Hi,
I solved the problem by using the hb_gui on the other node of the cluster.

In that case the IPaddr resource shows the correct ocf/heartbeat 
CLASS/PROVIDER (on the other node only heartbeat was reported) and the 
new RG with the new IP are starting correctly.

In effect seems a bug in the hb_gui...someone had the same problem on 
DEBIAN 2.1.3?
In your opinion can be a problem affecting the cluster functionality or 
only the management?

Thank you
Bye
Alessandra


On 10/28/2011 12:32 AM, Alessandra Giovanardi wrote:
> On 10/28/2011 12:12 AM, Andreas Kurz wrote:
>> Hello,
>>
>> On 10/27/2011 11:54 PM, Alessandra Giovanardi wrote:
>>> Hi,
>>> I followed your suggestion and all was ok thanks...both nodes update
>>> correclty your configuration and takeover works fine.
>>>
>>> Anyway, after those operations, we tried to add another RG to our
>>> cluster, named group_univdrupal_prod with only one IP and we had some
>>> problem.
>>> The RG is correctly created but at the start (we disabled stonith first)
>>> an error occurs. I paste you the log at the end.
>>>
>>> Why the resource_univdrupal_IP_1 seems unmanaged?
>> For any reason it was unable to start and then to stop the IP on
>> gicdrupal01 ...
>>
> Quite strange since if I start this IP on gicdrupal01 with:
> ifconfig eth1:0 130.186.99.43 netmask 255.255.255.0 up
>
> the IP is correctly configured on the interface...
> So it seems a heartbeat problem....
>
>>> The only strange think I observe is that the IP resource still present
>>> into the first RG was heartbeat::ocf:IPaddr, while the new is created as:
>>> heartbeat:IPaddr
>> These are different resource-agents ... use heartbeat:ocf:IPaddr2 ...
>> yes, the one with the "2" at the end ;-)
>>
> I tried also IPaddr2 without success (also in that case the resource is
> created without the :ocf: field-->  why?, from hb_gui the only choices
> are IPaddr and IPaddr2 with "hearbeat" and not "ocf/heartbeat" as
> Class/Provider:
>
>    pengine[27190]: 2011/10/27_22:35:53 notice: native_print:
> resource_univdrupal_IP_1    (heartbeat:IPaddr2):    Stopped
> pengine[27190]: 2011/10/27_22:35:53 notice: native_print:
> resource_univdrupal_IP_1    (heartbeat:IPaddr2):    Stopped
> pengine[27190]: 2011/10/27_22:35:54 notice: native_print:
> resource_univdrupal_IP_1    (heartbeat:IPaddr2):    Stopped
> pengine[27190]: 2011/10/27_22:36:05 notice: native_print:
> resource_univdrupal_IP_1    (heartbeat:IPaddr2):    Stopped
> pengine[27190]: 2011/10/27_22:36:05 notice: native_print:
> resource_univdrupal_IP_1    (heartbeat:IPaddr2):    Started gicdrupal01
> FAILED
> pengine[27190]: 2011/10/27_22:36:06 notice: native_print:
> resource_univdrupal_IP_1    (heartbeat:IPaddr2):    Started gicdrupal01
> (unmanaged) FAILED
> pengine[27190]: 2011/10/27_22:36:14 notice: native_print:
> resource_univdrupal_IP_1    (heartbeat:IPaddr2):    Started gicdrupal01
> (unmanaged) FAILED
> pengine[27190]: 2011/10/27_22:36:18 notice: native_print:
> resource_univdrupal_IP_1    (heartbeat:IPaddr2):    Stopped
> pengine[27190]: 2011/10/27_22:36:21 notice: native_print:
> resource_univdrupal_IP_1    (heartbeat:IPaddr2):    Started gicdrupal01
> FAILED
> pengine[27190]: 2011/10/27_22:36:22 notice: native_print:
> resource_univdrupal_IP_1    (heartbeat:IPaddr2):    Started gicdrupal01
> (unmanaged) FAILE
>
> It sound me like a problem of the hb_gui, which does not compromise the
> cluster features, but does not permit the RG creation...
> Some other suggestions to by-pass the problem?
>>> even if I select from the hb_GUI the IPaddr heartbeat...
>>> Furthermore the heartbeat release version seems the last available into
>>> the DEBIAN lenny stable repository:
>>>
>>> ii  heartbeat
>>> 2.1.3-6lenny4              Subsystem for High-Availability Linux
>>> ii  heartbeat-2
>>> 2.1.3-6lenny4              Subsystem for High-Availability Linux
>>> ii  heartbeat-2-gui
>>> 2.1.3-6lenny4              Provides a gui interface to manage heartbeat
>>> clusters
>>> ii  heartbeat-gui                              2.1.3-6lenny4
>> really, really, really consider an upgrade ... even if you are on lenny,
>> use latest pacemaker backports packages.
>>
>> Regards,
>> Andreas
>>
> Is quite complicated for us ;-)
>
> Thanks
> A.
>


-- 
Alessandra Giovanardi
e-mail [email protected]
Cineca - DSET http://www.cineca.it
Phone +39 051 6171939

_______________________________________________
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