Checking the status of zFS file systems and NFS

There are various commands with which you can check the status and further attributes of a zFS file system.

You can check the attributes of a zFS file system with the following command:

df -kv <filename>

See an example of the output: The z/OS® UNIX file system ownership is on COH3 and the file system is movable:

ha2adm> df -kv /usr/sap/HA2
Mounted on     Filesystem                Avail/Total    Files      Status
/usr/sap/HA2   (OMVS.ZFS.COHPLEX.HA2.USRSAP) 203197/1024560 4294966793 Available
ZFS, Read/Write, Device:24, ACLS=Y
File System Owner : COH3        Automove=Y      Client=Y
Filetag : T=off   codeset=0
Aggregate Name : OMVS.ZFS.COHPLEX.HA2.USRSAP

To verify, whether a file system is defined sysplex-aware, use the following command:

zfsadm aggrinfo -long -aggregate OMVS.ZFS.COHPLEX.HA2.USRSAP

For a sysplex-aware file system, the command output is similar to the following:

OMVS.ZFS.COHPLEX.HA2.USRSAP (R/W COMP): 13626955 K free out of total 14353200
version 1.5
auditfid C3D6C8D6 D7C50E14 0000
sysplex-aware
       1703355 free 8k blocks;        115 free 1K fragments
         10248 K log file;             72 K filesystem table
          1992 K bitmap file

The following command allows the operator to check whether NFS clients have mounted a file system, (<MVSNFS> stands for the job name of the NFS server):

F <MVSNFS>,LIST=MOUNTS

Consider the case where clients may not have done an explicit unmount (for example, if the connection was disrupted, or the client system was switched off). This usually does not impact the NFS server.

However, if an HFS data set is unmounted and then remounted to the z/OS system, the NFS server does not allow NFS mounts to the newly available file system if any old NFS mounts are active.

The mount count is reset, and unmounts are forced with the following command:

F <MVSNFS>,UNMOUNT='/HFS/<mountpoint>'