arl from a300 to a400

[ChristianBossi]

Hi all, we have to perform an Head Swap from A300 to A400.

I have seen that a fundamental prerequisite is to have at least one common port for cluster networks.(from old node to new one)

On a300, cluster network are on e0a and e0b.

instead on a400 are on e3a and e3b (100 GB).

so i need to use a temporary cluster network interface on it.

the problem is that on a400 e0a and e0b are used only for ha nvram and must be connected together (e0a node1-->e0anode2,e0bnode1--->e0b node2).

seen that during head swap procedure with arl, ha is disable, is it possible to use e0a and e0b as a temporat cluster network interface?.

From a400 sysconfig i saw that e0a and e0b are 10/25/100 gb network, so I think they can be used as a cluster network.

any suggestions

many thanks for your time

Christian

[AlexDawson]

Hi there!

Which document are you using to prepare for this task? I think the requirement means that all four systems should be on the same cluster switch, not that you must use e0a on all of the nodes, but I admit I could be wrong on that one. I have reviewed this document and can not see it in there

Has a cluster switch been made available for the process?

In relation to your question, no, my understanding is that on the A400, those ports cannot be used for cluster interconnect at any time.

[aborzenkov]

I vaguely remember similar problem I had during ARL head swap. Old and new head had different Ethernet ports, and after head swap new node inherits configuration of old head so it had no working cluster ports. I had to manually adjust configuration before node could join the cluster, but I do not remember how.

For next head swap with this customer I configured cluster ports in advance to ensure there was some overlap so at least one working cluster port was available after head swap.

[ttran · Accepted solution]

Hi ChristianBossi,

The below guide can be used to upgrade controllers of different platforms even though the versions and specific platforms you mentioned are not the same, but the process is essentially the same. Specifically, page 8 discusses preparing the ports of the new controller for the Cluster LIF.

Upgrading controller hardware in a pair of nodes running clustered Data ONTAP 8.3 by moving storage 

Regards,

Team NetApp

[hilke_p]

Sadly the document is no longer available. Is there a new Link for that?

[AlexDawson]

Hi there! Sorry for the delay in response - probably going to be more useful to someone else at this point, but start looking at https://kb.netapp.com/onprem%2Fontap%2Fhardware%2FController_hardware_upgrade_for_AFF_and_FAS_systems_resolution_guide to find the most appropriate guide! 

Hope this helps.