> "Steven N. Hirsch" <shirsch@adelphia.net> writes:
>
>
> > Yes, it does go rather fast. Unfortunately, where it leads me is into
> > oblivion. Running 'iozone 64 8192' on a dual-PPro SMP client (2.2.1-ac3)
> > against a knfsd server locks up solid after a few seconds. No log
> > messages, no fanfare.
> >
>
> If you have time, I'd like to work with you on this.
>
> I had reports from Thierry Danis that seemed to imply that the
> nfs_write_errors list might be getting corrupted at times. He said
> that he was seeing return values of '1' rather than 8192 when he
> 'strace'd iozone. Furthermore, the kernel NFS debugging seemed to
> imply that 'nfs_updatepage' was not being called when this happened
> (although nfs_file_write' was). Are you seeing the same thing?
I'm not sure I have enough free disk space for an strace of iozone <g>.
At any rate, I can't seem to trigger the lockup again after the initial
occurance.
Regarding NFS debugging, is there a centralized, coherent explanation of
the various NFS and RPC debug flags? If you have a suggestion for
configuring this, I'll glady feed you the resulting data.
Currently, a dual-PPro SMP client running ac4 is giving Bonnie benchmarks
of > 6 MB/sec. write and > 4 MB/sec. reads against a P166 server over
half-duplex 100Base-TX. Not too shabby.
I'll continue to watch it closely.
Steve
-
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/