Re: NFSv3 client for Linux-2.2.5 ready for alpha testing...

Andrew Schretter (schrett@math.duke.edu)
Fri, 16 Apr 1999 10:12:33 -0400 (EDT)


> This argument half-convinced me that support for rsize>4k is needed
> for the Solaris case. I've therefore implemented it in the latest
> version (0.7.0) of the NFSv3 patches:
> Please could Andrew and others who've been seeing slow reading wrt
> Solaris machines test this code out. If it makes no difference, I'll
> scrap it...

I installed the new patches and tried out the kernel.
The first reboot resulted in a kernel panic and a reboot.
(Sorry I couldn't catch it). The second reboot didn't
panic but gave a message about the interrupt handler then
rebooted.

I tracked both of these down to trying to use NFSv2 mounts with 8192 read
sizes from a linux server. So I booted without mounting anything. Tried
using NFSv3 mounts to a solaris server. This worked, mostly, but produced
the following errors :

Writing to the solaris server was fine.
Reading produced the errors :

NFS: Odd RPC header size in read reply: 128

This repeated several times.

Then I tried mounting NFSv2 from a solaris server with
32768 rsize. It mounted and wrote with no errors.

Reading produced the errors :
RPC: garbage, retrying 2553
RPC: garbage, retrying 2553
RPC: garbage, exit EIO

These repeated with different increasing values of 9 (the next value
was 2562 and so on).

Finally, to try to reproduce the crashes above, I tried
mounting a linux server via NFSv2 with an 8192 rsize,
then copy a file from it. It locked up the system hard with
no error messages.

Hope this helps.

Andrew Schretter
Systems Programmer, Duke University
Dept. of Mathematics (919) 660-2866

-
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/