I'm using an application in a global cluster environment formed by 2 mini-cluster with one node each one, configured by a supplier.
this application is using 2 NICs: one for the LAN of the management of the servers and one for the LAN of the management of the network elements.
The supplier only configured a resource associated to the NIC of the management so, in case of a fault of this NIC the application can switch, while in case of a fault of the NIC of the network element the application freezes.
This behaviour is obviously unacceptable, but the answer of the supplier was: the application doesn't support this feature (2 NIC resources together, apart from the heartbeat NIC).
Is this possible ?
I know that you cannot know how the application works, but I think that the configuration of the NIC resources is totally indipendent of the application itself, in whatever way it works.
Is this correct ?
thanks in advance and BR
Sorry for the late reply. As you said we cannot know how the applications work but if you want to can configure another NIC resource in the service group. That just means VCS will monitor it and depending on how you configure it, it could switch the service group for you.
no problem for the delay.
Thanks for your answer.
Effectively in a not production environment I added another NIC resource and it worked.
The question was that the supplier did not support me probably because he did not want to vaildate it (due to a lot of effort, probably)
But, in theory, a monitoring of a NIC resource should not influence in any case a sotware application, expecially if the hint of the supplier was:
manually switch the resource group in case of failure of the additional NIC resource.
thanks and best regards.
Like i said, its really up to you, you can make the NIC non-critical meaning it won't affect the service group if you want to, or you can make it failover the group. VCS is flexible :)
good idea, I can add a second NIC resource as NOT critical resource and in the future only, after a supplier approval, I can transform it in a critical resource.
So I can identify a NIC failure immediately and then switch the resource group manually
Bye and thanks a lot.