El sáb, 21-06-2008 a las 07:27 +0200, Michael Schwartzkopff escribió:
> Hi,

Hi Michael,

Firstly, thanks for answering my questions. :)

> strange.
> Have you ever tries to configure three separate IPaddr2-clones? Do they run? 

Yes, I have set up a new different configuration. The previous one was
not reliable with network connectivity. Now I have two IPaddr2 clones
(the third HA IPaddr2 is a virtual address of 10.20.1.12). The problems
have not aroused so far. They appear from time to time.

> What about running the iptables command manually on the node without the 
> cluster?

No problem, there. As I say previously no problem was found when
introducing iptables rules by hand.

> 
> Please correct me, but when I am looking through your CIB you have three 
> instance atribute sections for your server_ip, which all set ip to 
> 10.20.1.12. Please reduce that to 1 instance attribute section.

Thanks, you're right. I haven't noticed this misleading configuration.

> Why do your have a start and a stop operation all resources? Start and stop 
> is 
> done the cluster. You should only have a monitor operation.

I have read from clusterlabs documentation [1] to set the start/stop
operation to specify a fine-grained timeout or other attributes.

> You also have two instace_attribute secitons in networking_ip and alcampo_ip.

Again, thanks so much!

> Any changes in behaviour?

No problem until now.

Other question, if there is a network connectivity problem such as
network card/cable failure, IPaddr2 resource does not check these
problems to migrate to other nodes, doesn't it?.

I think I must set up `pingd` configuration. Am I right?

Thanks so much for your suggestions!


[1] http://clusterlabs.org/mw/Image:Configuration_Explained.pdf

_______________________________________________
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