I let my 2.2.0-final box run other the last days with 'normal'
stress. It acts as a NFS server and NFS client as well. ypbind runs
also. It is based on a RH 5.2 distribution, running the latest knfsd,
autofs etc.
Upon start up, memory used (-/+ buffers/cache line) is less than 5Mb.
After 4 days, only me beeing logged in, with no more processes than
4 days ago, I get more than 20Mb used :
> total used free shared buffers cached
> Mem: 31016 29840 1176 4688 864 7924
> -/+ buffers/cache: 21052 9964
> Swap: 129020 3400 125620
The same box running a 2.0.36 kernel never shows more than 10Mb after weeks
of intense stress (NFS server, NFS client, HTTP, FTP, numerous compilations,
etc.) (peaks can be of course higher (~30Mb), but memory gets properly
freed).
My question may sound silly (or impertinent ;-), but is the
memory management really clean now (it was really bad in the 2.1.129,
if I remember well the swap overflow I had, and improved greatly
with the latest kernels).
Thanks for any advice,
-- Thierry Danis danis@sagem.fr- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/