Working With File Modules That Report A Stale Nfs File Handle - IBM Storwize V7000 Unified Problem Determination Manual

About this task
After a client IP change, a df -h command can return no results, as shown in the
following example:
The ls command can return the following error:
ls: .: Stale NFS file handle
The Storwize V7000 Unified system hosting file module might display the
following error:
If one of these errors occurs, complete the following steps.
Procedure
1. Access the file module, using root privileges, that hosts the active management
2. Issue the onnode all /usr/sbin/exportfs -a command to flush the NFS cache
3. Verify that the NFS mount is successful. If the problem persists, restart the NFS
4. Verify that the NFS share mount is successful.

Working with file modules that report a stale NFS file handle

To recover from the "Stale NFS file handle" file system state on a file module, you
must suspend, reboot, and resume the file module.
About this task
Note: If the "Stale NFS file handle" message was displayed after a client IP
change, refer to "Working with NFS clients that fail to mount NFS shares after a
client IP change" on page 275.
Because of errors or conditions related to this file module, the file module
disconnected itself from the file system that was shared with the other nodes. All
of the file descriptors that were opened to the file system through this file module
have become "stale", as indicated by command output or a Stale NFS file handle
error message, and cannot access their corresponding files. When this occurs, all
affected file modules enter an unhealthy state, and a CIM similar to the following
is sent to the alert log:
GPFS Error - check stale file handle failed with error code 1:
If you receive the error above, complete the following steps:
Procedure
1. Open the CLI in the active management node using root privileges, and issue
276
Storwize V7000 Unified: Problem Determination Guide Version
Filesystem
machinename: filename: -
mgmt002st001 mountd[3055867]: refused mount
request from hostname for sharename (/): not exported
node role.
in each file module.
service on the file module that is refusing the mount requests from that client.
see stale file handle on /ibm/gpfs0 on file module: mgmt001st001
/usr/sbin/exportfs -a to flush the NFS cache in each file module. Verify that
the state of each affected file module is healthy and that no new "Stale NFS file
handle" CIMs are displayed in the alert log after you resume the file module. If
the problem persists, continue with the following steps.
Size Used Avail Use% Mounted on
- - - /sharename