On bootup, the computer (bios) says that it has 64MB of ram, and 1 64MB DIMM chip is installed. When the 6.0 or 6.1 (original 6.1, not the RHEA boot images) bootable install CD's start up, they only find 16MB. After the system is installed and booting from the hard drive /proc/meminfo still shows only 16 MB. If I try booting with "linux mem=64m" the systems crashes before as the first filesystem is loaded. I can make it boot with "linux mem=62m" but why isn't it finding all of the installed memory in the first place? Since these machines are going to be servers, I really don't want the systems to crash because of memory (available/usable vs specified) issues. Wasn't the 6.1 release supposed to detect memory past (and one would believe up to) the 64MB limit? I have tested this on two identical boxes, so I don't belive that it is a hardware problem (unless the hardware problem is identical as well 8-)). Then again, it the bios (rev. A01) in these boxes could be really screwy. -- rouilj
Assigned to dledford
I'm having the same problem with an assembled machine. It has 128 mb but in red hat I only could see 64mb. The machine uses the AMD Athlon CPU. Booting the computer with the trick rouilj said, I can see up to 120 mb. The other 8 mb are used by the integrated video card. I added the option to the lilo.conf file in order to avoid typing it every time the machine boots. Isaac
This is vendor BIOS problems not kernel. Linux 2.2 supports E801 large memory queries. 2.4 supports the E820 stuff which helps for some of these vendor problems
Commits pushed to master at https://github.com/openshift/openshift-docs https://github.com/openshift/openshift-docs/commit/973c71946bfbe263bdf0bda52b7054749f7c7fa4 Issue 7105, removed from cluster resource override example https://github.com/openshift/openshift-docs/commit/c695e5b91b9d739b17ae6c387b1b341459ec0d1b Merge pull request #7106 from ahardin-rh/overcommit-fix Issue 7105, removed kubernetesMasterConfig from cluster resource override example