[prev in list] [next in list] [prev in thread] [next in thread] 

List:       linux-ha-dev
Subject:    Re: [Linux-ha-dev] serializing actions of resource scripts
From:       Andrew Beekhof <beekhof () gmail ! com>
Date:       2007-11-20 12:29:54
Message-ID: 171B5901-4E41-4151-9AF4-4F0F54D37078 () gmail ! com
[Download RAW message or body]


On Nov 20, 2007, at 1:25 PM, Sebastian Reitenbach wrote:

> Andrew Beekhof <beekhof@gmail.com> wrote:
>>
>> On Nov 20, 2007, at 12:43 PM, Sebastian Reitenbach wrote:
>>
>>> Hi,
>>>
>>> Andrew Beekhof <beekhof@gmail.com> wrote:
>>>>
>>>> On Nov 20, 2007, at 10:38 AM, Sebastian Reitenbach wrote:
>>>>
>>>>> Hi,
>>>>> Andrew Beekhof <beekhof@gmail.com> wrote:
>>>>>>>>
>>>>>>>> that we cant do - at the crm level its cluster-wide or not at  
>>>>>>>> all
>>>>> Anyway, a cluster wide solution on the crm level would work for  
>>>>> me.
>>>>
>>>> that what rsc_order constraints are for.
>>> Then, how would I configure that? I want to start/stop Xen domains  
>>> in
>>> arbitrary order, but I cannot seem to get it to work with rsc_order?
>>> I tried rsc_orders with symmetrical=no, and a score of -INFINITY and
>>> also 0,
>>> but I don't get it that flexible. The best success I had with these
>>> rsc_orders:
>>>
>>> <constraints>
>>>  <rsc_order id="HOME_before_PUBLIC" from="NFS_HOME" type="before"
>>> to="NFS_PUBLIC" symmetrical="no" score="-INFINITY"/>
>>>  <rsc_order id="PUBLIC_before_FTP" from="NFS_PUBLIC" type="before"
>>> to="FTP" symmetrical="no" score="-INFINITY"/>
>>>  <rsc_order id="FTP_before_LDAP" from="FTP" type="before" to="LDAP"
>>> symmetrical="no" score="-INFINITY"/>
>>> </constraints>
>>>
>>> But when I had e.g. NFS_PUBLIC target_role=stopped, all the other
>>> three had
>>> target_role=started. Then marked the node active,  the NFS_HOME
>>> started, as
>>> expected, but not the FTP and LDAP domU's.
>>>
>>> With a score="0", the xend got overloaded, and the startup of the  
>>> xen
>>> domains failed at all.
>>
>> 0 should be the right score here.
>> what version do you have?  i'd recommend the latest interim build
>> which had a bunch of cleanups in this regard.
> It is a 2.1.2-20.1 on suse 10.3 x86_64.

ah, excellent

>
> I had the constraints configured with a score=0, all Xen domains were
> target_role started. Then I changed the node from standby to active.
> I saw "xm list" hanging for a while, I think it was too busy,  
> therefore I
> assume it tried to start all resources at the same time.

its possible
you might want to increase the timeouts anyway

> I could do it again, and run hb_report afterwards, if that would help.

yes please

_______________________________________________________
Linux-HA-Dev: Linux-HA-Dev@lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/
[prev in list] [next in list] [prev in thread] [next in thread]