doubts about if the virtualized environments are
really stables, and im trying to demostrate if the problem was caused
for the some dom0 failture, or if really can exist some problem with
my virtualitation scenarie.

Im searching in all the logs, and the only strange issue founded is a
lot strange dns request in the bind server (installed in dom0).

My question is:  What can collapse all the domU servers?  is possible
install some package or anything to prevent or detect any possible
problem for the domU''s and give a faster answer to prevent a collapse?

Thanks in advance.

_______________________________________________
Xen-users mailing list
Xen-users@lists.xensource.com
http://lists.xensource.com/xen-users

On Wed, Aug 05, 2009 at 07:23:33AM -0500, kazabe wrote:
> Hi.
> 
> I have a dom0 running etch and 2 domU:  my webserver and mailserver.
> 
What Xen version? What dom0 kernel version? Installed from sources, or from
some debian packages? 
> In the lastest 3 months i was experimenting a complete collapse in all
> the domU, forcing us to reboot the dom0 to restore the virtual
> servers.  The latest issue was at the weekend, and the company was
> without email and website for 2 days.
> 
Definitely something wrong here. Either you''re running buggy version of
Xen or
dmo0 kernel, or then your hardware if faulty.
> 
> Now my boss have some doubts about if the virtualized environments are
> really stables, and im trying to demostrate if the problem was caused
> for the some dom0 failture, or if really can exist some problem with
> my virtualitation scenarie.
> 
> Im searching in all the logs, and the only strange issue founded is a
> lot strange dns request in the bind server (installed in dom0).
> 
> My question is:  What can collapse all the domU servers?  is possible
> install some package or anything to prevent or detect any possible
> problem for the domU''s and give a faster answer to prevent a
collapse?
> 
Well you should check logs from dom0: dom0 kernel dmesg, /var/log/messages,
/var/log/syslog, "xm dmesg", "xm log".

Also, check all the logs from domUs. 

And monitor CPU/memory/disk usage from dom0, and also from all the domUs.

-- Pasi

_______________________________________________
Xen-users mailing list
Xen-users@lists.xensource.com
http://lists.xensource.com/xen-users

On Wed, Aug 5, 2009 at 7:23 PM, kazabe<kazabe@gmail.com>
wrote:
> Hi.
>
> I have a dom0 running etch and 2 domU:  my webserver and mailserver.
>
> In the lastest 3 months i was experimenting a complete collapse in all
> the domU, forcing us to reboot the dom0 to restore the virtual
> servers.  The latest issue was at the weekend, and the company was
> without email and website for 2 days.
You didn''t mention what "complete collapse" is. Is it hang?
Is it
kernel panic? Is it slow response? Is it rebooting?
What does the system console (which usually is the best source of
information) say when it happens?

Sometimes the problem is as simple as a faulty external storage.
>
>
> Now my boss have some doubts about if the virtualized environments are
> really stables, and im trying to demostrate if the problem was caused
> for the some dom0 failture, or if really can exist some problem with
> my virtualitation scenarie.
Well, if you need support, sticking with vendor-supported solutions
(RHEL for example) is a good idea. They''re generally better tested and
can help you in certain problems.

-- 
Fajar

_______________________________________________
Xen-users mailing list
Xen-users@lists.xensource.com
http://lists.xensource.com/xen-users