Solaris

Hi,

I have a Solaris 10 server and I want to export a filesystem to a linux client and give the client's root user root priviliges on the filesystem.
The client is an ubuntu 14.04 LTS server.
the dfstab on the server looks lik this:

Code:

 /usr/sbin/share -F      nfs     -o      sec=sys,rw,root=something.somewhere.nl:@192.168.1.0 -d "user home dirs"    /home/t4   -

The linux server is on the 192.168. network. The fstab line looks like:

Code:

tmku201:/home/t4    /home/tmku201    nfs    defaults    0    0

The mounting works perfectly, but root of the Linux server can not delete files on /home/tmku201.

Do you have any ideas?
Regards,
Nico

---------- Post updated at 09:27 AM ---------- Previous update was at 09:00 AM ----------

Replying to my own post I think I fixed it.
I changed the fstab line on the Linux server:

Code:

192.168.1.101:/home/t4    /home/tmku201    nfs    defaults    0    0

You know, my servers have the first network interface attached to the university network and the second to the 192.168. backend.

And in the first fstab I mixed things up a bit. tmku201 is the hostname on the university network for the Solaris server and I want to mount over the backend.
Strange though that the mount succeeded at all

Rregards,
Nico

Thanks for sharing your finding.
You can also give a hostname that is displayed with

Code:

getent hosts 192.168.1.101

You are right, it is a bit strange that the mount works. The request is routed to the NFS server's external address and returned from the other address. Maybe this is checked/rejected by the NFS client.
Note that usually the mount request uses UDP for rpcbind and mount RPC, while the NFS RPC uses tcp. But there shouldn't be a difference in the routing (but there could be one in a fire wall).