Port-sgimips archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: mec0 dies under NFS load



Michael Lorenz wrote:
On Feb 27, 2008, at 15:25, Stephen M. Rumble wrote:

Quoting Martin Husemann <martin%duskware.de@localhost>:

When I do a "build.sh build sets" on another machine and then copy the
binary sets via NFS, my machine silently resets. The same happens when
doing a "installworld".

Copying the sets via scp works though.

I haven't yet tested other archs, but my gut feeling blames if_mec ;-)
Please prove me wrong.
[...]
I've got an ex in my O2, never used it since tsutsui-san committed something that made mec go fast enough for me ( now I get 4MB/s sustained, before it would eventually slow down to something like 8kB/s ).

Likewise here; my O2 is my music server and sucks the entire music spool off my sparc64 box via NFS, and while this probably isn't heavy load, it does quite fine and even survived the NFS server rebooting nicely (where as my other sparc64 box paniced in a sympathy death ;)). I haven't done any NetBSD builds on that machine since it's now above my pain threshold (too damn slow!), but I did upgrade it from a build done on the Netra over NFS without any hiccups. It's also built some pkgsrc stuff over NFS (since we can't yet cross-build all of that ;)) without issues.. I've got some 75MB of packages on it.

In fact before kern/38088 (or maybe me upgrading it, don't remember which came first), my O2 had been up about a month straight. It's a R5kSC @ 180Mhz.


have fun
Michael


Home | Main Index | Thread Index | Old Index