Incorrect or unexpected stacks are being monitored

When you configured the IBM® Z OMEGAMON® AI for Networks monitoring agent, you specified what data you wanted collected for each TCP/IP stack to be monitored, but after the agent is initialized, you find that the agent is monitoring all stacks.

By default, the IBM® Z OMEGAMON® AI for Networks monitor agent monitors all TCP/IP stacks discovered for an LPAR. If you do not provide data collection options for a specific stack, the global options you specified are used. The list of stacks in &hilev.&midlev.&rtename.RKANPARU member KN3TCPMO does not identify the stacks to be monitored. Instead it identifies monitoring overrides for the stacks listed.

When you use the Configuration Tool to configure the IBM® Z OMEGAMON® AI for Networks agent, you provide global monitoring options by navigating to the SPECIFY CONFIGURATION PARAMETERS panel and then navigating to the SPECIFY COMPONENT CONFIGURATION panel and selecting option 1, Specify component configuration. These options are used to monitor all TCP/IP stacks discovered unless you provide overrides for specific stacks.

To update global values using the Configuration Tool, perform the following steps:
  1. Start the Configuration Tool and go to the CONFIGURE IBM® Z OMEGAMON® AI for Networks/RTE: rtename panel.
  2. From the CONFIGURE IBM® Z OMEGAMON® AI for Networks/RTE: rtename, choose 2 SPECIFY CONFIGURATION PARAMETERS.
  3. From the SPECIFY CONFIGURATION PARAMETERS/RTE: rtename panel, choose 2 SPECIFY TCP/IP MONITORED SYSTEMS INFORMATION. This panel is where global values are set.
  4. Specify an A or a U to display the ADD/UPDATE TCP/IP MONITORED SYSTEMS INFO panels.
  5. Check whether the value for Do you wish to monitor this stack is No. If it is, change it to Yes. All stacks defined in the RTE are then monitored.
  6. Choose 3 Create monitoring agent runtime members.
  7. Submit the JCL that is presented.
  8. Exit the Configuration Tool.

To provide overrides for a specific stack, navigate to the SPECIFY CONFIGURATION PARAMETERS panel and then navigate to the SPECIFY TCP/IP MONITORED SYSTEMS INFORMATION panel. On this panel you can specify an A or a U to display the ADD/UPDATE TCP/IP MONITORED SYSTEMS INFO panels. You use the panels to specify monitoring options that are saved in &hilev.&midlev.&rtename.RKANPARU member KN3TCPMO and run the create runtime.

When you use PARMGEN, to configure the IBM® Z OMEGAMON® AI for Networks agent, perform the following steps:
  1. Locate the configuration profile for this instance of the monitoring agent. See Step 1 under Cannot connect to location server or find running CMS on CT_CMSLIST to determine how to locate the configuration profile.
  2. Open the configuration profile and look for the following parameters to identify which TCP/IP stack is being monitored. Two sets of values are found: default values for all TCP/IP stacks, which are used unless other values are explicitly configures, and the override values. These parameters define the name of the address space and the global values:
    ** Values that describe the address space:
KN3_AGT_CONFIGURATION_MODE        STANDALONE
KN3_AGT_STC                       CANSN3

** IBM® Z OMEGAMON® AI for Networks Agent parameters:
** TCP/IP Information:
KN3_TCP_VIO_UNIT                  VIO
KN3_TCP_SAMPLE_INTERVAL           5
KN3_TCP_COLLECT_STACK             Y
KN3_TCP_CONN                      Y
KN3_TCP_IPSEC                     N
KN3_TCP_ROUTE_TBL                 Y
KN3_TCP_ROUTE_TBL_FREQ            10
KN3_TCP_FTP                       Y
KN3_TCP_FTP_DSPINTV               2
KN3_TCP_TN3270                    Y
KN3_TCP_TN3270_DSPINTV            2
KN3_SNMP_CONFIG_FILE "MFN.N3510BAS.NEW510.RKANSAMU(KN3SNMP)"
KN3_TCP_EEHPR                     Y
KN3_TCP_ALLHPR                    N
KN3_TCP_CSM                       Y
KN3_TCP_OSA                       Y
KN3_TCP_INTS                      Y
KN3_TCP_INTE                      Y
KN3_TCP_GLBS                      Y
KN3_SNA_VTAM_COLLECT_DATA         Y
KN3_SNA_VTAM_SNAC_SNACINTV        5
    These parameters are used to specify the override values:
    ** Define TCP monitoring systems member:
** Specify KN3_TCPXnn_* row for each TCP/IP monitored system:
** Global default is $$$$ (monitored all TCP/IP stack):
** Note: Each KN3_TCPXnn_ROW and KN3_TCPXnn_* set equates to a table
**       row.  Increment the nn value accordingly.
KN3_TCPX                         BEGIN                 * Table begin  *
KN3_TCPX01_ROW                   01
KN3_TCPX01_SYS_NAME              $$$$
KN3_TCPX01_TCP_STC               $$$$$$$$
KN3_TCPX01_TCPIP_PROFILES_DSN    TCPIP.PROFILE.TCPIP
KN3_TCPX01_OVRD_GLOBAL_FLAG      N
KN3_TCPX01_OVRD_COLLECT_STACK    Y
KN3_TCPX01_OVRD_CONN             Y
KN3_TCPX01_OVRD_IPSEC            N
KN3_TCPX01_OVRD_ROUTE_TBL        Y
KN3_TCPX01_OVRD_ROUTE_TBL_FREQ   10
KN3_TCPX01_OVRD_FTP              Y
KN3_TCPX01_OVRD_FTP_DSPINTV      2
KN3_TCPX01_OVRD_TN3270           Y
KN3_TCPX01_OVRD_TN3270_DSPINTV   2
KN3_TCPX01_OVRD_OSA              Y
KN3_TCPX01_OVRD_INTS             Y
KN3_TCPX01_OVRD_INTE             Y
KN3_TCPX01_OVRD_GLBS             Y
KN3_TCPX01_TCPIP_PROFILES_MBR    ""
**** KN3_TCPX02_* is a sample row for a second instance.
**KN3_TCPX02_ROW                   02
**KN3_TCPX02_SYS_NAME              0128
**KN3_TCPX02_TCP_STC               &CNMTCPN. 
**KN3_TCPX02_TCPIP_PROFILES_DSN    SYS1.TCPIP.PROFILES 
**KN3_TCPX02_OVRD_GLOBAL_FLAG      N 
**KN3_TCPX02_OVRD_COLLECT_STACK    Y 
**KN3_TCPX02_OVRD_CONN             Y 
**KN3_TCPX02_OVRD_IPSEC            N 
**KN3_TCPX02_OVRD_ROUTE_TBL        Y 
**KN3_TCPX02_OVRD_ROUTE_TBL_FREQ   10 
**KN3_TCPX02_OVRD_FTP              Y 
**KN3_TCPX02_OVRD_FTP_DSPINTV      2 
**KN3_TCPX02_OVRD_TN3270           Y 
**KN3_TCPX02_OVRD_TN3270_DSPINTV   2 
**KN3_TCPX02_OVRD_OSA              Y 
**KN3_TCPX02_OVRD_INTS             Y 
**KN3_TCPX02_OVRD_INTE             Y 
**KN3_TCPX02_OVRD_GLBS             Y 
**KN3_TCPX02_TCPIP_PROFILES_MBR    "NMP128" 
KN3_TCPX                         END
    The name of the system to be monitored is represented as four dollar signs ($$$$). Whenever there is not a specific entry in the table for a detected TCP/IP address space, the configuration options provided in the $$$$ entry are used. The configuration options provided in the $$$$ entry are used whenever a specific entry in not included in the table for a detected TCP/IP address space. The $$$$ entry is referred to as the global default.
  3. Examine the values for KN3_AGT_STC for the global values and KN3_TCPX01_SYS_NAME and KN3_TCPX01_TCP_STC for the override values. If these are not the systems you wanted to monitor, change these values.
  4. If you change any values, you must update the configuration values in the RTE. See PARMGEN Step 4 underCannot connect to location server or find running CMS on CT_CMSLIST.

Beginning in V4.2 or later, you can eliminate a TCP/IP stack from the set of stacks to be monitored by responding N to the question Do you wish to monitor this stack on the ADD/UPDATE TCP/IP MONITORED SYSTEMS INFO panels or specifying N for the values of the KN3_TCP_COLLECT_STACK PARMGEN parameter.

The IBM® Z OMEGAMON® AI for Networks monitoring agent discovers all active stacks when it initializes. By default, the monitoring agent attempts to monitor all stacks that have been discovered. Every minute, the monitoring agent checks to see whether the stacks that are present in the previous monitoring interval are still active. If a stack has gone from active to inactive, the monitoring agent restarts the registration process for that stack. If the stack cannot be monitored, it is deregistered and checked again after a minute. This results in messages KN3I002I through KN3IR006 being written to the RKLVLOG repeatedly: 
KN3I002I TCP SUBNODE MONITOR STARTING FOR ORIGINNODE TCPIP5:SYS1
KN3I003I TCP SUBNODE MONITOR STARTED FOR ORIGINNODE TCPIP5:SYS1
KN3IR004 TCP SUBNODE REGISTRATION SUCCESSFUL FOR ORIGINNODE TCPIP5:SYS1
KN3IR005 TCP SUBNODE DEREGISTERED FOR ORIGINNODE TCPIP5:SYS1
KN3IR006 TCP SUBNODE MONITOR STOPPED FOR ORIGINNODE TCPIP5:SYS

The product is working as designed and the messages can be ignored. To avoid repeated attempts to register a stack that is not active most of the time, you can indicate that it not be monitored using the Configuration Tool or PARMGEN.