NonStop Operations Guide for H-Series and J-Series RVUs

Anticipating and Planning for Change
Anticipating and planning for change is a key requirement for maintaining an enterprise-level, 24
x 7 operation. To avoid taking a NonStop NS-series system down unnecessarily:
• Evaluate system performance and growth—Track system usage and anticipate system capacity
and performance requirements as new applications are introduced.
• Provide adequate computer room resources—Avoid unnecessary downtime by ensuring you
have enough physical space and power and cooling capacity to handle future growth.
• Configure the system with change in mind—Configure the system in a way that easily
accommodates future growth. One way to do this is to select limits that allow for growth. For
example, by configuring enough objects to provide for the anticipated growth of your online
transaction processing environment, you can increase the maximum number of objects controlled
by PATHMON objects without a system shutdown.
Stopping Applications, Devices, and Processes
Whenever possible, schedule system shutdowns in advance so that system users are prepared.
Then, stop applications, devices, and processes in an orderly fashion. To include shutdown
commands in a shutdown file, see “Stopping the System” (page 191).
Unless you stop a system in a careful and systematic manner, you can introduce abnormalities in
the system state. Such abnormalities can affect disk file directories and can cause the processors
to hang in an endless loop when you attempt to load your system.
You must be aware of which processes must not or cannot be stopped. For example, some TCP
processes must not be stopped. System processes must not be stopped. Generic processes configured
to be persistent cannot be stopped.
Note the effect on the system when you stop these applications:
• Stopping Pathway applications begins shutdown of all TCP objects (shutting down TERM
objects and then themselves) in parallel. New work is disallowed. The PATHMON process
logs the start and completion of SHUTDOWN2. It does not log status messages during
shutdown.
• Stopping DSM/SCM stops the CNFGINFO server process, the Pathway environment for
DSM/SCM, the alternate EMS collector $ZPHI, and TCP/IP processes for DSM/SCM, as this
example shows:
STOP CNFGINFO server process $ZPHC
STOP DSM/SCM Pathway system $YPHI
PATHCOM $YPHI;SHUTDOWN !,WAIT
$Z02H: TCP TCP-H, STOPPED
$Z02H: TCP TCP-T, STOPPED
STOP DSM/SCM Alternate EMS Collector $ZPHI
• Following the SPOOLER DRAIN command, the collectors allow current jobs to finish but reject
new opens with a file-system error 66 (device downed). When you drain the spooler, each
collector stops when it has no more open jobs. Each print process finishes printing any active
jobs and then stops. After all collectors and print processes have stopped, the supervisor stops.
The spooler enters the dormant state, ready to be warm started.
• Following the SCF CONTROL, DISK REFRESH command, all other disk I/O is suspended. The
amount of time a refresh operation takes to finish depends on the amount of disk cache
containing dirty pages in use at the time, and writing to disk can take several minutes.
Stop processes and applications in this order:
1. After you send a message alerting users of the shutdown, stop all user applications.
2. If your system is equipped with Pathway, stop Pathway applications. At the Pathway prompt:
190 Starting and Stopping the System