On Sun, Oct 04, 1998 at 02:05:53AM -0700, Toomas Losin wrote: > On production boxes with 128 megs I usually allocate 256 megs swap so > that if a bizarre situation arises that chews RAM the box can survive > for a while while I figure out what caused the problem. Who else has > watched a Linux box recover from a load average of 600? :-)
One of ours made it to 280 or so a few weeks ago, and came back. (some
boxes should just NEVER be allowed to start swapping ;).
Prior to that I don't think I'd ever broken 100 on a system that
was actually going to recover from the experience.
VanLUG Email Archive