IBM SC33-1683-02 Server User Manual


 
related regions to be taken over, because the new active region can continue
communication with the other active regions. Takeover is therefore likely to be your
preferred course of action.
Enabling and disabling restart in place:
The restart-in-place function of the
overseer program can be enabled and disabled using the Restart command. When
you enter this command, restart processing is enabled or disabled for all generic
applids that the overseer is monitoring. You can also specify that particular
active-alternate pairs are not to be automatically restarted in place, regardless of
whether restart processing is enabled or disabled. This is described in “How to tell
the overseer which actives and alternates to monitor” on page 613.
The Restart command works like an ON/OFF switch. Restart in place is enabled
when the sample program is initialized. When the Restart command is first entered,
restart in place is disabled. If you issue the command again, restart is enabled
again, and so on. If a region fails while restart in place is disabled, no attempt to
restart it is made, even if restart in place is enabled again.
Rules that control restart in place:
The sample overseer program concludes that
a region has failed if both:
v The region is not executing now, and was known to have been executing during
the previous examination of the relevant CAVM data sets by the overseer.
v The region did not sign off normally from the CICS availability manager (CAVM).
The overseer program can restart a failed active region in place, if all the following
conditions are met:
v Restart in place is enabled for this overseer.
v Restart in place is enabled for this active-alternate pair.
v There is no executing alternate region for this active, or the alternate region is
currently defined with TAKEOVER=COMMAND. If the alternate region is defined
with TAKEOVER=AUTO or TAKEOVER=MANUAL, the overseer assumes that
the alternate will initiate a takeover or that the console operator will decide what
action to take.
v The failing region was running in the same MVS image as the overseer.
v An attempt to restart the region in place is not already in progress.
v If the failing region belongs to a group of related regions (an MRO environment,
for example), a takeover to another MVS image, perhaps initiated by another
region, is not under way.
When a failed active region is restarted in place, whether by the operator or by the
overseer, the corresponding alternate region cannot continue to support the new
active region, and must be restarted. The overseer program restarts the alternate
region automatically in these circumstances, if restart processing is enabled for both
the failing region and the overseer.
If you want to be able to restart regions in place in both MVS images in a
two-image environment, an overseer program must execute in each image.
If the failed region was started originally as a started task, the overseer program
restarts it as a started task, and if the failed region was started as a job, the
overseer restarts it as a job. For more guidance information about how the sample
overseer program restarts failed regions in place, refer to the
CICS Operations and
Utilities Guide
.
the XRF overseer program
612
CICS TS for OS/390: CICS Customization Guide