Administrator does not start after a DB2 database failover.
Article ID: KB0084711
Updated On:
| Products | Versions |
|---|---|
| TIBCO Administrator | - |
| Not Applicable | - |
Description
Resolution:
Description:
=========
Administrator fails over to a secondary DB2 database. Standard tasks run fine until the Administrator is restarted. If you fail back to the primary database the Administrator works normally.
Environment:
===========
Operating System(s): All
TIBCO Administrator: All
Database Server: DB2
Database Replication: High availability disaster recovery (HADR)
Symptoms:
========
1). A similiar error to the following would be found in the Administrator.log
Error [EntityInputStream] AESDKJ-0000 [main] java.io.StreamCorruptedException: invalid stream header: 00000000
at java.io.ObjectInputStream.readStreamHeader(ObjectInputStream.java:784)
at java.io.ObjectInputStream.<init>(ObjectInputStream.java:281)
at com.tibco.pof.entitystore.db.DBEntityInputStream.getObject(DBEntityInputStream.java:584)
2). All the BLOB type columns in the TIBCO database tables (default prefixed with AD or AU) will be set to "NOT LOGGED".
Cause:
======
DB2 HADR uses logging when replicating data to the secondary database. Since BLOB data is not being replicated Administrator is unable to locate the necessary data during nitialization. More inforamtion about DB2 HADR can be found at IBM's web site: http://www.redbooks.ibm.com/redbooks/SG247363/wwhelp/wwhimpl/js/html/wwhelp.htm.
Workaround:
==========
A possible workaround would be to manually change the attribute to "LOGGED" for each of the BLOB columns. You should bear in mind that BLOB data larger than 1GB may impact the performance of data replication.
Description:
=========
Administrator fails over to a secondary DB2 database. Standard tasks run fine until the Administrator is restarted. If you fail back to the primary database the Administrator works normally.
Environment:
===========
Operating System(s): All
TIBCO Administrator: All
Database Server: DB2
Database Replication: High availability disaster recovery (HADR)
Symptoms:
========
1). A similiar error to the following would be found in the Administrator.log
Error [EntityInputStream] AESDKJ-0000 [main] java.io.StreamCorruptedException: invalid stream header: 00000000
at java.io.ObjectInputStream.readStreamHeader(ObjectInputStream.java:784)
at java.io.ObjectInputStream.<init>(ObjectInputStream.java:281)
at com.tibco.pof.entitystore.db.DBEntityInputStream.getObject(DBEntityInputStream.java:584)
2). All the BLOB type columns in the TIBCO database tables (default prefixed with AD or AU) will be set to "NOT LOGGED".
Cause:
======
DB2 HADR uses logging when replicating data to the secondary database. Since BLOB data is not being replicated Administrator is unable to locate the necessary data during nitialization. More inforamtion about DB2 HADR can be found at IBM's web site: http://www.redbooks.ibm.com/redbooks/SG247363/wwhelp/wwhimpl/js/html/wwhelp.htm.
Workaround:
==========
A possible workaround would be to manually change the attribute to "LOGGED" for each of the BLOB columns. You should bear in mind that BLOB data larger than 1GB may impact the performance of data replication.
Issue/Introduction
Administrator does not start after a DB2 database failover.
Feedback
Yes
No
Powered by
