Status information for error analyses

Health reporting

When IBM Integrated Synchronization is running, it regularly provides performance and health information in replication event messages. This information is refreshed every 15 minutes. The health information consists of DSNX messages with a severity of INFO. (In product versions earlier than 7.5.9, the interval was just 5 minutes, and the severity was WARNING). The messages are written to the SYSLOG of the accelerator database. You can also view these messages from the replication event viewer of your administration client. In addition, you can collect and display the information on the IBM Db2 Analytics Accelerator Console (see Retrieving status information from the console). The message content varies with respect to the status. Hence two types of messages exist: messages that are issued in regular operation mode, and messages that are issued after a restart of IBM Integrated Synchronization.

Messages in regular operation mode

These messages contain the following information:

Up to product version 7.5.6	Product version 7.5.7 and after
Replication latency in seconds	See left column
Relative byte address (RBA) or log record sequence number (LRSN) of the latest commit	See left column
Number of open transactions	See left column
If open transactions exist, the earliest open RBA or LRSN	See left column
Log parser statistics: number of INSERT, UPDATE, and DELETE operations	Parsed source operations: number of INSERT, UPDATE, and DELETE operations
Database operations: number of INSERT and DELETE operations	Applied target operations: number of INSERT and DELETE operations
If the Java Virtual Machine (JVM) is operating in a low heap condition and the risk of an out-of-memory-error is increased, the tenured heap usage is displayed in percent.	See left column

Outline:

Up to product version 7.5.6	Product version 7.5.7 and after
Integrated Synchronization status: Latency X seconds. Latest commit RBA/LRSN X. Number of open transactions X. Earliest open RBA/LRSN X. Log parser statistics: X insert, X update, X delete. Database operations: X insert, X delete. Tenured heap usage X%.	Integrated Synchronization status: Latency X seconds. Latest commit RBA/LRSN X. Number of open transactions X. Earliest open RBA/LRSN X. Parsed source operations: X insert, X update, X delete. Applied target operations: X insert, X delete. Tenured heap usage X%.

Messages issued after a restart

These messages contain the following information:

• Replication latency in seconds
• Relative byte address (RBA) or log record sequence number (LRSN) of the latest commit
• Number of open transactions
• If open transactions exist, the earliest open RBA or LRSN
• If the Java Virtual Machine (JVM) is operating in a low heap condition and the risk of an out-of-memory-error is increased, the tenured heap usage is displayed in percent.

Outline:

Integrated Synchronization status after restart 
at <last applied transaction source timestamp>.
Latency X seconds.
Latest commit RBA/LRSN X.
Number of open transactions X.
Earliest open RBA/LRSN X.
Tenured heap usage X%.

Service Request Blocks

When IBM Integrated Synchronization has been configured for a Db2 subsystem and an accelerator, or for a data sharing member and an accelerator, an additional service request block (SRB) is started in the DBM1 address space. This SRB collects log records for the tables that need to be synchronized.

The SRB remains active as long as the accelerator is connected to the Db2 subsystem or data sharing member. If the SRB is no longer kept active by incoming requests, it is removed after a timeout period of 60 seconds.

Each time a new SRB is started, the message DSNI090I is issued. For example:

08.46.07 STC18562  DSNI090I  -DB11 DSNILGRT -STARTING ASYNCHRONOUS LOG  753
  753             READER TASK FOR
  753                     SESSION ID  43299E29761A0805
  753                     STARTING AT 00000000026A1D883F4A
  753                     WITH        1 QUALIFIERS

Each time an SRB is terminated, either on request by the accelerator or through a timeout, the message DSNI092I is issued. For example:

08.46.48 STC18562  DSNI092I  -DB11 DSNILGRT -NORMAL TERMINATION OF  772
  772             ASYNCHRONOUS LOG READER TASK FOR
  772                     SESSION ID    43299E29761A0805
  772                     ENDING AT     00000000026A1D89405E
  772                     AFTER READING 92 LOG RECORDS

Errors during log reading

If an error occurs during log reading, the SRB is terminated, and the message DSNI091I is issued. The z/OS system log probably contains more information about the error.

Displaying the status of log reader tasks

-DISPLAY STATS(LOGREADERTASKS)

or, short:

-DISPLAY STATS(LRT)

These commands display all SRBs that are collecting log records in the currently active Db2 subsystem or data sharing member. For example:

DSNT788I  -DB11                                                       
 SESSIONID        STATUS     CURR. POSITION       NUM RECS   AGE       
 ---------------- ---------- -------------------- ---------- ------    
 164FA983947B0801 SUSP EOS   0000000000018898F744 4          167s      
 ******* DISPLAY OF STATS TERMINATED ********************************* 
 DSN9022I  -DB11 DSNTDSTS 'DISPLAY STATS' NORMAL COMPLETION  

The following STATUS values might be shown for log reader tasks:

RUNNING

Task is processing log records.

READING

Task is reading log records.

SUSP READ

Task suspended due to a full buffer. The task is waiting for the accelerator to fetch more records, so that processing can continue.

SUSP EOS

Task suspended because the end of the log has been reached. The task will be woken up at regular intervals to check if new log records have arrived.

CANCEL

Task is in the process of being terminated.