A good lesson learnt on my 3PAR

(Posted LinkedIn July 2013)

We have a couple of 3PAR P10400 SANs with Synchronous Replication (Remote Copy) of volumes between our DCs. All of our Remote Copy volumes sit on the same disk tier on both sides, with AO moving hot spots up and down as required.
I tuned some volumes a few weeks back on the destination 3PAR from FC to NL disk, but had an oversight that a couple of the volumes I tuned were the RC replicas. The net result once the tuning completed was oversubscription of IOPs on the NL disks, which in turn created high disk latency… and the knock on effect of this is that it affected the production volumes on the source 3PAR due to the write latency, or the delay of acknowledgements.
I am sure this configuration would work in some situations, however in our case the source volumes in question backing for VM datastores and have high IO requirements.
The moral of the story is, if you are using RC copy, make sure your volumes sit on same tier of disk on both sides, unless your IO requirement determines otherwise.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: