[Samba] CTDB: Clustered NFS, reboot, requires me to exportfs -r(a)

[Charles Hammitt]

Hi Samba,

    I hope you are doing well.  I run a cifs / nfs CTDB clustered NAS 
solution, and I find that when I reboot any of the nodes in the cluster, 
I must re-export the nfs mounts so they show up properly.  Perhaps this 
is a general linux nfs bug and I am barking up the wrong tree, but I 
haven't found any problem / solution mentioning this as of yet besides 
my own known workaround of re-exporting once the service comes up.

    As you can see below, it exports the first host defined, but skips 
over the second two until I re-export.  Not sure if it is the space 
between the first host and the second host and the third, or if it is 
because my first host is defined by IP, and the second two are define by 
hostname.  But the cause is likely one or the other.

hostnames and ips are omitted, but you should get the idea.

*%cat /etc/exports *

/gpfs/nfs/share n.n.n.n(rw,root_squash,fsid=nnn) 
n.n.n.n(rw,root_squash,fsid=nnn) n.n.n.n(rw,root_squash,fsid=nnn)

*%showmount -e nnnn*

Export list for nnnn:
/gpfs/nfs/share n.n.n.n

*%exportfs -r*

*%showmount -e nnnn*

Export list for nnnn:
/gpfs/nfs/share n.n.n.n, n.n.n.n, n.n.n.n




regards,

Charles