N4000 Boot-time Messages after Mem upgrade

[Sean McConkey_2]

Hi

Recently added another 4Gb to an N-Class, using non-HP but certified memory.

I get the following on boot and am sure there is a way to disable it. IT does time out but takes sometime so would prefer to either disable or set timeout to about 2 seconds.

Any help gladly appreciated.

Cheers

Sean

***** EARLY BOOT VFP : SYSTEM ALERT *****
SYSTEM NAME: uninitialized
DATE: 02/20/2002 TIME: 13:58:44
ALERT LEVEL: 6 = Boot possible, pending failure - action required

REASON FOR ALERT
SOURCE: 7 = memory
SOURCE DETAIL: 4 = SIMM or DIMM SOURCE ID: FF
PROBLEM DETAIL: 7 = added to configuration

LEDs: RUN ATTENTION FAULT REMOTE POWER
FLASH FLASH OFF OFF ON

0x3000186774FF226F 00000000 000004DB - type 6 = Tag
0x5800186774FF226F 00006601 140D3A2C - type 11 = Timestamp 02/20/2002 13:58:44
A/a: ack read of this entry - Q/q: quit Virtual Front Panel Display
Anything else redisplay the log entry
->Choice:

[Paula J Frazer-Campbell]

Sean

Are you are saying that you intend to ignore the server warning you about a memory problems?

Even if the memory is not HP then the server is checking it on boot and saying that it has problems.

Is this the correct memory? running at the same speed as other memory?

Does the server boot OK without this new memory?

Is it filled correctly ballanced accross the slots?


Paula

[Paula J Frazer-Campbell]

Sean

Are you running from a UPS?

LEDs: RUN ATTENTION FAULT REMOTE POWER
FLASH FLASH OFF OFF ON

0x3000186774FF226F 00000000 000004DB - type 6 = Tag
0x5800186774FF226F 00006601 140D3A2C - type 11 = Timestamp 02/20/2002 13:58:44

The REMOTE POWER bit needs investigating.


Paula

[Stefan Farrelly]

The proper way to test if this new memory is ok is to boot the server (if it will boot - if it wont boot obviously the new memory is no good), and run XSTM (online diagd). Select the memory and run the exerciser on it (this will take a while and put a heavy load on your server). If there are problems you will now get entries in syslog looking like this;

Feb 20 14:15:53 hal EMS [2778]: ------ EMS Event Notification ------ Value: "MAJORWARNING (3)" for Resource: "/system/events/memory/49" (Threshold: >= " 3") Execute the following command to obtain event details: /opt/resmon/bin/resdata -R 182059010 -r /system/events/memory/49 -n 182059009 -a
Feb 20 15:16:15 hal EMS [2778]: ------ EMS Event Notification ------ Value: "SERIOUS (4)" for Resource: "/system/events/memory/49" (Threshold: >= " 3") Execute the following command to obtain event details: /opt/resmon/bin/resdata -R 182059010 -r /system/events/memory/49 -n 182059010 -a

This shows you definitely have memory problems. Things to try are reseating the memory SIMS or swapping them around. We had to do this twice on an N class we added memory to recently before we stopped the memory errors, but they did stop, and its been fine for ages now.

[Ian Dennison_1]

To me, this message says "Found new memory - updating configuration". Does it happen on the second boot?

Share and Enjoy! Ian

[Paula J Frazer-Campbell]

Hi

Memory info:-

Memory Configuration Rules

N-Class systems support up to 4 memory carriers (numbered 0, 1, 2, and 3), each with slots for 4 memory DIMM pairs (numbered 0a/b, 1a/b, 2a/b, and 3a/b). For the DIMMs to work, both DIMMs in a slot pair must be the same type (i.e., the same part number).
?? Performance considerations. For best performance, install DIMMs evenly among the carriers, loading carriers 0, 1, 2, and 3 in sequence. For example, four 512MB memory DIMM pairs should be loaded into slot 0a/b on carrier 0, then slot 0a/b on carrier 1, and so on. Although not a requirement for performance, the largest memory sizes should be installed first, making it more convenient if the memory DIMMs are re-sequenced at a later time.

?? Cooling considerations. The following will help optimize air flow cooling within the system.

Always install memory carriers in the sequence 0, 1, 2, than 3. For example, if you have only two memory carriers, make sure they are installed as carrier 0 and carrier 1.

Install DIMMs into the carriers in the following slot order: 0a/b, 1a/b, 2a/b, 3a/b.

The most important rule is that DIMMs of the same type are distributed evenly amongst the carriers. If necessary, remove and re-sequence existing memory modules to balance the distribution; otherwise, a PDC error may occur.


HTH

Paula

[Sean McConkey_2]

Paula,

I took this message to read that it was reporting that new memory had been installed, and not an actual problem with the memory. Ian in answer to ? yes it happens on every boot.

i.e. PROBLEM DETAIL: 7 = added to configuration

I could be wrong though.


Stefan, I'll give online diags a whiz, no problem about performance issues, not a live system.

Sean

FYI

The memory is installed as below, and was done by a certified engineer.

MEMORY INFORMATION

WARNING: The non-destructive test bit was set, so memory was not tested
destructively. Information only, no action required.

WARNING: One or more DIMMs in the system were not made by HP.


MEMORY STATUS TABLE (MB) (Current Boot Status)

Carrier->0 Installed 1 Installed 2 Not installed 3 Not installed

Slot 0a 512M Active 512M Active-MBE - -
Slot 0b 512M Active 512M Active-MBE - -
Slot 1a 512M Active 512M Active - -
Slot 1b 512M Active 512M Active - -

Slot 2a 1024M Act-Non HP 1024M Act-Non HP - -
Slot 2b 1024M Act-Non HP 1024M Act-Non HP - -
Slot 3a - - - -
Slot 3b - - - -

Subtotal 4096M 4096M 0 0

TOTAL = 8192 MB
---------


MEMORY FAULT TABLE

Carrier->0 Installed 1 Installed 2 Not installed 3 Not installed

Slot 0a 512M Active-MBE - -
Slot 0b 512M Active-MBE - -
Slot 1a - -
Slot 1b - -

Slot 2a - -
Slot 2b - -
Slot 3a - - - -
Slot 3b - - - -


Memory Installation Guidelines
------------------------------

- For DIMMs to work, both DIMMs in a slot pair (a/b) must be the same type.
(Same part number = same type)

- For best performance, install DIMMs evenly by type among the carriers,
loading them onto carriers in the following order 0 1 2 3.

- For proper cooling, install DIMMs onto carriers in the following order:
0a/b 1a/b 2a/b 3a/b.

- For proper cooling, install carriers in the following order: 0 1 2 3.


Active, Installed Memory : 8192 MB of SDRAM
Deallocated Pages : 0 Pages
-----------
Available Memory : 8192 MB

Good Memory Required by OS : 14 MB

Memory
HVERSION SVERSION
-------- ----------
0x0900 0x00000900

[Paula J Frazer-Campbell]

Sean

The system is not happy with this memory:-

ALERT LEVEL: 6 = Boot possible, pending failure - action required

As previously stated fire up stm and exercise the memory.

It may be just that the system is seeing non HP memory!

Your log is from GSP - what is appearing on the console during boot?

Also check out this link - same problem.

http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0x218fd219fb60d5118ff00090279cd0f9,00.html


HTH

Paula

[Sandip Ghosh]

Hi,

First thing you should put the high density memory on the first available slot that means 0a and 0b. Put 2 GB (1024MB SIMM) module on the slot 0a and 0b of both the memory carrier. Then you put the 1 GB Module ( 512MB SIMM ).

And regarding the problem what you getting, because it has not seated properly. The error message will go off when you seat the memory properly.

Sandip

[Tom Geudens · Accepted solution]

Hi,
This might be NORMAL !
I got the following mail from DataRam.


As discussed, I am sending a brief description of what customers can expect when they install Non-HP memory into their HP L Class Servers.

1. Systems with Non-HP memory will halt the boot process for several seconds, before continuing.

2. The delay will last for several seconds with a display alerting the user of the NON_HP memory installed.

3. During this wait, the system will give the user the option to (A)cknowledge the message.

4. If the customer presses the letter 'A' to acknowledge the message, the ALERT LAMP will turn off and the system will complete the boot process.

5. If the customer does not press the letter A, the system will continue to boot after the waiting period, but with the ALERT LAMP still activated.

6. The Operating System will load and run properly even thought the ALERT LAMP remains activated. This will have no effect on normal operation.



The messages you have are EXACTLY those on my system. The memory I installed works fine !

So ...
Hope this helps,
Tom Geudens

[Tom Geudens]

Hi,
extra "add-on"
I had this checked, the behavoir on N-Servers is the same.

Regards,
Tom Geudens