Don't know by more info might help. A catch 22 is that usually
the top of the derby.log usually has all the environment info
that could be useful.
Can you list OS, JVM. I/O handling to the log is likely an OS
thing. What properties do you have set, any special for error
logging?
I would likely look first for multiple starts and stops of derby
and a setting default for one of them to truncate the derby.log.
Maybe leading to multiple processes trying to write to same derby.log.
/mikem
Bergquist, Brett wrote:
Our customer called and said the server was not working. Before
restarting, I retrieved the derby.log and it seems strange because this
was right at the top of the log:
----------------------------------------------------------------
Tue Feb 28 15:33:58 EST 2012: Shutting down Derby engine
----------------------------------------------------------------
Tue Feb 28 15:37:28 EST 2012 Thread[DRDAConnThread_245,5,main] (XID =
1482003981), (SESSIONID = 133257672), (DATABASE = csemdb), (DRDAID =
????????.??-4471791624540787385{335696}), Cleanup action starting
Tue Feb 28 15:37:28 EST 2012 Thread[DRDAConnThread_245,5,main] (XID =
1482003981), (SESSIONID = 133257672), (DATABASE = csemdb), (DRDAID =
????????.??-4471791624540787385{335696}), Failed Statement is: null
java.lang.NullPointerException
Cleanup action completed
The strange part is “Shutting down Derby engine” is the first thing in
the log. The server was up and running since 2/16/2012 with no
problems with about 60M transactions processed, a backup performed each
night, etc. And then this.
So any thoughts on how the derby.log could be recreated and have this as
the first thing in the log? I am thinking maybe an OutOfMemory
condition but all traces of whatever when wrong are gone, so I am trying
to work backwards and see how the derby.log could be created with this
as the first thing in the log.
Thanks for any help!
Brett
