Thanks Andrew and Ivan,
 by combining your tips I got rid of this restart issue.

My quite basic configuration is working now with following constraints like:
 <rsc_order id="imagestore_before_vm-skaodc1" from="vm-skaodc1"
to="imagestorecloneset" score="0"/>
 <rsc_order id="configstore_before_vm-skaodc1" from="vm-skaodc1"
to="configstorecloneset" score="0"/>
 <rsc_colocation id="skaodc1-with-configstore" to="configstorecloneset"
from="vm-skaodc1" score="INFINITY"/>
 <rsc_colocation id="skaodc1-with-imagestore" to="imagestorecloneset"
from="vm-skaodc1" score="INFINITY"/>

Cheers,
Petri

Andrew Beekhof wrote:
> On Thu, Oct 30, 2008 at 21:54, Petri Asikainen <[EMAIL PROTECTED]> wrote:
>   
>> Thanks for tip. ( I posted cib.xml earlier today 081030  at 8:09)
>> Do you mean that I should add  score to existing order constraints
>> like "imagestore_before_vm" or create new ones?
>>     
>
> add to the existing ones.
> also, you'll want to add colocation constraints for the VMs: eg.
>
>        <rsc_colocation id="skaodc1-with-configstore"
> to="configstorecloneset" from="vm-skaoas1" score="INFINITY"/>
>        <rsc_colocation id="skaodc1-with-imagestore"
> to="imagestorecloneset" from="vm-skaodc1" score="INFINITY"/>
>
>   
>> Ivan wrote:
>>     
>>> Without the CIB it's hard to tell but I had fairly similar issue with
>>> the SLES 2.1.3 version which needed "score=0" for all resources as order
>>> constraint and all came right after that.
>>>
>>> TA,
>>> Ivan
>>>
>>>
>>> On Thu, 2008-10-30 at 18:51 +0200, Petri Asikainen wrote:
>>>
>>>       
>>>> And heres part of /var/log/messages from time when rebooted node xen2
>>>> comes up.
>>>> Canf find reason why resources on xen1 are restarted...
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> Andrew Beekhof wrote:
>>>>
>>>>         
>>>>> version? logs? config?
>>>>>
>>>>> On Thu, Oct 30, 2008 at 06:53, Petri Asikainen <[EMAIL PROTECTED]> wrote:
>>>>>
>>>>>
>>>>>           
>>>>>> I have setup basic two node xen domu cluster with iscsi san.
>>>>>> Nodes are communicating over two separate ethernet segments using
>>>>>> separate nics for each connection.
>>>>>> Nodes are using shared ocfs2 storage for domu settings and disk-images.
>>>>>> Xen storage and domus are configured as cluster resources.
>>>>>>
>>>>>> When reboot other node and it comes up, all resources alredy running on
>>>>>> second node are restarted.
>>>>>>
>>>>>> Could some one point me where I could look for, or what could cause this?
>>>>>>
>>>>>> I tried even to set no-quorum-policy to freeze and startup-fencing to
>>>>>> false , but either is helping.
>>>>>>
>>>>>>
>>>>>> I'll post cib.xml on request. ;)
>>>>>>
>>>>>> Regards,
>>>>>>
>>>>>> Petri Asikainen
>>>>>>
>>>>>> _______________________________________________
>>>>>> Linux-HA mailing list
>>>>>> [email protected]
>>>>>> http://lists.linux-ha.org/mailman/listinfo/linux-ha
>>>>>> See also: http://linux-ha.org/ReportingProblems
>>>>>>
>>>>>>
>>>>>>
>>>>>>             
>>>>> _______________________________________________
>>>>> Linux-HA mailing list
>>>>> [email protected]
>>>>> http://lists.linux-ha.org/mailman/listinfo/linux-ha
>>>>> See also: http://linux-ha.org/ReportingProblems
>>>>>
>>>>>
>>>>>
>>>>>           
>>>> _______________________________________________
>>>> Linux-HA mailing list
>>>> [email protected]
>>>> http://lists.linux-ha.org/mailman/listinfo/linux-ha
>>>> See also: http://linux-ha.org/ReportingProblems
>>>>
>>>>         
>>> _______________________________________________
>>> Linux-HA mailing list
>>> [email protected]
>>> http://lists.linux-ha.org/mailman/listinfo/linux-ha
>>> See also: http://linux-ha.org/ReportingProblems
>>>
>>>
>>>       
>> _______________________________________________
>> Linux-HA mailing list
>> [email protected]
>> http://lists.linux-ha.org/mailman/listinfo/linux-ha
>> See also: http://linux-ha.org/ReportingProblems
>>
>>     
> _______________________________________________
> Linux-HA mailing list
> [email protected]
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems
>
>   

_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Reply via email to