Job Management Partner 1/Base User's Guide
An event database consists of files that accumulate JP1 events occurring on hosts running JP1/Base. An event database consists of the following files:
- Data (IMEvent0.dat and IMEvent1.dat)
- Indexes (IMEvent0.idx and IMEvent1.idx)
- Transfer information (IMEvent0.fwd and IMEvent1.fwd)
- Duplication prevention table (IMEvent.rep)
The files above are generated automatically when an event service starts. Two data files, two index files, and two transfer information files are generated. When the first file reaches the size specified by the db-size parameter in the event server settings file (conf), the second file is swapped in. When the second file reaches the maximum size, the contents of the first file are cleared and new JP1 events are accumulated in the first file.
The following figure shows how the event databases are swapped over.
Figure 1-7 Swapping of event databases
The event database is swapped over when the data in the current database reaches the size specified in the event server settings file, or when the time limit for keeping JP1 events specified in the event server settings file has expired. You can also use a command to manually swap the databases.
You can view the contents of the event database by browsing them in the Event Console window or by outputting them to a CSV file. For information on JP1/IM - View, see the manual Job Management Partner 1/Integrated Management - Manager Administration Guide. For details on how to output the contents of the event database to a CSV file, see 8.3 Outputting the event database to a CSV file.
- Organization of this subsection
- (1) When event databases are checked for possible corruption
- (2) Checking duplicate registrations
- (3) Duplication prevention table
(1) When event databases are checked for possible corruption
Note that the event database might become corrupted if you edit it directly, or if you use an OS command or backup software to back up or restore the event database while the event service is active.
JP1/Base checks whether the event database is corrupted at the following times:
- When the event service starts up
- When JP1 events are transferred
- When JP1 events are acquired by the event acquisition function#
- When an event search is performed from JP1/IM - View#
- #
- A message reporting that the database is corrupted is output only once for both the active and standby event databases. If an active event database is corrupted, this message appears only once when you attempt to acquire or retrieve a JP1 event from that database. The same is also true for a standby event database corruption.
To check messages in JP1/IM - View, convert the messages to JP1 events and send them to the manager host. For details on event conversion, see 9. Setting Up the Event Converters.
(2) Checking duplicate registrations
The event service enables you to check whether any duplicate JP1 events exist in the event database when registering a JP1 event. Unless duplicate registration is checked, a duplicate JP1 event might be registered when the following situation occurs:
- When a communication error occurs between hosts sending and receiving the forwarded JP1 events
- The JP1 events being forwarded to multiple hosts are aggregated to a re-forwarding host
- The transfer route of JP1 events circulates.
(3) Duplication prevention table
The duplication prevention table enables you to check whether any duplicate JP1 event exists. The transfer records of JP1 events for each sending host are written in the duplication prevention table. When the event server receives a JP1 event, the transfer record of the JP1 event will be added to the duplication prevention table or updated the appropriate record.
If the save-rep flag has been specified in the options parameter in the event server settings file (conf), the duplication prevention table is kept in a file. If the save-rep flag has not been specified, the duplication prevention table is kept in memory. The behavior of the JP1 event server that receives JP1 events depends on whether the JP1 event is kept in a file or in memory. We recommend that the duplication prevention table be kept in a file.
The differences in the behavior of the JP1 event server are as follows:
- When kept in a file
- The duplication prevention table is kept in a non-volatile state. Therefore, the data in the duplication prevention table is not erased even if the event server is restarted. When the event server receives a JP1 event from a host that has never been recorded in the duplication prevention table, the JP1 event is considered unknown and recorded in the duplication prevention table. The time necessary for recording a forwarded JP1 event is always the same, regardless of whether the JP1 event has already been forwarded or not.
- When kept in memory
- The duplication prevention table is kept in a volatile state. Therefore, the data in the duplication prevention table is erased when the event server is restarted. When the event server receives a JP1 event from a host that has never been recorded in the duplication prevention table, the JP1 event is searched for in the event database, and then recorded in the duplication prevention table. The time necessary for recording a forwarded JP1 event depends on the fact that the JP1 event has already been forwarded or not.
- Also, if the event server receives a JP1 event forwarded from a new agent, the JP1 event is searched for among all of the JP1 events registered in the event database. This means that delays might occur when operating on JP1 events. The delay becomes greater in proportion to the size of the event database.
All Rights Reserved. Copyright (C) 2009, Hitachi, Ltd.