OSI/MHS Configuration and Management Manual

Sizing and Tuning Your OSI/MHS Subsystem
OSI/MHS Configuration and Management Manual—424827-003
8-17
Peripheral Utility Program (PUP)
Peripheral Utility Program (PUP)
Use the PUP LISTCACHE command to monitor disk cache usage. If the proportion of
cache hits is too low (less than 96-97 percent), dedicate more memory to the cache.
Subsystem Control Facility (SCF)
Use SCF to monitor the following aspects of OSI/MHS:
•
Link and route-retry queues: STATUS ENTRY command. Long queues,
accompanied by nondelivery reports to applications, suggest that communication
links are not reliable enough; that you need to define more links, more routes, or
both; or that attributes affecting retry timers have values that are too low.
•
MR group internal queues: STATS PROCESS MRP command. Long internal
queues can sometimes suggest that there are too few of a certain kind of
process—an MS or GI group, for example—to handle the actual message traffic, or
that work is spread unevenly among the available processes. For example, there
might be too many APPL objects or too large a range of addresses assigned to the
same gateway.
•
Connection rejections: STATS PROCESS RTS command. Section 9,
Troubleshooting Your OSI/MHS Subsystem, includes an extensive discussion of
how to troubleshoot connection rejections.
Event Management Service (EMS)
Use Viewpoint, a printing distributor, or the EMS Analyzer to monitor event messages
from OSI/MHS and related subsystems. Especially significant are event messages
related to routing and resource management:
ZMHS-EVT-ACCT-MRP-ROUTE
ZMHS-EVT-CROSS-THRESHOLD
ZMHS-EVT-UNROUTE
ZMHS-EVT-UNROUTE-RECIP
Another event of interest is ZMHS-EVT-INTL-WARN with any of the following error
codes:
See the OSI/MHS Management Programming Manual for detailed descriptions of
these and other OSI/MHS event messages.
7110 TRYING-NEW-PDU-STORE Can indicate that the PDU store is full.
7111 TRYING-NEW-CONNECT Can indicate that the MR group is overloaded.
7112 TRYING-NEW-MRP-SEND Can indicate that the MR group is overloaded.
7113 TMF-ERROR Can indicate that the audit file is too small.