[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Sarge not showing all RAM

JerryN wrote:

Intriguing!  So many words in this track being said but nothing really
of any value whatever...
Very well, Jerry, I feel from your words that my answer has not been helpfull... "nothing really of any value whatever...."
So, either you're stating that my english is so bad that I can't read, understand or write a sentence that makes sense to other, or maybe you're trying to say that I'm just dumping garbage to the mailing list... Before my "clarification" I must say I felt and I feel ofended by what you've said. I have tried to be helpfull. If you believe I have not been, please post your reasons for that claim. If you do that, maybe I can improve the contents of my answers and you improve your ways to comment other peoples attempts without offending. 

Well, I'll clarify my answer then:
1) I have read a post from someone that claims that can't see the 2Gb memory it has on the system. He sees less than 1GB. The person said he tried kernel 2.4.27-1-386 and 2.6.8-1-386. 2) That person claims it has another machine in wich he can see the full 1GB kernel. 3) The person stated "Do I miss anything ? I believe that the standard kernel should work with 2GB..."
4) When I posted there had been already a reply stating that to access the full 1GB thr kernel must have the "highmem" option enabled.
I replied SHOWING MY MACHINE DETAILS (therefore providing PROOF of my claims) that a) I am running a 2.6.8-1-686-smp kernel (a standart debian kernel, not a kernel compiled by myself) . b) In my system I can see the full 1 GB memory that is installed (so, SHOWING the problem does not occur if you use the same kernel as I do) c) the kernel parameters that my kernel has enabled and that are allowing the system to see the full 1GB (so that if you decide to compile the kernel you're using presently you know what options to look for) 

THEN I SUGGESTED
a) Check the running kernel on the good machine (You're probably using a different kernel so you may try to install in the bad machine the one that is on the good machine - "uname -a" would be a way to check the kernel version -> I HAVE SHOWN WHAT MY SYSTEM REPLYES SO THAT YOU HAVE AN EXAMPLE) b) Install a 686 or a 686-smp kernel. I am not sure if all combinations of kernel-image-2.x.y[-smp] have HIGHMEM options enabled (the ones I SHOWED were enables in my kernel) BUT I BELIEVE they all must have. 


SO THIS IS NOT GOOD? NOT A GOOD REPLY? NOT INFORMATIVE? NOT HELPFULL?
Yes, I'm shouting! IF you haven't understood some of my info you could just ASK! Instead of DEGRADING my attempt to help! 

"nothing really of any value whatever...."  bah!

What did you wanted me to say?
How to install a package? How to compile a kernel? How to check your kernel version? How to list the available packages? How to check wich processor you have? How do you clean your ass?
I hope you solve YOUR problems! They're not mine! And, for your new-year, I hope you learn how to ask things, or comment on answers, so that people feel good by helping you! 

.....
Joao Clemente


On Thu, 2004-12-23 at 16:02 +0000, Joao Clemente wrote:


Björn Abt wrote:


Hello List,

I have a HP DL360 G4 with 2GB ECC RAM and a P4-Xeon with 3 GHz. I installed Debian Sarge netinst on this machine and after examining the system i found out that top only shows 906MB RAM, even /proc/meminfo shows only 906MB. I tried this with both the 2.4 (kernel-image-2.4.27-1-386) and 2.6 (kernel-image-2.6.8-1-386) Kernels. After that I ran memtestx86 to see if I had malicious Memory, but it showed the full 2GB and it ran a whole day without showing any errors.

Do I miss anything ? I believe that the standard kernel should work with 2GB...

Another machine with 1GB Non-ECC-RAM shows the full 1GB in /proc/meminfo


jpcl@linux:~$ uname -a
Linux linux 2.6.8-1-686-smp #1 SMP Thu Nov 25 04:55:00 UTC 2004 i686 GNU/Linux 
jpcl@linux:~$ free
             total       used       free     shared    buffers     cached
Mem:       1035796     863212     172584          0      83236     677888
-/+ buffers/cache:     102088     933708
Swap:       979832          0     979832
As you can see, I have access to the full 1GB (well, not exactly 1024MB but I supose the remaing is eated up by the kernel) 

My kernel has
# CONFIG_NOHIGHMEM is not set
CONFIG_HIGHMEM4G=y
# CONFIG_HIGHMEM64G is not set
CONFIG_HIGHMEM=y

and it IS a stock debian kernel....
I would suggest you to check wich kernl is running in that 1GB non-ECC-RAM machine... I (am not sure but) believe the 686 or 686-smp kernels have high-mem enabled. 

Good luck
Joao Clemente
Reply to: