>> On 06/10/2010 07:04 PM, Caspar Smit wrote:
>>> Don't get me wrong, the failover IS WORKING. Only the connection
>>> re-instatement to MS iSCSI Initiator isn't when using the RAs.
>>
>> Can you try setting an explicit SCSI ID and SN on your logical units
>> please, and see if that improves the situation? As in:
>>
>> primitive res_lu_lun1 ocf:heartbeat:iSCSILogicalUnit \
>>   params target_iqn="iqn.2001-04.com.example:storage.example.iscsivg01"
>> lun="1" \
>>   path="/dev/iscsivg01/lun1" \
>>   scsi_id="iscsivg01.lun1" scsi_sn="4711"
>
> I already tested this, but did a re-check to no avail.
> Still only node01->node02 failover works.
>
> I thought maybe the DC role of pacemaker had something to do with it
> because when node01 is DC the services (failover-ip, iscsitarget,
> iscsilogicalunit, portblock) start/stop almost instantly on node1 (when
> monitoring with crm_mon) start/stop on node02 takes a second each.
> Switching DC didn't change anything.
>
> Still only node01->node02 failover works.
>
> Now to test open-iscsi.

Ok, I can now DEFINATLY say this is an MS iSCSI Initiator issue.
I tested with open-iscsi and now the connection re-instatement works both
ways, tested like 10 failovers and no disconnect! :-)

Used same pacemaker config like before.

Still weird why MS initiator only works one way. Guess I have to ditch MS
in my setup.

Kind regards,
Caspar

>
> Kind regards,
> Caspar Smit
>
>>
>> Thanks.
>>
>> Cheers,
>> Florian
>>
>> _______________________________________________________
>> Linux-HA-Dev: [email protected]
>> http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
>> Home Page: http://linux-ha.org/
>>
>
> _______________________________________________________
> Linux-HA-Dev: [email protected]
> http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
> Home Page: http://linux-ha.org/
>

_______________________________________________________
Linux-HA-Dev: [email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/

Reply via email to