I'm trying to figure out if it's possible to customize the behavior of SRM/SRA. My scenario is I have a 3PAR T800 at my prod and DR sites replicating. When I try to run a test recovery plan and I tell it to replicate changes it sits there for hours while it waits and on the array we are running into a snowball effect where some groups are taking longer than the specified window, and the 3PAR rcopy only allows 20 concurrent LUNs to replicate at a given time. So what we run into is when any of the 20 finish the times are snowballing and causing the just finished replicated LUN to go back into a synching state since by the time it finishes the 30 minute timer has passed so it thinks it's time to replicate again. So basically at no time do the LUNs ever sit in a "Syned" state so in a "replicate changes" mode of a test recovery plan it sits there forever, or if you tell it to not replicate changes it immediately fails because it can't snapshot the LUN when it's in a syncing state. I thought SRM would stop replication in a test scenario when you tell it to not replicate changes, but come to find out apparently that's not the case. The only work around we can get from VMware or 3PAR is to either change the replication window to a larger window so they complete, which of course we can't do since that would violate our RPO times, increase the network pipe, or stop replication manually. We are doing the last one. My ultimate question is this: Is it possible to run a custom command with SRM to tell the SRA to pause replication so the sync stops and it's able to get a snapshot of the LUN with no errors, then restart the sync once it gets to the end of the job?
I found the link below which has some commands but not many, have you spoken to HP about this, I was under the impression that the storage vendor writes the SRA and VMware certifies it, but I could be wrong.
I have a document from HP that shows what the SRA commands are. I had VMware and HP on the phone yesterday talking about it and the level 2 guys from both sides couldn't give me a good answer either way. That's why I went this route to see if anyone had done it or knows how.