HPE 9000 and HPE e3000 Servers
cancel
Showing results for 
Search instead for 
Did you mean: 

Care to offer a dignosis? L -class-power monitor card problem?

 
Ed Lawrence_1
Frequent Advisor

Care to offer a dignosis? L -class-power monitor card problem?

I have an L class server.
When I turn on the power, the attn light flashes.
No normal output appears on the attached ASCII console.

PS in GSP shows fans and power supplies as 'normal'.

VFP shows:
Run - off
Attn - flashing
Fault - off
Remote - on
Power - on
and....
"power monitor 4400"

The SL Activity log shows:
Alert level: 1
Source= 1 Processor
Source Detail= 1 processor general
Caller Activity: C = System Initialization
Caller subactivity=00

Has anyone encountered this before? Anybody care to offer a diagnosis? I'm wondering if I need either a new power monitor card or system board.
12 REPLIES 12
Sajid_1
Honored Contributor

Re: Care to offer a dignosis? L -class-power monitor card problem?

Hi,

If the it says the source is Processor, then I won't suspect a power monitor issue there. DO you have any chassis codes ? any error messages or warnings ?

I would advice to place a hardware support call and get it checked.

good luck
learn unix ..
Jeff Schussele
Honored Contributor

Re: Care to offer a dignosis? L -class-power monitor card problem?

Hi Ed,

Know this sounds goofy, but pull all power cables from the system - leave 'em off for a minute. Then replug them & try to power-up.
Don't know why, but Ls have been known to get in a weird state after power failures.

If that doesn't do it, I'd guess CPU 0 has gone bad & you can't even start the POST.

HTH,
Jeff
PERSEVERANCE -- Remember, whatever does not kill you only makes you stronger!
Robert_Jewell
Honored Contributor

Re: Care to offer a dignosis? L -class-power monitor card problem?

The LED pattern (Run - off, ATTN - flash, FLT - off) indicates that no PDC code is executing or a non-critical error has been detected. If your system reporting this at power on then I would suspect the first state.

You could try reseating the CPU's and if possible cleaning the contacts (carefully) as these connectors are very sensitive to dirt and dust. This would be especially true if the CPUs are being moved around every now and then (as in a test or training system).

Good luck!

-Bob
----------------
Was this helpful? Like this post by giving me a thumbs up below!
Bill McNAMARA_1
Honored Contributor

Re: Care to offer a dignosis? L -class-power monitor card problem?

I'd go with Jeff's suggestion..
This is a known issue.
I believe it's corrected via firmware.

Later,
Bill
It works for me (tm)
Ed Lawrence_1
Frequent Advisor

Re: Care to offer a dignosis? L -class-power monitor card problem?

update:

1) Have tried unpluging power cords for a few mins.
2) Have tried swapping the two cpus.
3) Have tried reseating cpus, power monitor card, and cpusm, and GSP/console card.
4) Have noted that new log entries have appeared in the GSP. None at all; have checked all log types.

No change in status.
Ed Lawrence_1
Frequent Advisor

Re: Care to offer a dignosis? L -class-power monitor card problem?

There was a typo in my last update. (The word "no" got left out of my last sentence)

NO NEW LOG ENTRIES have appeared at all in the GSP logs.
Thanks.
Jeff Schussele
Honored Contributor

Re: Care to offer a dignosis? L -class-power monitor card problem?

Hi Ed,

Well, I certainly think you've covered everything. I *doubt* you have 2 failed CPUs (although it's possible...) so I have to believe that your initial diag was probably correct.
Better call HP & have them come out to replace that Pwr Mon board.....hint that they should probably bring a CPU, too, just to be safe.

Good Luck,
Jeff
PERSEVERANCE -- Remember, whatever does not kill you only makes you stronger!
Patrick Wessel
Honored Contributor

Re: Care to offer a dignosis? L -class-power monitor card problem?

Hello Ed,
Did the server ever run normal or did this failure appear during the installation.
Can you give us the last logs from the GSP error log, this is usually more helpful than the activity log
What do you mean by "Have noted that (no?) new log entries have appeared in the GSP" Does the GSP log the same entries as every time or does it log nothing? If the "nothing" option is true, try to log an entry with the CE command and see if it appears in the GSP (to verify that it functional)

There is no good troubleshooting with bad data
Ed Lawrence_1
Frequent Advisor

Re: Care to offer a dignosis? L -class-power monitor card problem?

Hi everyone, here is an update with answers to Patrick's questions.

1) The server is used by our Support and Training staffs. I'm in Training. I use the system in an HP course that my company offers. According to my Support staff the system was working last week.

I took it into my classroom on Wednesday morning. It initially exhibited the symptoms I have described. I thought the problem might be IO-related or monitor-related. I changed the monitor type from HP to EM100 and the the system seemed fine. I could see output and it booted.

We used the system in class and at lunchtime shut it down. After lunch we turned it on and encountered the symptoms I have described.
This time, changing monitor type made no difference.

2) Today, I used the CE GSP command to log a 'repair' log message as Patrick suggested trying.
It was logged with a date of 1/1/1970 and a time of 00:09:47. the Alert Level was 6, the Source was 6, and the Reporting Entity was 7 (cusotmer engineer). My msg appeared on the far right.

3) I then noted how a lot of messages had dates of 1/1/1970 and times close to 00:00.

I checked the Activity logs and noted how the last log msg with a valid date was 7/31/02 at 13;16;31.
Alert level: 1
Source:1 Source detail: 1
Source ID:0
Problem detail: 0
caller activity: C
Caller subactivity: 00
reporting Entity Type: E

There were numerous logs msgs that matched this one. I found some for 7/31, 7/10. 7/5, 7/3, and previous.


3a) Here is the most recent ERROR log entry:

ERROR LOG:
1/1/1970 00:00:31
Alert level: 2
Source: 6
Source detail: 3
Source ID: 7
Problem detail:5
Caller activity: 4
Status : F
Caller subactivity: 5 = fan
Reporting entity type" 2=power monitor

BTW, the PS command shows two working power supplies and all fans present. Visual inspection indicates that all fans and power supplies seem to be working.

Support staff says it is time for a new system board. I was hoping to try everything we can before doing that, including trying to figure out if the prblem might be just with the Power Monitor card.

Thanks again.
Patrick Wessel
Honored Contributor

Re: Care to offer a dignosis? L -class-power monitor card problem?

Ed,

Sorry for asking so many question at once, but I thought it make more sense because of the time delay to ask one more question than waiting an additional day.

The GSP has no RTC. Every time you disconnect the power the date will fall back to 01.01.1970. At bootup synchronizes the diagnostics daemon the clock of the GSP. This is why you find so many strange dates in the logs.

A strange thing about your most resent error log is the Alert level. An Alert level 2 is normally not logged into the error logs, because it's not an error. A typically alert 2 is a message like "fan functioning"
This sounds like an firmware issue, that has nothing to do with your current problem, but you should keep it in mind.

I'm afraid that finding a hint what causes your system to not-cooperate is more time consuming and it's not as efficient that trouble shoot it with your local support.
But if you would like us to get a deeper look into your problem we would need the latest 'real' error logs of your system

Good luck
Patrick

There is no good troubleshooting with bad data
Ed Lawrence_1
Frequent Advisor

Re: Care to offer a dignosis? L -class-power monitor card problem?

PROBLEM SOLVED

First I want to thank everyone for your suggestions, ideas, and support.

On Friday evening a new system board was delivered, and Bob my support person, installed it. As we both feared/suspected the problem remained.

We had both heard/read that parts between N's and L's are not inter-changeable, but he took a chance and took the CPUSM from our N and put it in the L.

Yes, the problem was a bad CPUSM.

Thank you all again for your support. I learned a lot through this experience and your helpful replies.

EL


Ed Lawrence_1
Frequent Advisor

Re: Care to offer a dignosis? L -class-power monitor card problem?

Bad CPUSM.