Quite a few of us out here are using the EL releases. Right now, I'm
typing this using K12LTSP 4.2EL, and just last week, I did my yum
update, which pulled down all the new CentOS 4.6 updates. If this is a
new deployment, you would do very well to go with the 5.0EL release.|
Michael Blinn wrote:
RAM issue is a limitation of physical address space in 32-bit systems.
You have two options to get around it: Run 64-bit kernel (if processor
supports it) or run the bigmem kernel, which implements PAE (Physical
address extensions). There is a performance hit associated with using
PAE, however I'm not sure of the specifics. I shouldn't expect it to be
really noticeable unless you're on a loaded server. You may want to
pass a mem=4GB kernel boot options and see if that makes a difference.
- Perhaps PAE (bigmem) is slowing you down a lot?
And yes, I would suggest using CentOS as FC6 is already EOL.
Shawn Powers wrote:
I remember having to use a bigmem kernel a
while back to get support for my 6GB of RAM in my 32 bit xeon server,
but for some reason I though K12LTSP 6.0 had support for it with the
SMP kernel. Maybe I'm doing it wrong...
That said, I installed K12LTSP v6 over Christmas break (updating from
my FC1 version, whatever it was), and today has been a nightmare!
Everything is so darn slow... I'm using xfce, without the session
manager (it caused crashes), but everything about the systems are
slow. (I installed 2 servers, identical hardware, and firefox is being
I'm considering installing 5.0.0EL, especially since FC6 is about to
die... or maybe it's already EOL, I'm not sure.
Pay no attention to the man lamenting in the latter half of that email,
but I'm curious about the bigmem issue -- anyone have any ideas?
K12OSN mailing list
K12OSN redhat com
For more info see <http://www.k12os.org>