Re: [coLinux-devel] Getting better and better

[Jason A. P. <pat...@pc...>]

Jason A. Pattie wrote:

> Jason A. Pattie wrote:
>
>> However, NFS does not seem to work properly (another thing I compiled 
>> into my custom kernel).  I was able to get it to mount the directory 
>> from the server, but the moment I issued an 'ls' on the mounted NFS 
>> filesystem, the network traffic indicators for the network 
>> connections had lots of activity, but the 'ls' commands died, cannot 
>> be killed, even with a -9, and umount on the NFS mount point does not 
>> work.  Says the device is busy.
>
>
>
> I think I'm getting closer.  Doing tcpdump's on both ends reveals that 
> my laptop is sending back to the NFS server an 'icmp: <laptop> 
> unreachable - need to frag (mtu 1500)'.  Don't know exactly what this 
> means, but possibly we might need to reduce the mtu size on eth0 in 
> coLinux?


I did an ifconfig on eth0 and noticed that MTU is one of the values.  I 
was able to alter it by specifying 'mtu' and a value on the ifconfig 
commandline.  So, I dropped it to 1400.  This time, I did an ls on a 
directory that I knew existed off the main mount point and had few files 
in it.  It worked!  I then increased the MTU back up to 1500 and did the 
same ls command on the small directory again.  It worked also!  I then 
dropped the MTU back down to 1400 and attempted the ls on the big, main 
mount point directory again, and it locked up again.  Don't know what's 
goin' on.  And I still got fragment messages coming back to the server 
for an (mtu 1500).


-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.
MailScanner thanks transtec Computers for their support.