Compute Module: MFS5520VIR
Boot Drive: Storage pool consisting of 2 Intel 320 Series 80GB SSDs (part # SSDSA2CW08) . Two virtual drives configured as RAID 1 arrays
OS: VMWare ESX 5.1
There are two compute modules that each have their boot drives configured on this storage pool. There are other compute modules in this chassis that use non-SSDs for their storage pools & virtual drives.
The two compute modules using the SSD virtual drives consistently have issues booting. The watchdog timer always performs a hard reset at least once on these compute modules during reboots, often two and three times. The two events that appear in the event log are:
3522 - BIOS FRB2 Watchdog timer expired - The server was hard reset
3520 - Watchdog timer expired - The server was hard reset
The compute modules that boot to drives in the non-SSD storage pools never have this issue. Has anyone else experienced this?
Did you monitor the boot process and see in which stage was the server rest? Did you assign a non-SSD VD to this compute module and see if it still has this issue?
Also please check the following BOIS settings:
Server Management => FRB-2 Enable [Enabled] It will reset the server if it doesn't complete the POST in ~6 mins.
Server Management => OS Boot Watchdog Timer [Disabled] This should be disabled unless Intel System Management Software is installed (not available on VMware ESXi)
Also, do you have AXXTM3SATA installed? See http://www.intel.com/support/motherboards/server/mfsys25/sb/CS-030937.htm Intel� Modular Server System MFSYS25 — AXXTM3SATA board and bracket assembly install guide
The drives do have the AXXTM3SATA adapters installed.
I was monitoring the boot process via remote KVM, but during the first two attempts nothing actually displayed in the KVM remote screen.
I will double check the BIOS settings you mentioned and also attempt booting to a non-SSD device.
You can also check the log on CMM and see if there're anything related to POST failure. Doesn't look related to SSD if nothing is displaying on the screen.