There can be times when PPRC cannot replicate an update to a primary
device onto the secondary device. In these cases, the primary device
sends channel end, device end, and unit check status to the host that
encountered the error. With the PPRC error recovery procedures (ERP)
PTF that is installed, the host system performs the following actions:
- The system stops all application I/O, for that
host, to the primary volume in error. The host system also prevents
other system images from accessing this volume.
- The system writes an IEA49xx message to
the log to indicate that the PPRC volume is now in suspended state
or duplex state. (Ensure that the systems' log is common to both
the primary system and the recovery system.)
- The system puts information that is related to the specific failure
into the SYS1.LOGREC data set for service personnel reference. SYS1.LOGREC
can be common to both the primary system and the recovery system.
- The system waits for enough time to let the IEA49xx message
reach the recovery system. You can introduce an automation routine
here that would receive control before the I/O operation can complete.
- The system resumes all host application I/O to the primary volume.
If you specified the CRIT(NO) parameter when you established the primary
volume, the storage control allows subsequent I/O operations to continue
without being unit checked. If you specified the CRIT(YES) parameter
on the CESTPAIR command for the primary volume, then the storage control
does the following:
- Checks all subsequent write I/O operations.
- Allows the host application to follow its own recovery actions.
Note: An "A" in the unformatted CQUERY message
ANTP0091I indicates that the volume pair is designated as CRIT(YES-ALL).
This means that the pair was established with the CRIT(YES) option.
The storage control maintenance panel is set to inhibit writes on
any failure, which includes secondary device failures.