Re: GNOME Servers were down

[Owen Taylor <otaylor redhat com>]

On Thu, 2010-04-29 at 16:51 +0800, Ray Wang wrote:
> Hi Owen,
> 
> Today, we experienced many gnome servers were down, and Olav figure
> out that label or other servers were rebooted, that lead to many
> services were not available, for example, git, bugzilla, puppet0, wiki
> and NFS etc.
> 
> Olav guided me that bugzilla is on vbox as a VM, and I added him on
> vbox, we figured out that those VMs will not able to start when
> domain0 is rebooted, the configuration files are not in /etc/xen/auto,
> we are not sure you have done with these on vbox, but please notice
> that and make sure those necessary services can start when host
> restart?  We left other services to you, and please decide which VM
> should be started, thanks a lot! :)
> 
> I just wonder what VMs on combobox, are they KVM based? how can we
> start them if the host is down?

On both combobox and vbox, the primary way we should be manipulating VMs
is through 'virsh'.

 # list running domains
 virsh list 

 # list running and not running domains
 virsh list --all

 # start a domain
 virsh start <domain>

 # stop a domain
 virsh shutdown <domain>

 # force-stop a domain
 virsh destroy <domain>

 # get information about a domain
 virsh dominfo <domain>

 # Set a domain to autostart
 virsh autostart <domain>

 # unset a domain to autostart
 virsh autostart --disable <domain>

I've now set all the domains that should be autostarted to autostart
(the exception is the 'online' domain on vbox, which is the old
mugshot.org, and shouldn't be started.)

So hopefully things should be better next time the machines get
rebooted.

> For the record, label has a fault hard drive, so that raid1 is failed.

Yes, it's known.

- Owen