A different kernel, because of more memory?

[Hans]

After i had added some extra memory in my Linux server with Debian Etch, the following error occurred during starting up the system:

"no such device open sock (?)" or something like that.

After that message, the system hangs.
Within the BIOS setup screen i could see that the system recognizes all the memory banks within the server, so that's ok. I am sure that the memory chipsets are ok for the server, because Dell advised me to use the ones i've bought.

So it seems that the current Linux kernel (AMD64) does not recognise the extra memory.
Now i want to know, how to solve this problem.

I am thinking about installing a different Linux kernel, but on the moment i don't know which one.
(The total memory of the server will be 8 GB)

Can anyone give me advise what to do in this situation?

Thanks in advance!

[falko]

What do

free

and

cat /proc/meminfo

display?

[Hans]

Thanks for your reply Falko,
As you understand, within the atachment is the output without the additional memory installed.
(Because the server does not boot with the new memory added, or with only the new memory installed)...

[falko]

Are those additional modules from the same manufacturer, and if so, are they the same model? Using RAM modules from different manufacturers can make a system unstable.

[Hans]

Hi Falko,
Thank you for your reply. I know that the same memory modules are important.
And yes they are all from the same manufacturer and model.

My problem occurred because the rather old Linux 2.6.x-EM64T-P4 SMP kernel did not recognise more then 2 GB. Last night, i installed the 2.6.26-1-AMD64 kernel from Debian and now all memory (8GB) can be used.

The reason i still had an EM64T kernel was from the period that Debian AMD64 was unofficial. I was still using it, because it is very stable and i had never problems with it. From now on i will use AMD64 which comes with Lenny.