When zFS can no longer obtain sufficient storage to complete a request, it issues message IOEZ00188A, possibly creates a dump, and restarts. If you see message IOEZ00188A before zFS initialization is complete (before message IOEZ00055I), either increase the REGION size in the ZFS PROC or decrease some cache sizes in the IOEFSPRM configuration file.

In addition, the zFS hang detector periodically checks a warning limit and a critical limit. When it reaches the warning limit, message IOEZ00662I displays and remains on the console until the situation is resolved, or until the critical limit is reached. If the critical limit is reached, message IOEZ00663I displays and remains on the console until storage usage goes below the critical limit to the warning limit, and then message IOEZ00662I displays again. See STOR for more information about how to determine the amount of storage being used in the zFS address space.

A zFS storage shortage can be caused by the number of active vnodes in use in zFS. You can query the number of held vnodes using either the MODIFY ZFS,QUERY,LFS system command or the zfsadm query -vnodecache command. You can also query the current sizes of the zFS caches in the zFS address space using the zfsadm configquery command with its cache size parameters, such as -tran_cache_size and -vnode_cache_size. For example, zfsadm configquery -meta_cache_size returns the metadata cache size. When zFS is running in a shared file system environment, you can query the client reply storage using zfsadm configquery -client_reply_storage. You can also determine cache sizes by using the MODIFY ZFS,QUERY,STORAGE command. Decreasing one or more cache sizes might relieve the zFS storage shortage.

In general, if you see a return code of 132 (ENOMEM), zFS is short on storage; take steps to reduce zFS storage usage. When storage shortages become critical, you can also see 157 (EMVSERR) and mounts might begin to fail.