Does services.sh restart cause the vmnics to revert back to the highest priority vmnic in a vmkernel adapter?

[bhanekom]

I was just qondering if anyone could confirm wheteher or not running the services.sh restart on a host will force the host to revert back to it highest priority/active vmnic in a vmkernel. We have a situation where it is bets to tell the vmkernel to not failback to another nic (small business switches which seem to inform devices it is available then fail again whilst restartarting in a stack causing the vmnics to flap between ports), so I was just curiosu as to whether running the command has any impact on whcih vmnic is in use, specifically if it will force the vmkernel to switch to the vmnic marked as active (from the vmnic marked as standyby).

[Nick_Andreev]

Standby uplink is used only if active has failed. You may not worry, restarting services.sh won't trigger it.

[bhanekom]

Thanks Nick, Appreciate the clarification on that point. What I am trying to achieve is a simple way of forcing the VMNics back to their assigned active NIC (highest Priority) after a failure of a physical swtch which it is connected to. We have had to set the NIC to FAILBACK = NO as the Switch Stack "flaps" between avialable and not available whilst busy powering back on causing the VMNics to switch between the active and standby NICs too frequently.

[Nick_Andreev]

Ok, now I see what you mean. There is no such thing as a highest priority NIC, though. If you disable failback, it will not fail back. NICs just change from active/standby order to standby/active. I'm not aware of any other way to change it back other than manually reconfiguring it.

Why is it not an option though? Switch reboot is a rare event. Shouldn't hurt to change it manually once in a blue moon. Or do you have many hosts and you want to automate the process?

[bhanekom]

Thanks NIck,

Manual reconfiguration is an option. We only have 3 hosts to deal with so not an issue with that. The reason I was looking for something to make it simpler was that the people who will be called out when a failure happens have no previous experience with VMWare (or any virtualisation) as it is hosting the core product they are skilled at working on, so I was trying to find a way to make their lives a bit easier for the rare occasion this needs to happen. It has only really been an issue at this point because we are in the Process of Factory Acceptance Testing the system and there have been multiple switch "failures" in a day.

[Nick_Andreev]

Then I can suggest writing a PowerCLI script. Changing NIC teaming using PowerCLI would be no longer than 10 lines of code.

[bhanekom]

Thanks Nick. I have limieted PowerCLI experience, but will definitely look into this. I appreciate you taking the time to assist.