Eric Chris Garrison wrote:
> One more interesting fact, we suspected a MTU mismatch... our AFS servers
> are 1500 MTU (running servers with -nojumbo arguments), and the
> supercomputer in question are set to 9000 MTU.
> 
> When we had the admins set a node to 1500 MTU, the problem went away.
> Since this isn't practical for the whole cluster, we had them change it
> back to 9000 and ran the client with the following set:
> 
>    AFSD_ARGS="-rxmaxmtu 1500"
> 
> However, this (9000 MTU + rxmaxmtu 1500) hangs just as before.
> 
> Any more ideas, with this new information?

IP MTU 1500 == Rx Mtu 1444.

Try setting it to 1444.






_______________________________________________
OpenAFS-info mailing list
[email protected]
https://lists.openafs.org/mailman/listinfo/openafs-info

Reply via email to