Unsolved
This post is more than 5 years old
614 Posts
1
7829
solaris super cluster server cannot mount /nfs mount on vnx even though...
Hello,
I created an NFS mount point on the VNX, added the hosts for read/write, root privileges. A new Oracle Super Cluster has been installed and I added the Super Cluster hosts.
We opened network ports for NFS to/from the the hosts to the VNX control station and primary data mover. Yet they still cannot mount the slices. Other hosts(not on the Super Cluster) on a different network can mount the nfs slices.
They are also using data pump.
Any input you have is greatly appreciated.
Admin
dynamox
2 Intern
2 Intern
•
20.4K Posts
0
June 16th, 2016 14:00
super cluster, sounds important
can they run "showmount -e datamover_interface" from one of these hosts ? This command will display all nfs exports available on that datamover.
admingirl
614 Posts
0
June 16th, 2016 14:00
Thank you Dynamox! I will ask him to do that now.
admingirl
614 Posts
0
June 16th, 2016 15:00
I got the group to give me a login and I executed the command it just hangs. The version is Solaris 11.
zhaos2
643 Posts
1
June 16th, 2016 19:00
If other hosts can mount the NFS from the same VNX, it is more likely that the issue is with Solaris itself or network side. Here are steps for troubleshooting:
# /usr/sbin/rpcinfo -p server_2 |egrep "portmap|nfs|mount"
100003 2 udp 2049 nfs
100003 3 udp 2049 nfs
100003 2 tcp 2049 nfs
100003 3 tcp 2049 nfs
100005 3 tcp 1234 mountd
100005 2 tcp 1234 mountd
100005 1 tcp 1234 mountd
100005 3 udp 1234 mountd
100005 2 udp 1234 mountd
100005 1 udp 1234 mountd
100000 2 udp 111 portmapper
100000 2 tcp 111 portmapper
If there is no output or it is timed out, the issue should be that the network router between Solaris and VNX filters traffic on these ports.
As per the problem description, I suppose the issue would be found from step 1 or step 2.
admingirl
614 Posts
0
June 16th, 2016 19:00
2049, 4045 are for data pump
admingirl
614 Posts
0
June 16th, 2016 19:00
baulisano@efsdzdbclient010101:/etc$ cat vfstab
#device device mount FS fsck mount mount
#to mount to fsck point type pass at boot options
#
/devices - /devices devfs - no -
/proc - /proc proc - no -
ctfs - /system/contract ctfs - no -
objfs - /system/object objfs - no -
sharefs - /etc/dfs/sharetab sharefs - no -
fd - /dev/fd fd - no -
swap - /tmp tmpfs -
dynamox
2 Intern
2 Intern
•
20.4K Posts
1
June 16th, 2016 19:00
what firewall ports did you open ? Something is still blocking it.
admingirl
614 Posts
0
June 16th, 2016 19:00
Thank you Dynamox, we opened
(UDP, TCP both ways) on the primary data mover and control station. Did not open for the failover datamover. Could that be causing it?
111
4045
1110
admingirl
614 Posts
0
June 16th, 2016 19:00
Thank you! Here is my output:
baulisano@efsdzdbclient010101:~$ ping 10.129.226.16
10.129.226.16 is alive
baulisano@efsdzdbclient010101:~$ /usr/sbin/rpcinfo -p 10.129.226.16
-bash: /usr/sbin/rpcinfo: No such file or directory
baulisano@efsdzdbclient010101:~$ rpcinfo -p 10.129.226.16
program vers proto port service
824395111 1 tcp 50682
824395111 1 udp 49628
100011 1 tcp 55233 rquotad
100011 1 udp 62062 rquotad
536870914 1 udp 4658
536870914 1 tcp 4658
100021 3 udp 63498 nlockmgr
100021 2 udp 63498 nlockmgr
100021 1 udp 63498 nlockmgr
100021 4 udp 63498 nlockmgr
100021 3 tcp 64240 nlockmgr
100021 2 tcp 64240 nlockmgr
100021 1 tcp 64240 nlockmgr
100021 4 tcp 64240 nlockmgr
100024 1 udp 49162 status
100024 1 tcp 49162 status
100003 2 udp 2049 nfs
100003 3 udp 2049 nfs
100003 2 tcp 2049 nfs
100003 3 tcp 2049 nfs
140391 1 udp 31491
100005 3 tcp 1234 mountd
100005 2 tcp 1234 mountd
100005 1 tcp 1234 mountd
100005 3 udp 1234 mountd
100005 2 udp 1234 mountd
100005 1 udp 1234 mountd
536870919 3 tcp 12345
536870919 1 tcp 12345
536870919 3 udp 12345
536870919 1 udp 12345
102660 1 udp 50170
102660 1 tcp 56752
100000 2 udp 111 rpcbind
100000 2 tcp 111 rpcbind
baulisano@efsdzdbclient010101:~$
It looks good, however, showmount -e just hangs. It's so odd, again, the nfs mount point is working on the other network.
Thanks so much!
zhaos2
643 Posts
0
June 16th, 2016 21:00
It sounds interesting. Is this a production system? Is it possible to have SSH access to login have a look?
dynamox
2 Intern
2 Intern
•
20.4K Posts
0
June 17th, 2016 10:00
4045
1110
you don't need control station for NFS/CIFS connectivity, management only. Standby datamover will assume primary datamover's IP address so nothing special there either.
Could your firewall guys temporarily create a rule to all all ports between that host and the interface on VNX. If firewall is wide open for that host that it could be host related, is there a local firewall on the host itself ?
admingirl
614 Posts
0
June 17th, 2016 17:00
So I can login, but I found that the VNX file NFS default port is 2049. The Oracle super cluster is using 4045. My legacy linux servers use 2049.
I don't think I can change the default port and if I do, other stuff will not be working I don't think.
Admingirl