We’re testing our first soon-to-be production servers with vSphere, and running into some performance issues.
First off, the setup:
- Two-tier Sage CRM system, with one dedicated web server (IIS) and one dedicated DB server (SQL).
- Both servers are running Windows server 2008 R2 64-bit, with their own VMFS volume on the SAN. Each VM has two virtual disks, consisting of a thin 30GB OS partition, and a larger (50 and 250GB)thin data partition. Both partitions are ‘VM Aligned’ since Server 2008 starts at 1024K, and the VMFS volumes were created in vSphere.
- No snapshots currently on volumes.
- The DB is SQL Server 2008 R2 64-bit, with 2 vCPUs, 10GB RAM, pvscsi/vmxnet3. The web server is set up with 1 vCPU, 8GB RAM, and pvscsi/vmxnet3.
- Both are running on the Dell R610 servers, Intel Xeon E5540, dual quad-core 2.53, 32GB RAM.
- iSCSI SAN: Dell EqualLogic PS4000E SAN with PowerConnect 6224 switches, 4x Broadcom 5709 nics running as HBAs with 1500 MTU (to utilize the Dell EL MEM).
Performance of the SQL server is less than ideal. Executing large queries takes longer on this setup than on our old server (IIS and SQL on the same box with local RAID5 array, dual core Xeon 5130 2.0GHx with 4GB ram).
I would think that it should be blazing fast. Pretty much a canned install of SQL, but we have the SQL priority boost turned on currently to see if that helps.
What things should I be looking at in terms of finding bottlenecks? The SAN is fast, the procs are fast, and it’s got plenty of memory, so something seems off.
Thanks a lot for any suggestions!
Aaron