Diagnosing disabled aggregates

If zFS detects a problem on an aggregate that is mounted read/write, zFS attempts to isolate the failure. As a result, zFS might mark an aggregate unavailable and issue message IOEZ00422E, as shown in the following example.

IOEZ00422E Aggregate PLEX.JMS.AGGR001.LDS0001 disabled

In addition, a dump and possibly zFS trace information might be generated. You can contact IBM® service and provide the dump and the trace and any other information that is useful for diagnosing the problem (for example, what was running on the system when the problem occurred).

When an aggregate is disabled, applications cannot read from, or write to, the aggregate. Other aggregates that are not involved in the failure remain available. However, the disabled aggregate is unavailable for reading and writing until it is unmounted and mounted. Beginning with z/OS® V1R13, if the disabled aggregate is zFS owned on a zFS V1R13 or later system, zFS attempts to automatically re-enable the disabled aggregate and make it available again for use.

zFS attempts an internal remount samemode on the zFS-owning z/OS V1R13 or later system in the following situations:
- It is in a non-shared file system environment
- The file system is non-sysplex aware
- The file system is sysplex-aware, but no other z/OS V1R13 or later system in the shared file system environment can take it over
Alternatively, in a shared file system environment where the file system is sysplex-aware, the zFS owning system requests that another system that is running z/OS V1R11 or later take over the aggregate.

The preceding re-enablement actions (aggregate movement or internal remount samemode) are taken only if the file system became disabled due to an internal zFS error or a corruption.

Even though the aggregate is disabled, z/OS UNIX System Services continues to display the aggregate mounted as R/W. To determine whether the aggregate has been marked as disabled, use the zfsadm lsaggr command or the zfsadm aggrinfo command.

An aggregate that was disabled might be corrupted, even if it was disabled and remounted. To be sure that the aggregate is internally consistent, run the ioefsutl salvage batch utility against the aggregate that was disabled, to repair any corruption, and prevent loss of data. See ioefsutl for more information.