7.3 Cautions on integrated scope

• Do not stop the service while updating the tree. Inconsistencies may occur with the database contents.

  If consistencies occur with the database, messages such as "KAVB7247-E JP1/IM-CS could not execute the operation request (<request name>) from JP1/IM-View. (Cause: The record in the database is invalid)", and "KAVB7248-E JP1/IM-CS could not execute the operation request (<request name>) from JP1/IM-View. (Cause: The database cannot be operated, <code>)" are displayed.

  If this occurs, restore the backup of the database, or recreate the database by jcsdbsetup command. For details about the procedure for recovering the monitored object database, see 1.2.2(2) Monitored object database backup and recovery procedures, and 1.2.2(3) Host information database backup and recovery procedures.

• If the central scope service abnormally terminates or is forcibly terminated, the database used by the central scope service may become corrupted. For this reason, back up the database on a regular basis. For details about the procedure for backing up the monitored object database, see 1.2.2(2) Monitored object database backup and recovery procedures, and 1.2.2(3) Host information database backup and recovery procedures.

• When you reinstall JP1/IM - Manager after uninstalling it, or if you connect JP1/IM - View after executing jcsdbsetup command, the [Monitoring tree] window may display the monitoring tree that existed before JP1/IM - Manager was reinstalled or the monitoring tree that existed before the jcsdbsetup command was executed. If that happens, stop the JP1/IM - View that is running, and delete the following folder before starting JP1/IM - View.

  The folder to be deleted:

  View-path\log\output

• If you deleted the event DB for JP1/Base, follow these procedures:

  • Start the [Monitoring Tree (being edited)] window of JP1/IM - View.

  • Select [File], and then [Obtain tree from server], to obtain the monitoring tree from the server.

  If you don't perform the above procedures, the following phenomenon occurs.

  • In the [Status Change Event Search] of the [Monitoring Tree] window, a wrong event is displayed.

  • In the guide function, another guide is displayed.

• If the host name defined in the system configuration definition of JP1/IM and the host name included in the information collected by the automatic generation function are different, a monitoring object that is identical to the one included in the existing monitoring tree is output as a difference.

  For example, if the host name written in FQDN format in the system configuration definition of JP1/IM is defined in a format other than FQDN on the job execution host of JP1/AJS2, a jobnet monitoring object is output as a difference. If it is not necessary, manually delete the monitoring object that has been output as a difference.

• If you start the [Monitoring Tree] window of Jp1/IM - View soon after you started JP1/IM - Manager or while the tree is being updated, a message "KAVB7240-W JP1/IM - CS was temporarily unable to perform the operation request (monitoring tree acquisition) from JP1/IM - View (Cause: The database is being used by another user) is displayed, and the [Monitoring Tree] window is not displayed. If that happens, wait for a while and then restart the [Monitoring Tree] window.

• When you monitor an event in JP1/SES format, you are not allowed to specify a message and detailed information of a basic attribute as a status change condition for an individual condition or a common condition.

• Regarding the message, detailed information and extended attribute of JP1 events, the character codes supported in status change conditions of the monitoring node are C (English code).

• If the status of the monitoring group is other than the initial status, and basic attributes or extended attributes of JP1 event are included in the guide viewing conditions for the guide information displayed in the monitoring group, the guide information will not change unless the status of the applicable monitoring group changes to the initial status.

  For example, consider a situation where a monitoring object with an error status and a monitoring object with a warning status exist directly under a monitoring group.

  In this situation, it is assumed that the monitoring group has the error status, and a guide, whose viewing condition is a JP1 event that meets a status change condition for a monitoring object with the error status, is displayed in the monitoring group.

  In this situation, changing the status of the monitoring object from error to initial, therefore changing the status of the monitoring group to warning, will not change the guide information displayed in the monitoring group.

  Changing the status of the monitoring object from warning to initial, therefore changing the status of the monitoring group to initial, will change or hide the guide information. We recommend that you specify only the monitoring node ID for a viewing guide condition that is displayed in the monitoring group, and provide a definition that does not include basic attributes or extended attributes of a JP1 event. Providing a definition like this allows you to always display the same guide, when you display guide in a monitoring group, regardless of the status of the monitoring group. Furthermore, be sure to provide a definition for a guide in which you specified only a monitoring node ID as a guide viewing condition, before providing a definition for a guide that includes basic attributes or extended attributes of a JP1 event as its viewing condition.

• The default setting for updating data to a monitoring object DB on a server is asynchronous writing.

• The following settings in a cluster configuration enables synchronous writing when the data is updated to the server.

  (a) Stop all the JP1/IM - Managers in physical host and logical host environments.

  (b) Create a text file with the following contents. (Fill in the name of the logical host to be configured, for "Logical host name" below).

  [Logical host name\JP1SCOPE\BMS]

  "DB_ACC_MODE_SYNC"=dword:00000000

  (c) Specify the text file you created in (b) for the argument, and execute the following command on the active server and on the standby server.

  jbssetcnf <The name of the file created>

  If you configure settings to enable synchronous writing to a monitoring object DB, the performance of updating data to the server decreases. However, because this makes it less likely that a monitoring object DB will be damaged at the time of logical host failover, the above configuration is recommended for a cluster environment.

• To change the configuration from synchronous writing back to asynchronous writing, use the following procedures:

  (a) Stop all the JP1/IM - Manager in physical host and logical host environments.

  (b) Create a text file with the following contents. (Fill in the name of the logical host to be configured, for "Logical host name" below).

  [Logical host name\JP1SCOPE\BMS]

  "DB_ACC_MODE_SYNC"=dword:00000001

  (c) Specify the text file you created in (b) for the argument, and execute the following command on the active server and on the standby server.

  jbssetcnf <The name of the file created>

• Specify one of the following names for the host name to be specified as the status change condition for the monitoring object. If you don't specify one of the following host names, the status of the monitoring object may not change even if a JP1 event occurs.

  (a) The host name returned by hostname command.

  (b) The host name you registered with the host information DB.

For details on how to configure (b), see jcshostsimport command (in Chapter 1. Command) in the manual JP1/Integrated Management 2- Manager Command, Definition File and API Reference.

• Specify the following host name for a managed host and a manager host of JP1/IM - Manager. If you don't specify the following host name, the monitoring node of the applicable host may not be displayed when using the automatic generation function of the monitoring tree.

  (a) The host name returned by hostname command

For details on how to configure a managed host and a manager host of JP1/IM - Manager, see the description of the configuration definition file (jbs_route.conf) in the JP1/Base User's Guide.

• If you automatically generate a large number of monitoring nodes using the automatic generation function of the monitoring tree, the automatic generation of the monitoring tree may time out. If that happens, execute the setup command for the automatic generation on the managed host and disable the automatic generation to decrease the number of the monitoring nodes that are going to be generated with the automatic generation function.

  Furthermore, if there is more than one managed host of JP1/IM - Manager where JP1/Base has not been started, the automatic generation of the monitoring tree may time out. If that happens, start JP1/Base, or exclude the hosts where JP1/Base has not been started from the managed hosts of JP1/IM - Manager.

• If you automatically generate the monitoring tree using the "business-oriented tree" template, monitoring objects for JP1/PFM placed under the monitoring group named <AJS2 host name> are generated to correspond to resources or applications monitored by JP1/PFM on the <AJS2 Host name> host. Manually delete the monitoring objects for resources and applications that are not related to the jobnet represented by the monitoring group named <jobnet name> to which <AJS2 host name> belongs.

  Similarly, with regard to the monitoring objects for JP1/PFM placed under the monitoring group named <Cosminexus operation management agent host name>, manually delete the monitoring objects for resources and applications that are not related to J2EE application represented by the monitoring group for <Cosminexus business operation>.

• If you edit the filter definition file (snmpfilter.conf) for the SNMP trap conversion function of JP1/Base, in order to monitor JP1/Cm2/NNM, add only definitions regarding the SNMP trap to be monitored in your environment from the definitions in the sample file (snmpfilter_im_sample.conf) to the filter definition file. Configure the size of the definition within 900 bytes in the filter definition file, by using the following formula.

  ((a1+1)+(a2+1)+(a3+1)+(a4+1).........(an+1))+34 < 900 bytes

  • an: The OID length of the SNMP trap defined in snmpfilter.conf (If OID is [.1.2.3.4.5], an is 10 bytes.)

  • If general traps are defined in snmpfilter.conf, the filter size is the [Result of the above calculation + (Number of general traps * 2)].

  • For the filter, based on what is defined in the filter file (snmpfilter.conf), the applicable object ID (OID) is obtained from trapd.conf of NNM.

• In the list of common conditions in the Common Condition Settings window, the following items are displayed but are not used in an English environment:

  • Common conditions related to JP1/Cm2/SSO, such as a System Alert Event (SSO) or an Application Alert Event (SSO)

  • Common conditions related to SCIM, such as a System Error Event (SCIM) or a System Warning Event (SCIM)

  • Common conditions related to System Manager, such as a Physical Host Emergency Event (System Manager)

• In the list of monitoring node types in the Create New Monitoring Node window, the following items are displayed but are not used in an English environment:

  • Monitoring objects related to JP1/Cm2/SSO, such as a System Alert Event (SSO) or Category Monitoring (SSO)

  • Monitoring objects related to SCIM, such as a System Error Event (SCIM) or a System Warning Event (SCIM)

  • Monitoring objects related to System Manager, such as a Physical Host Emergency Event (System Manager)

• In the list of generation trees in the Auto-generation - Select Configuration window the following item is displayed but is not used in an English environment:

  • System Configuration Tree

• Tuning to improve performance of JP1 event reception of central scope

  When JP1/IM - Manager receives a JP1 event, the received JP1 event is compared by central scope with the state change conditions of each monitored object in the monitoring tree. If the number of monitored objects increases, the number of comparisons also increases. This deteriorates the performance of not only central scope, but also JP1/IM - Manager as a whole.

  Tuning to prevent performance deterioration, and to enhance the JP1 event reception performance of central scope, is described below.

  (a) By validating the definition file for on-memory mode of status change conditions, JP1 event reception performance can be expected to improve because the number of times the disk is accessed for central scope processing when a JP1 event is received can be reduced.

  (b) To reduce the state change conditions that are compared, central scope performs filter processing when a JP1 event is received by using the following state change conditions:

  • Common conditions

  • Host name as an individual condition (the individual condition that set host name comparison as the comparison method)

  By setting the items described above in the state change conditions of the monitored object, JP1 event reception performance can be expected to improve because of the number of state change conditions for comparison can be reduced.

• When the system monitoring object [NNMi Monitor] or [Node Monitor (NNMi)] is used, please set the extended attribute NNMI_FAMILY_UK for the JP1 event, which is issued by JP1/IM-EG and converted to an NNMi incident (This is not set by default).

  When the extended attribute NNMI_FAMILY_UK is not set, the system monitoring object [NNMi Monitor] or [Node Monitor (NNMi)] cannot be monitored. For details on how to set up the extended attribute NNMI_FAMILY_UK, see the manual JP1/Integrated Management 2 - Event Gateway for Network Node Manager i.

To Page Top