I have two CA API Gateway clusters[A and B] at two different datacenters. Cluster A has been operational for past two years. Cluster B has been setup recently. All the gateway entities from cluster A was migrated to cluster B using GMU[migratein -all].
My questions are
1.How do I validate if both the cluster are absolutely in sync? What is the approach that I need to follow to validate this?
2.Cluster A keeps having constant changes.How to ensure that these delta changes gets pushed to cluster B?
I don't think there is a super easy way to see if two different clusters are in sync with each other - since they're separate clusters for a reason. However, I suspect you could do something like export the ssg database to a SQL file, compare it's size, then do the same thing on the other cluster. If the databases were synced, I believe the file sizes should be very similar to each other as well, if not identical to each other. However, this also only counts for the database (which is where most of the data is held anyways), you'd still need to manually compare the system.properties and such to ensure they are all setup with the same system variables, same goes for appliancedefs.sh in case any memory management changes were made, and more.
Can you please expand on your second question? Are you referring to just the database level of detail, or everything including the system properties, etc.? If just the database, then I think you can simply have MySQL forward-only data to cluster B's database node, outside of the actual Gateway application itself. If you're looking at other local files, this may prove to be more difficult but then you could perhaps look to use something like rsync (this would not be a supported method though - there isn't really any supported method that I'm currently aware of for this particular task). I'd need to dive further to see if there is something I'm missing, as there may be a better way to do this than what I'm thinking of right now.
Ultimately though, it sounds like you want to setup Cluster A to replicate to Cluster B. Can I ask, would Cluster B effectively be used as a failover for Cluster A, in a different data centre perhaps? Setting up a second cluster to essentially be a "hot standby” of the first cluster can be pretty complicated and tends to go a bit outside the scope of what the Gateway application does out of the box. We can always look to get CA Services involved if you'd like though too, as they typically specialize in this sort of thing.
If you're perhaps able to expand on what you're ultimately trying to achieve, I think we can perhaps try to look at this a bit differently and hopefully find a more supported method of achieving your use case.
Were you able to resolve the issue? What was the final result?
Stephen HughesBroadcom Support