zFS running in the z/OS UNIX address space
In releases before z/OS V2R2, the amount of 31-bit virtual storage that was needed by both z/OS UNIX and zFS combined would have exceeded the size of a 2 GB address space. Due to that size limitation, zFS and z/OS UNIX could not coexist in the same address space.
In z/OS V2R2, zFS caches were moved above the 2 GB bar into 64-bit storage. You can now choose to have zFS run in its own colony address space or in the address space that is used by z/OS UNIX, which is OMVS.
When running zFS in the OMVS address space, each file system vnode operation (such as creating a directory entry, removing a directory entry, or reading from a file) will have better overall performance. Each operation will take the same amount of time while inside zFS itself. The performance benefit occurs because z/OS UNIX can call zFS for each operation in a more efficient manner.
- MODIFY commands must be passed to zFS through z/OS UNIX. Use the form
MODIFY OMVS,pfs=zfs,cmd. For more information, see Passing a MODIFY command string to a physical file system in z/OS MVS System Commands. This form of the MODIFY command can be used whether zFS is in its own address space or in the OMVS address space.Note: When zFS is running in the OMVS address space, any zFS MODIFY commands that are issued through an automated process or system automation must be changed to accommodate the new command format. - The CANCEL ZFS command is not available.
- When the IOEFSPRM configuration file location is defined by the IOEZPRM DD card, it must be placed in the OMVS PROC. For more information, see The zFS configuration options file (IOEPRMxx or IOEFSPRM).
zFS will run under the OMVS user ID.
- You can determine if zFS is in its own address space by issuing D OMVS,PFS. If the output shows an ASNAME value, zFS is running as a colony address space. Otherwise, the lack of an ASNAME value means that zFS is running in the OMVS address space.