After upgrade to SAN IQ 9.5, write peformance issue

[oikjn]

it is...  it has to be that way...  if you are synchronous the write doesn't "count" until it exists at both sites.  That definitely adds to latency since the write has to go to both sites and be confirmed at both sites, but that is the definition of synchronous... otherwise it would be asynchronous.

[Amana Living]

Hi, is this still an issue? I was thinking of upgrading our SAN's from 8.5. Should I just go to 9.0 then?

[Besty_Boy]

Hi, I have just upgraded our multi site san from from v9 to 9.5 patch set 3 without an issue. We had to reduce our volumes to under 2TB as we were having performance issues while the nodes were resynchronising before hand. Once completed everything went very smoothly. I have had no reported performance issues.

[Couchman]

Yes we still have a problem, a case is still open which is now with the HP VMware team as the storage team think everything is OK with the SAN. Are you running asynchronous multi-site configuration Besty_Boy?

[Besty_Boy]

We are running a 6 node hyper-v synchronous multi-site SAN. This is split over 3 site with the quorum on the 3rd site. The distance between storage nodes is about 2km.

[David_Tocker]

Yes but what is the latency between the sites? it should be bugger-all for good performance.
2km tells us nothing. How are they linked? Bandwidth?
I only run a single site cluster, but as far as i can tell 9.5 has not affected the performance (SCOM logging confirms this)