Details
-
Type:
Bug
-
Status: Closed (View Workflow)
-
Priority:
Major
-
Resolution: Fixed
-
Affects Version/s: 1.8.9, 1.9.5
-
Component/s: Build / Packaging
-
Security Level: Default (Default Security Scheme)
-
Labels:None
-
Environment:Any OpenNMS server configured with its datasources on a PostgreSQL 9.0 or later cluster
Description
Steps to reproduce:
1. Start with a DB dump from a PostgreSQL 8.4 or earlier cluster that's been backing up an OpenNMS instance that has been running long enough for some batch reports to have run
2. Load that dump into a PostgreSQL 9.0 cluster
3. Run the OpenNMS installer against the DB in the 9.0 cluster
4. Start OpenNMS, log in to the web UI, works fine
5. Restart OpenNMS and hit the web UI
Expected result: web UI works fine
Actual result: 503 error from JettyServer
Check logs/webapp/jetty.log and logs/daemon/spring.log and find many exceptions (I'll provide full traces shortly) that mention:
java.io.StreamCorruptedException: invalid stream header: BB656430
This problem has a mention on the opennms-discuss mailing list in late November 2010:
http://comments.gmane.org/gmane.network.opennms.general/38880
There's an issue in the Alfresco tracker that references the same message:
The resolution in that case was to switch to a newer PostgreSQL JDBC JAR. I tried this on my system and it appears to have eliminated the problem.