HPE EVA Storage

EVA 8100 - Set cache policy to write-through for replication destinations?

 
Jeremy C
Regular Advisor

EVA 8100 - Set cache policy to write-through for replication destinations?

I'd like to bounce this idea around.  Perhaps someone can tell me if there is a major pitfall that I'm unaware of.

 

Problem:  One of our EVA 8100's is suffering from occasional periods of high latency (in the hundreds of ms) several times a day.

 

Configuration:  Two EVA 8100's with async CA.   The EVA "B" with latency issues hosts both Dev/QAQC stuff as well as replicated vdisks from the production site.

 

My solution:  Change the write cache policy on the destination vdisks to "write-through".   Periods of high latency often coincide with high IO on the vdisks being replicated to EVA "B".  If I change those destination vdisks to write-through cache then it will slow down the replication when EVA "B" can't keep up and, if necessary, start the DR groups logging rather than cause EVA "B" to suffer from high latency.  Remember, it's async CA.

 

Does this sound reasonable?   I realize it's a compromise.

 

Thanks!

 

1 REPLY 1
Sheldon Smith
HPE Pro

Re: EVA 8100 - Set cache policy to write-through for replication destinations?

Been a Very Long Time since I saw an 8100, however:

 

Are the hosts on EVA "B" sharing the FPs with the CA replication traffic?
If so, consider updating the switch zoning so the CA replication is exclusive on a couple of the FPs?


Note: While I am an HPE Employee, all of my comments (whether noted or not), are my own and are not any official representation of the company

Accept or Kudo