>>> One thing I don't like about your setup is that you have 2 different >>> machines serving NFS directories to each other.
> This seems not to have elicited any response, pro or con. I know of > no reasons in principle why two machines can't simultaneously act as > NFS clients and NFS servers - are there any? AFAIK the two subsystems > are separate and have no dependencies or interactions; does anybody > know otherwise? The only good reason to avoid that involves non-backgroundable mounts. If system A mounts something from system B and system B mounts something from system A, things work fine as long as one system is up. If there's a power failure then at powerup the two systems will deadlock if they do their NFS mounts before starting the NFS server. And I think that's the typical system startup semantics, at least on Unixes. Also, woe to the NIS server that depends on an NFS mount from one of its clients. Sysadmins have have mostly learned to do backgroundable mounts in startup scripts and automount/autofs has improved things quite a bit further with they mount-as-needed semantics. One thing I do strongly recommend against is using the SGI (copied in Linux) "nohide" option in /etc/exports as that leads to file systems where an inode number may not be unique and all my experiences with it produce more confusion than it's worth. -Ric _______________________________________________ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
