I have a question about MCC FC Upgrade sequence

[wzndragon1]

My customer has a 4-node AFF700 metrocluster FC to upgrade the operating system,9.5P14 to 9.7P21,

node 0185---node 0186   HA  pair.

node 0187---node 0188   HA  pair

--------------------------------------------------

node 0185--node 0188 DR partner

node 0186--node 0187 DR  partner

---------------------------------------------

       I know that according to the standard document, the MCC FC upgrade sequence should be node 0185-0188 sequential or parallel upgrade,Next, node  0186-0187 is upgraded sequentially or in parallel.

      Due to some special reasons, the customer requires the upgrade sequence to be 0185-0187-0186-0188,My question is whether this upgrade is okay,What risks will it bring?

[akiendl]

You should stick to the order that is documented / that ANDU (Automated None Disruptive Upgrade does). 0185 and 0186 will be upgraded at the same time then 0187 and 0188 at the same time. This is to ensure that at any time the DR relationship is in order. Meaning you do not loose DR capability in the event of a disaster during an ONTAP upgrade.

ak.

[wzndragon1]

Hi,akiendl:

     Thank you for your answer,I guess you may have made a mistake,

node 0185--node 0188 DR partner

node 0186--node 0187 DR  partner

         According to document,0185 and 0188 will be upgraded at the same time then 0186 and 0187 at the same time.There's really no way now,If you upgrade in the wrong order, , if a site level disaster occurs, you will lose DR capabilities?If I am lucky enough, this upgrade is also possible?

PREVIEW

[akiendl · Accepted solution]

Sorry misread that. Yes, it is always local node and dr partner updating at the same time.
As said if DR and HA are on different levels things get really complicated when a disaster strikes.

ak.