Proliant 6500: Boot halted on Error 1610

[elverk1]

This unit has been in storage for 18 months or so. We pulled it out to run as a test server for a project. Boot proceeds normally until after processor initialisation. At this point the boot process stops and outputs to screen "1610 - temperature violation detected, waiting 5 minutes to cool"
There is no physical sign of anything running hot. All fans are operating normally.
The boot does not resume after 5 minutes, but is halted indefinately (well, at least 25hours anyway) The system continues to operate, however, as it paints a dot on the screen once each minute whilst waiting to cool. As the boot has halted before completion, there is no access to keyboard, cd or floppy drives, so no chance to load any diagnostics.
Does anybody know where the 1610 error originates? where is the temperature sensor(s)? Is there any way of forcing the boot process to continue?
Your help will be appreciated.
Peter

 

[TheLad]

I would suggest reseating the processors and checking that the heatsink is seated correctly. If the heatsink has a fan on it, make sure the fan is running (I know you said that all fans are working, this one is worth an extra check.

It sounds like it is overheating the processors. Bear in mind that processors do run very hot during normal operating which is why the heatsink is important.

--------------------------------------
"Insert funny comment in here!"
--------------------------------------

 

[burtsbees]

I concur---however, I would also put new thermal compound on the procs, making sure to clean the old crap off with alcohol. Is it yellow, and hardened?. Can't tell you how many times I have seen that.

Burt

 

[elverk1]

Thanks for the tips - concensus seems to be that the problem is with processor overheating.
The unit has 4 processors. I have only 2 processor terminator modules so can only swap out procs two at a time. I have done this and the error perpetuates - which suggests to me that if procs are the problem then all 4 have the fault. I will sit down somewhere quiet and dis-assemble the procs and re-grease, as suggested.
Is it possible that this issue arises because of corruption in the config data held in NVRAM? I have replaced the memory backup battery, but was not sure that I had purged any data (corrupt or not) out of memory. How is this done?
cheers
Peter

 

[burtsbees]

Nothing to do with NVRAM. Regrease the procs. Like I say, I have seen it a million times...I have had my techs go out to sites..."I swapped the processors and it still happens..." "Did you clean the old crap off and put new Arctic Silver on?"
"Uh, no---that wouldn't make sense...why would both procs do that then? "

"JUST DO IT!!!"

So, that has fixed it, about one million times.

Burt

 

[elverk1]

All 4 procs dis-assembled, old gunk cleaned off and new heatsink compound applied. The old stuff was very dry and yellowed, as you said.
Full of confidence, replaced procs in chassis and applied power. Booted into post and halted on error 1610, just as before.
I'm pretty confident I did a good job of rebuilding the procs - so any other suggestions for tracing the source of this error?
cheers
Peter

 

[burtsbees]

If you ensure that all proc heatsinks are properly clipped to the procs, and all fans are working, then I would look into replacing the motherboard, as the thermal sensor may be faulty. Beofre you do that, just try one proc (in socket one, of course) by itself. Also, try and update the BIOS with the firmware update cd.

Burt

 

[elverk1]

I tried running on two procs (I only have two terminator modules) in various sockets - all without any change. Is there a way of doing the BIOS update when the system halts before completing POST?
cheers
Peter

 

[burtsbees]

Oh yeah...there is that...crap.
I'm not sure off top---maybe a boot floppy that does it...look on HP's web site, I suppose.

Burt

 

[elverk1]

I think I'm at the end of the road on this one. This server was retired a couple of years ago so for sure we are not going to put a new mobo into it. Shame - always useful to have a big old boy like this available for trials etc.
Thanks for your help guys.
Peter

 

[TheLad]

Before you bin it, try running a single processor only and maybe swap the components until you can find one that works (if indeed it will work)?

--------------------------------------
"Insert funny comment in here!"
--------------------------------------

 

[burtsbees]

Lad---he said that he only has 2 terminators...

Burt

 

[elverk1]

Having realised that the 6400R terminator modules are the same as those used in the 6500 (same mobo) I tried booting with only one proc. No joy I am afraid. Tried all 4 proc's individually in all 4 sockets, but error 1610 every time. I think you are right Burt, the problem is on the mobo. Seems odd that an error like this, which is a complete showstopper, should be so poorly documented. I guess this means it is rarely encountered.
Thanks for the help.
Peter