Using the four-system complex shown earlier in Figure 1, assume that it is necessary
to purge SYS1 from the ring. SYS1 is an active system. The operator
on SYS2 issues the purge command; thus, SYS2 is the assisting system.
The required steps are:
- On SYS1, the operator, to avoid potential data integrity
exposures, must bring the workload to an orderly shutdown by taking
such actions as stopping the subsystems and terminating jobs.
- On SYS2, the operator then issues VARY GRS(SYS1),PURGE.
The following messages appear:
ISG100E SYSTEM SYS1 IS STILL AN ACTIVE GRS SYSTEM
ISG101D CONFIRM PURGE FOR ACTIVE SYSTEM SYS1 - REPLY NO OR YES
In
this example, the operator can safely reply YES. These messages remind
the operator that SYS1 is still active; purging it from the ring might
create a data integrity exposure.
- On SYS2, the operator replies YES to message ISG101D.
- On all active systems, the following messages appear:
ISG011I SYSTEM SYS1 - QUIESCING GLOBAL RESOURCE SERIALIZATION
ISG013I SYSTEM SYS1 - QUIESCED GLOBAL RESOURCE SERIALIZATION
ISG011I SYSTEM SYS1 - BEING PURGED FROM GRS COMPLEX
ISG013I SYSTEM SYS1 - PURGED FROM GRS COMPLEX
- On SYS1, the operator must stop the system. At this point,
SYS1 is no longer known to global resource serialization. The ring
consists of SYS2, SYS3, and SYS4. To rejoin the ring, SYS1 must reIPL
with GRS=JOIN.
While purging SYS1, global resource serialization
might detect a potential data integrity exposure. After purging an
active system, the operator should follow the installation's procedures
for resolving a data integrity exposure, such as contacting the system
programmer responsible for investigating the state of resources.
More information about the problem normally appears in SYSLOG, where
the following message appears to describe any resources that might
have been damaged by purging the system:
ISG018I REQUESTORS FROM SYSTEM SYS1 HAVE BEEN PURGED FROM RESOURCES
NAMED xxxx,yyyy