Updating WAS JDBC Data Source Passwords - Don't forget to reboot

In the words of The IT Crowd ( back on for a final time on September 27 ), "Have you tried turning it off and on again?".

As part of a normal password expiration process, we needed to update the JDBC Data Source passwords ( actually, the Java2 Authentication Aliases = J2C Aliases ) for a BPM Standard 7.5.1Process Server installation.

Having changed the passwords, remembering to add a comment indicating when and who, I tested the JDBC data sources, which immediately failed with: -

The test connection operation failed for data source LocalZoneDB on server nodeagent at node gbrdsr000000918_base with the following exception: java.sql.SQLException: [jcc][t4][2013][11249][3.61.65] Connection authorization failure occurred. Reason: User ID or Password invalid. ERRORCODE=-4214, SQLSTATE=28000DSRA0010E: SQL State = 28000, Error Code = -4,214. View JVM logs for further details.

The test connection operation failed for data source LocalZoneDB on server nodeagent at node gbrdsr000000917_base with the following exception: java.sql.SQLException: [jcc][t4][2013][11249][3.61.65] Connection authorization failure occurred. Reason: User ID or Password invalid. ERRORCODE=-4214, SQLSTATE=28000DSRA0010E: SQL State = 28000, Error Code = -4,214. View JVM logs for further details.

Hmmm, I thought - what have I missed ?

Initially, I restarted the Deployment Manager ( from where I was running the Test Connection ), but to no avail.

I then checked and rechecked the passwords that I'd entered via the Integrated Solutions Console GUI, by comparing them to the encoded passwords here: -

/opt/IBM/WAS/wasadmin1/app/profiles/dmgr/config/cells/ProcServ02Cell/security.xml

and confirmed that they were all exactly as I'd intended to enter them.

I then asked the DBA to validate the passwords; again, they worked perfectly ( he was testing from the DB2 server using su - for the respective user ID ).

Finally ……

I restarted the Node Agents :-)

And…. guess what ….. ?

It worked.

Why, I hear you cry ?

Because the JDBC Data Source Test Connection process, whilst initiated by the Deployment Manager, actually runs in the Node Agent JVM, hence the need for the restart :-)