So im trying to do a cold migration. It works fine but its slower than it should be.
Im aware that it uses the management interface and am sure that this interface is not under excessive load.
Its gigabit and only passes through a single switch, its not routed or anything daft like that.
The NIC useage graphs in vcenter look a bit bizzare (see below) I tried the same setup in the lab and couldnt reproduced the spikes.
I've tried all the obvious things such as upgrading the firmware on the switch and all hosts and checking that the CPU's dont have high load averages.
Any opinions would be much appreciated.
p.s. the hosts are all dl380 G5's with 2 x Xeons and 8GB RAM.
I assume the VMFS datastore is on local storage. What kind of disks do you use (SAS/SATA, speed)? Which RAID controller (P400/E200)? How much cache does it have and does it contain a battery (BBWC) for write-back operation? How is the RAID configured (# of disks, RAID level)?
Thanks for your reply.
This is running on iSCSI, the iSCSI box is also a 580G5 it has a P400 controller with 512 BBWC.
The VMFS store is made up of 5x 300GB 6G DP 10K SAS drives in RAID 5.
I've checked its stats and CPU/MEM are fine.
iSCSI is running over a dedicated gigabit switch the utilisation on all the links is low i've even tried shutting down all the VM's except for the one im migrating and the same problem still occured.