This post is more than 5 years old
2 Posts
0
2056
July 28th, 2011 08:00
Recoverpoint - CRR Async RPO
I've got a solution that I'm trying to find out the lowest RPO I could possibly achieve.
Point to Point solution VNX to VNX with Recoverpoint/SE (3.4) (2 at each site)
IP link between the sites, 1Gb and typical latency of 7ms. The distance/latency was too much for Sync and I didn't have a FC link. I'll need to set this up for Async.
I've got plenty of bandwidht and I'm more interested in low RPO. I'd turn off compression and dedup, just to make sure they don't slow things down. I'd also setup optimise for Lag and set this with
Any ideas or experience of what's the best RPO I could achieve if I setup the CG to use fixed per Write. In the Recoverpoint Overview documentation it talks about the RPA deciding when to send the writes from the local to the remote RPA based on lag policy, load and resources.
I'm only intending to have 100G of data replicated using this low RPO CG. This has a low rate of change but high importance and this is why I'm looking for the low RPO.
There's only another 2TB of data to be replicated using CRR. This can have a higher RPO and I could set any other parameters to limit the impact of this CG replication of the priority CG needing low RPO.
Any help or feedback is appreciated.
RickBrunner1
117 Posts
1
July 28th, 2011 12:00
There are a number of factors that can impact the achievable RPO, latency and available bandwidth are a couple, but you also must consider the change rate (writes that need to be replicated). As you seem to be keenly aware that there is a tradeoff between RPO and bandwidth utilization and the options in RecoverPoint to allow you to set a preference.
Assuming you have the available bandwidth and the change rate you will be replicating is lower than the available bandwidth an RPO of seconds is not unrealistic.
The best and most accurate way to determine the achievable RPO would be to work with an EMC RecoverPoint SME to gather the application IO statistics and import these into BSCD which will generate a report showing the overall bandwidth requirements, # RPA’s and Journal throughput required as well as the RPO achievable . They can also assist with advising on the CG policy settings if needed.