I was hoping someone might be able to assist with some major performance issues we are having in our production environment. (NOTE: I inherited this setup, sans some recent changes to the swtich cabling, when I started with the company)
Setup:
1 EQL PS6100e iSCSI SAN (24x1TB NL-SAS 7200rpm drives) with one monolithic LUN
~50 VMs running off of the SAN with at lest 1 volume per VM, some with 2, spread across 10 VM hosts
iSCSI traffic is running through a PC7048P switch, on its own VLAN (FW 5.1.1.7)
VM traffic is running through a second PC7048, on 2 VLANs (FW 5.1.8.2)
VM hosts are running Free ESXi 5.5 update 2. Most have 2 NICs for VM traffic and 2 NICS/kernel ports for iSCSI with multipathing
The main issue we are seeing is that our hosted clients are experiencing extreme lag when running their primary application (2 applications, one runs against an Advantage database server that also houses the roaming profiles data directory, and the other runs against a separate POSTgreSQL database server). This has essentially rendered these applications nigh unusable.
Is this something that could be configuration based (i.e. not using a 1:1 LUN to Volume for the database servers) or some other inherent issue I could be looking for? At this point, we are probably going to migrate the database VMs to a local storage VM host setup (5 SATA 3Gb/s in RAID 5) to try and eliminate some of the severe latency issues while we look at and probably purchase new servers with sufficient 10K or 15K SAS drives to support our needs.