Any feedback would be appreciated. Thank you in advance.
We have Exchange Server 2013 and run two MBX servers in a DAG, with 2 databases. Each database has about 200 mailboxes
Our databases and logs are split into separate volumes with performance policies to match.
For the last 90 days our average latencies are (Read/Write):
DB1-Data: 0.79ms / 0.29ms
DB1-Logs: 0.92ms / 0.19ms
DB2-Data: 0.82ms / 0.31ms
DB2-Logs: 0.99ms / 0.19ms
We're achieving and average of 97% cache hit
Maybe a little more information from you would help get some answers. Which Exchange Server version? What roles? What type of volume (database, logs)? How many users?
Thank you for replying. We have Exchange 2013 using client access and DAG. We have about 300 users and our logs don't get too big so we placed the whole server in one volume. I was just curious what others on average were getting under the performance volume read/write latency.
Typically JetStress (the Exchange IO test tool) will want to see reads to the data volume complete under 20ms and logs volumes to complete under 10ms in order to pass
There is a good guide here
Thank you for replying and for the guide. As I replied to Nelson, I meant to ask what other Nimble customers were getting on their Exchange Volume DB/Log read/write average latency. I appreciate the guide though which I will take a look at.
Over 30 days:
As reported by the array:
Data: 1.58ms / 0.18ms
Logs: 1.07ms / 0.16ms
As reported by VMare (so includes OS drives too etc):
DB Server: 2.7ms / 0.14ms
Cashub: 1.5ms / 0.33ms
For a comparison, our previous storage (10k SAS enterprise array):
DB Server: 14ms / 2.6ms
Cashub: 18ms / 3.5ms
MDB1 3.53 ms Log1 under 1ms
All the others DB log pairs are broadly similar
Retrieving data ...