[ale] NFS mount problems

[Ryan Fish]

For the past couple days a RHEL 3 AS box running Oracle 9.x.x.x has been
having a lot of trouble maintaining NFS mounts with a NAS on the same
network (in the same VLAN).  When the RHEL box is rebooted the mounts are
fine but eventually attempts to access them time out.  The load averages on
the Oracle server just keep rising until it finally stops responding to
requests from the application.

 

- I am able to ping the NAS from the Oracle server (and vice-versa) with no
packet loss or other errors

- I am able to telnet on port 2049 (NFS) from the Oracle server to the NAS
(and vice-versa)

- "rpcinfo -p" shows all required services are running on both the Oracle
server and the NAS

- "mount" on the Oracle server shows the two mounts to the NAS as being
available

- When I run "mount" manually on the two mount points they are shown twice
by "df -h"

- All other servers in the network are able to keep their mount points with
the NAS without any trouble

 

- From /etc/fstab:

nas01:/mc/backups/systems/db02/exp /mc/backups/database/mc/exp_bk/exp nfs
rw,hard,intr 1 2

nas01:/mc/backups/systems/db02/rman /mc/backups/database/mc/exp_bk/expold
nfs rw,hard,intr 1 2

 

These mounts points have worked fine for months.  Nothing has changed with
them or the config of either server (Oracle or NAS).

 

As far as the network is concerned, the switch these boxes were using died
Thursday, April 26th so they were moved to a second switch.  The second
switch is known to have issues with invalid_crc errors that cause things to
go haywire so a brand new switch was configured and installed in the rack.
All servers were moved to this new switch on Friday, May 5th.  From what I
can see the new switch is not to blame as /var/log/messages on the Oracle
server show this problem going back through Tuesday, May 2nd.  Also there
are no errors, etc. reported in the logs on the switch and, as stated above,
none of the other servers on the switch are experiencing any difficulty
keeping their mounts with the NAS in place.

 

What am I missing here?

 

Thank you.

-Ryan

-------------- next part --------------
An HTML attachment was scrubbed...