Beginning in z/OS® V1R13, zFS provides enhanced sysplex-aware support. When a zFS read/write file system is mounted sysplex-aware in a shared file system environment where all systems are running z/OS V1R13 or later, zFS can directly read and write zFS data from all of the V1R13 or later systems. If both the owning system and the requesting system are running z/OS V1R13 or later (and the file system is sysplex-aware), zFS directly accesses the file system. While zFS data is directly read and written, zFS metadata is normally read and written through the zFS owning system (SY2 in Figure 1). In some cases, zFS metadata can be directly read. If the zFS owning system or the requesting system are running a release of z/OS earlier than z/OS V1R13, zFS function ships requests to the zFS owning system.

Figure 1. zFS sysplex-aware file system being directly accessed from all z/OS V1R13 systems

Figure 1 is similar to Figure 1; however, all systems (SY1, SY2 and SY3) are running z/OS V1R13 or later. In this case, the sysplex-aware file system (FS2) is zFS owned by SY2 and user data is being directly accessed from all three systems.

There are some cases when an application running on a system (SY1) that is doing direct I/O can be affected by problems on the zFS owning system (SY2) such as a failing system or having I/O failures on the owning system during metadata updates. The application can also be affected if it needs to traverse a higher level directory contained in a file system that is owned by the failing system.