SRM meesage

[Roro_2]

Hello,

Our customer has an ES40 Alpha server Model 2 running Tru64 Unix 5.1 with one CPU 667 MhZ.
We noticed that after powering up the server and before getting SRM prompt P00>> , the following message is dispalyed on console :
unexpexted system evrnt through vector 680 on CPU 0
os_flags 00000000000
cchip_dirx 0004000000000
tig_smir 00000000000008
tig_cpuir 00000000000001
|
|
reserved_1 0000000000000

NB : I changed the CPU and the problem still not fixed.

[kobylka]

Hello Roro!


You are having an environmental error, which causes a "logout frame" (those numbers you posted) to be printed on SRM. Each line corresponds to part (offset) within this frame but there seem to be some fields missing. Could you please post the fields (lines) between tig_cpuir and reserved_1 too?


Kind regards,

Kobylka

[Roro_2]

Hi Kobylka,

Thank you for your concern.

please find below the missing fields:

tig_cpuir 0000000000001
tig_psir 0000000000003
ln78_isr 0000000000000
door_open 0000000000000
temp_warning 0000000000000
fan_ctrl_fault 0004000000000
power_down_code 0000000000000
reserved_1 0000000000000

Regards

Roger

[kobylka · Accepted solution]

Hello Roro!


I've been checking your values and am a bit confused by the value in

fan_ctrl_fault 0004000000000


For sure, the problem is related to either:

- System fan (any of fans 1 - 6 not responding to RMC)
- CPU fan (fans 5 - 6 , speed issues)
- PCI fan (fans 1 - 4, speed issues)

Maybe some fan is not working correctly or your ES40 is simply overheating. For a description of the fans in the system refer to "ES40 Service Guide - 1.12 Fans"

Confusion arises when trying to decode the bitmask because valid bits for field "fan_ctrl_fault" range between 0 - 11 (12 bits in total, counting from right to left) and your fan_ctrl_fault bitmask has bit 38 set (!!!).

I would strongly suggest to use Compaq Analyzer to interpret this error fault.
Boot into the OS (Tru64 or OVMS) and start CA. It will display error events for localhost by default and give you way more information on the actual failure.


Kind regards,

Kobylka