> In the past someone mentioned that the ASYNC appender can deadlock. Has
> anyone else experienced this? When it deadlocked, what happened? Did log
> messages just stop getting delivered, or did it take the whole server down?
It deadlocked already at server startup so I guess that qualifies for t
If to
achieve synchronous writing of the logs via an MDB, you will have to configure
the container to only consume one message at a time from the
queue/topic. This could become a bottlekneck and you would run the risk of
overflowing the queue and never catching up. If you bring JMS into th
We log a ton of stuff and are having problems caused by synchronous logging
blocking the appserver when linux rotates other various logs in the
system. So we're considering using the ASYNC or JMS appenders. If we use
the JMS appender we'll write a quick MDB to log all messages synchronously
When it deadlocked, what happened? Did log messages just stop getting
delivered, or did it take the whole server down?
Has one else experiences such a problem?
.peter
Juha Lindfors wrote:
I've seen the async appender deadlock, so I'm not entirely
convinced of its robustness. I did not pur
I've seen the async appender deadlock, so I'm not entirely
convinced of its robustness. I did not pursue it any further to find the
cause. There's at least one open bug in the log4j tracker for this.
YMMV
-- Juha
On Fri, 19 Dec 2003, Peter Luttrell wrote:
> We log a ton of stuff and are having
We log a ton of stuff and are having problems caused by synchronous logging
blocking the appserver when linux rotates other various logs in the
system. So we're considering using the ASYNC or JMS appenders. If we use
the JMS appender we'll write a quick MDB to log all messages synchronously.
Does a