Hi There,
We have two servers with ESXi 5.1 installed. Each server has 2 NICs.
Two iSCSI LUNs are exported from Solaris OS to two ESXi servers. The Solaris system has 4 NIC.
NICs are connected as follows.
----------------------------------------------------------
Solaris ---- ESXi 5.1
------------------------------------------------------------
10.7.0.11/16 ------ Node1-10.7.0.12/16
10.8.0.11/16 ------ Node1- 10.8.0.12/16
10.9.0.11/16 ------ Node2- 10.9.0.12/16
10.10.0.11/16 ------Node2- 10.10.0.12/16
Each ESXi server sees two paths (one is active and another is standby). MTU is set to 9000 on all the NICs.
iSCSI port binding is enabled on each esxi servers.
What we observed is that, sometimes the rescan takes longer time and i/o latency of the LUN are very high. Could you please help to solve this issue?
====
vmkernel.log shows as below
cpu24:8216)WARNING: ScsiDeviceIO: 1224: Device naa.600144f06ce14300000052dfb77a0002 performance has deteriorated. I/O latency increased from average value of 144809 microseconds to 2940124 microseconds
===