The test scenario differs from a genuine failure in that the local storage system is active during the test. Processing on the source units takes place in the local Data Center. The recovery procedures are tested in the remote Data Center using the target units.
CAUTION!
If the source units are not protected locally by RAID level, data loss can occur while the source and target units are split. If a system is started from a target unit, the SYSID of the pubset must first be changed using /SET-PUBSET-ATTRIBUTES (see the "Commands" manual [1]). Otherwise, problems can occur with sharable units.
Overview
A recovery procedure consists of the following steps:
Failure detection
A check to establish whether processing can be continued on the local system (with the source units)
Interrupting remote copy mode and enabling the target path
Shutdown of the local system, if necessary
Startup of the standby system and continuation of processing
(IPL, attach, import, change network node, restart applications)
The actions needed to interrupt remote copy mode and enable the target path are described on "Interrupting remote copy mode and enabling the target path". The actions needed to recover the source and target units are described on "Reconstructing the source and target units after the test".