Replace a node by moving the drives to another chassis

[tonymcmillan]

We are having trouble with a SimpliVity node that keeps failing.

We would like to replace the entire node without performing any new node installation. So, we want to move the boot and storage drives to a new, identical, chassis with the expectation that the node will work like before, minus the failures. 

Has anyone done this before or have a reason to believe this would NOT work?

[DaveOb]

No the chassis node is linked to the server serial numer at deployment within the db and on the node itself,Theoretically you could start hacking but it would be a totally non documented and unsupported thing to do and personally I would not trust it to run production workloads even if I did get it to work.

[tonymcmillan]

Dave,

Thanks for your qjuick reply. 

Since we are a reseller, I'm curious to know the impact for us in the future if a mainboard needs to be replaced in a node. This sounds like it would be far from a straighforward part swap. So, we'd have some serious work do with adding and ejecting SimpliVity nodes?

Is there also any reliance on the S/N of the accelerator card?

Best,

Tony

[Stor_Mort]

Hi Tony,

The disk signatures are based on the raid controller WWN and the system GUID is based on the serial number. Changing either of those will usually cause the node to stop working. HPE Support can repair those inconsistencies. If you don't have HPE Support contract, the only way to bring the node back online is to do a factory reset, rejoin the cluster, and resync with the other cluster nodes.