Re: [Xen-API] Xapi problem

[George Shuklin <george.shuklin@xxxxxxxxx>]

I'd like to say it's pretty often problem. When something going wrong, 
xapi, squeezed, different kernel daemons start to flood messages. The 
logrotation is not very helpful in this case and host cease action.
Especially this dramatic when space is ran out on pool master.
Classic situation is storage error (f.e. list loose to some iscsi SR).

Right now we creates and mounts separate partition for /tmp (where 
kernels extracted by pygrub), /var/log (where flood occur) and /var/xapi 
(where messages and state.db stored). This allows to continue host 
operations even in case of 'log partition filled' situation.
We tried to hack XCP installation to do this in install time but it very 
configurable, so we do this in post installation stage.
And we do have some 'handmade' monitor reporting us via mail about space 
running out. From my statistic under heavy load (few thousands VM) it 
happens about 2-3 hosts per week.


On 03.09.2011 00:51, Nycko wrote:
> Hi, this is the second time I step.
> After that I filled the dom0 disk, and freed space, xapi stopped
> working, no xe command works, I can not restart xapi (for the same
> problem).
> The first one I could fix it then restart the server, I can not
> because I have a vm in production. Any ideas on this?
>
> Saludos
_______________________________________________
xen-api mailing list
xen-api@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/mailman/listinfo/xen-api