ESX Server Hangs on Boot-up

[spi200]

Hi all

If have been happily running ESX server on three HP blades for 18 months now with little or no problems, but we had a Blade crash a week ago and still have not recovered it. We run ESX server 2.5.4 on two BL25's and one BL45 connected to a HP MSA1500 SAN. Each ESX server boots off its local storage on each blade.

Here is what has occurred.
The motherboard failed on our production blade BL45 – HP have replaced the motherboard (after a number of visits) and the Blade will atleast start and now boots.

We moved our production vm servers across to two other blades (ESX servers) and they are functioning OK. – No problems here really however backup are a little slow.

The BL45 server will not complete a ESX boot sequence – hangs on “VMware wmkernel” – the server will complete a boot sequence in Linux or Linux-up mode only.

A check of log files in the \var directory show nothing worthwhile at all that I can see. Dates in the logs preceed the crash.

At the point where the VMkernel hangs the BL45 Blade displays a led (Front panel) that it is in a degraded state – HP technician is investigating this.

We would like to get the ESX server working again (without a rebuild if possible) but are unsure whether we have still have a hardware problem or a corruption of the VMKernel following the crash. I can find very little info on the Net regarding this issue.

Here is the last few lines that are displayed on the screen.

Starting vmkstart [ OK ]
Starting network parameters [ OK ]
Bringing up interface lo: [ OK ]
Bringing up interface eth0: [ Ok ]
Starting VMware services
VMware VMnix driver [ OK ]
VMware VMkernel


No further response is received for the Console.


Not sure if we still have a hardware problem or if we have a corrupt VMKernel.

Any help would be much appreciated.

Regards

David

 

[spi200]

Hi All

We HP came in a took our Bl45 Server away, a few days later they have returned it and the server now boots and the Sever health light is green. They said they found some faulty memory.
I loaded about 8 VM's up on the server and BAM - down she went - I thought it must have been me, so I recovered it and left only a couple of VM's on the Blade (ESX Server) - No problems for a couple of days so I put two more VM's on to it again - no problems for a couple of days. Today I put two more VM's on and BAM down it went again. I think it a call back to HP and I suspect its a memory problem again.

Regards

Dave

 

[spi200]

Hi

Well we ran a util called Memtest and it showed faulty memory so HP will be back tomorrow to change it.
See how we go then.


regards

Dave

 

[spi200]

Hi all

We HP returned to site this morning and found that they had installed incorrect memory into our BL45 server. This must have been done when the server was last at their office for repair - anyway he replace two of the 16 2Gb simms and it now seems to be working fine.
HP are not what they used to be here in OZ.

Dave