Re: core-dumping over NFS

From: Claus Guttesen <cguttesen_at_yahoo.dk>
Date: Mon, 12 Jan 2004 23:58:42 +0100 (CET)
> I've observed the following bad behaviour of
> -current mostly related
> to dumping core of a buggy program over the NFS.
> 
> The mounts are regular and default (v3?), except for
> the ``intr'' flag.
> No rpc.lock or anything...
> 
> Any comments?

Setting up an 5.2 (yehaa) nfs-server. Did some
network-testing witohout doing any changes to nfs on
client and server, just a plain mount.

Throughput was mere 16 Mbit pr. sec. on a 100 Mbit
network. One thread suggested the usage of TCP and
some changes to the rd-rw-size on the client. So I did
the changes. My fstab look like this:

jasmin:/home/claus /home/claus/jasmin nfs
rw,tcp,intr,nfsv3,-w=32768,-r=32768 0 0

The throughput rose to 77 Mbit per. sec.
/etc/rc.conf's nfs_server_flags changed to:

nfs_server_flags="-t -u -n 8"

Eventhough this may not be directly related to your
problem, you could try to switch to TCP and change the
client-fstab as mentioned above.

The server is a dual pentium III at 1 Ghz, 1.5 GB RAM,
client likewise. Running RC2 for a week now on both
client and server without any hickups. Server is
running two instances of seti_at_home to keep it at 100 %
CPU-util. Copied a 647 MB large file back and forth
while testing nfs-settings, diff'ed the file. No
problem what so ever.

regards
Claus


Yahoo! Mail (http://dk.mail.yahoo.com) - Gratis: 6 MB lagerplads, spamfilter og virusscan
Received on Mon Jan 12 2004 - 13:58:45 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:37:37 UTC