so did
I, and it was the reason I filed the bug #161.....
basically, orion never has 2 instances of the same
entity in memory, so whenever a transaction takes too long, or tries to find an
entity held by another transaction,
orion
persistence manager detects it and makes one of these transactions fail to avoid
an actual deadlock. However, during high and not so high stress, this kind
of
exception will happen along with possible server
crashes...
please
refer to bug #161 for more details
JP
-----Original Message-----
From: David Ekholm [mailto:[EMAIL PROTECTED]]
Sent: Viernes, 08 de Diciembre de 2000 11:59
To: Orion-Interest
Subject: Deadlock exception when accessing entity beanDuring heavy stress testing of my application I noticed that I get the following exception:com.evermind.server.DeadlockException: Transactions was rolled back: timed out, while waiting for <primary key for one of my CMP entity beans here>I can't see why this happens. The database has not reported any deadlock problem, and there is a row in the database for the primary key I am trying to look up.What can trigger such a problem. The CMP bean has no relations to other beans/tables what so ever, and just a primary-key index on the primary key./David
"The Las Vegas of Online Gaming"David Ekholm
System Architect
Hammarby Kajväg 14, 120 30 Stockholm
tel: +46 (0)8 55 69 67 11
mob: +46 (0)70 486 77 38
fax: +46 (0)8 55 69 67 07
icq: 410993
