On Fri, 2006-07-21 at 12:36 -0500, Larry Vaden wrote: > On 7/21/06, William L. Maltby <BillsCentOS at triad.rr.com> wrote: > > On Fri, 2006-07-21 at 12:10 -0500, Larry Vaden wrote: > > > On 7/21/06, Scott Silva <ssilva at sgvwater.com> wrote: > > > > Larry Vaden spake the following on 7/21/2006 8:05 AM: > > > > ><snip><snip> > Here's what got my curiosity up (and wondering why the swap'd item(s) > didn't eventually expire); the first machine is running sendmail, the > others are running postfix. > > Mem: 1025076k total, 602508k used, 422568k free, 42204k buffers > Swap: 2031608k total, 600k used, 2031008k free, 506496k cached > > Mem: 1295916k total, 719596k used, 576320k free, 70192k buffers > Swap: 2031608k total, 144k used, 2031464k free, 441984k cached > > <snip similar> In that thread was mentioned a "feature" that came to be considered a bug. That's why U4 is supposed to remove it. That *maight* affect what we are seeing. You can reduce even that by using the "swappiness" thing and considering the other actions mentioned in that thread. What you show is *not* a problem, IMO, but an irritation. I believe the memory manager just dumps some *really* inactive pages out so that buffers and cache (which should be *very* active areas) can get more memory. <straight face> Makes since: why maintain an error trapping routine in memory at the expense of buffers and/or cache? We *know* there are no bugs that need trapping, right? </straight face> > rgds/ldv > _______________________________________________ > CentOS mailing list > CentOS at centos.org > http://lists.centos.org/mailman/listinfo/centos -- Bill -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 189 bytes Desc: This is a digitally signed message part URL: <http://lists.centos.org/pipermail/centos/attachments/20060721/8c8694c1/attachment-0005.sig>