On Sat, 2007-08-11 at 20:46 +0200, wolf2k5 wrote: > On 8/10/07, Eric Sisler <[EMAIL PROTECTED]> wrote: > > In my case I was upgrading various VMware servers from RHEL3 / VMware > > GSX to RHEL4 / VMware Server. One of the virtual machines on a server > > with 16Gb of RAM kept getting whacked by the oom-killer. Needless to > > say, this was quite frustrating. > > Thanks for the detailed message, I'm going to try the suggested > changes on an host with RHEL4 and VMware Server that I had some VM > killing issues with. > > BTW, any ideas why you didn't have this problem with RHEL3 and VMware GSX?
I hate to answer for him but basically he was comparing a RHEL3 system with 6GB of RAM with a RHEL4 system with 16GB of RAM. That's not a fair comparison because the extra memory will cause more lowmem pressure even on RHEL3. > Also, does the same problem affect RHEL5 (I don't have an host handy > to test with)? Because of the way Redhat has reworked the kernels included in RHEL5 it shouldn't really have this issue. For the 32-bit version of RHEL5 there are really only two kernel (not counting the xen kernel). You can have either the standard kernel or the PAE kernel, but according to the release notes the standard kernel is limited to recognizing 4GB (I have not verified this) so if you have more memory than that you have to use the PAE kernel. The PAE kernel is basically the RHEL5 equivalent to the hugemem kernel in previous releases so effectively Redhat no longer gives you the option to run a non-hugemem kernel on systems with more than 4GB of RAM. That should effectively be the same as RHEL4 running the hugemem kernel which usually effectively prevents this issue. For a more accurate and detailed description of the problems involved some people might want to read the following article: http://kerneltrap.org/node/2450 Later, Tom _______________________________________________ rhelv5-list mailing list [email protected] https://www.redhat.com/mailman/listinfo/rhelv5-list
