Upgrading from RMM 6.0.x to RMM 7.x - Best Practice
I am undertaking an upgrade of RMM 6.0.6 to RMM 7.0.1. This upgrade faces me with some unpleasant decisions. Because the 6.0.6 environment has a CCM/EWM instance, I am faced with the following choices upon upgrading to 7.0.1:
- Do not upgrade RMM. Upgrade EWM 6.0.6 to 7.0.1, then install RMM on top of the upgraded EWM 7.0.1 instance. RMM 7.0.1 would be a blank slate. EWM data would be upgraded.
- Upgrade RMM and have two EWM instances as part of the 7.0.1 ELM solution. The reason this is the case is because there is no merge workflow supported for merging an existing EWM 6.0.x and RMM 6.0.x instances into a combined EWM/RMM 7.x instance. I would have the upgraded EWM instance with upgraded EWM data, as well as an upgraded RMM instance, now an EWM instance, thus causing the overall solution to have two competing EWM instances.
I wanted to see if anyone had any other ideas or advice as I undertake this activity?
Thanks!
One answer
The only answer is #2.
In what sense are the two EWM instances "competing"? In particular, in what way do they compete any more than the 6.x RMM instance competed with the 6.x RTC instance?
Note that many customers have multiple 6.x RTC instances, and then after an upgrade to 7.x, have multiple EWM instances (and do not expect those RTC instances to somehow be merged into one EWM instance as part of the upgrade).
Of course, a "server merge" would be a nice feature to have (once one figured out how to avoid all data conflicts that might occur when merging the data to a single server, as well as how to redirect all the incoming URL links to the original server to end up at the merged server). But this is not a feature that somehow comes for free when performing an upgrade.
Comments
Thank you for your response, Geoffrey.
I should have added the detail that this development deployment has all the applications running on the same physical server. (This is why I used the term "competing" in reference to the dual EWM instances.)
In general, for performance, it is the size of your workload that matters, rather than the number of servers. So having two servers each doing half of the workload uses a similar amount of resources as one server doing the whole workload.