logging-log4j-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ralph Goers <rgo...@apache.org>
Subject Re: Berkeley DB for Java
Date Sun, 03 Mar 2013 01:44:52 GMT
JMS guaranteed delivery is achieved by sending the message to the JMS server and then writing
it to disk similar to what Flume does. Doing it locally before sending is an order of magnitude
faster. At least that is the difference between the remote and embedded Flume appenders.

Ralph

On Mar 2, 2013, at 4:00 PM, Gary Gregory <garydgregory@gmail.com> wrote:

> Sounds like a job for a JMS provider like ActiveMQ. 
> 
> Gary
> 
> On Mar 2, 2013, at 18:50, Ralph Goers <ralph.goers@dslextreme.com> wrote:
> 
>> What I am looking for is simply a Queue interface that persists log events to disk
such that control can be returned to the caller immediately after writing to the queue and
even if the JVM crashes it is guaranteed that the log event will be delivered even after a
restart.  I've found some code that lets me do that with Berkeley DB for Java. I know that
that code base is quite a bit newer than Berkeley DB itself and I haven't ever heard discussions
from anyone using it to know what kinds of problems there may be. All the uses for HSQL I've
ever seen basically say not to use it for production stuff, although I don't know why. Plus,
I really don't need a full RDBMS to implement a queue.
>> 
>> Ralph
>> 
>> On Mar 2, 2013, at 12:00 PM, Gary Gregory wrote:
>> 
>>> If you are looking for a Java embeddable database, I've used H2 for unit testing
a JDBC piece instead of using heavier databases like MySQL, Oracle, MSSQL and so on.
>>> 
>>> G
>>> 
>>> 
>>> On Sat, Mar 2, 2013 at 12:15 PM, Ralph Goers <ralph.goers@dslextreme.com>
wrote:
>>>> Does anyone have experience using Berkeley DB for Java?  I am looking at
creating something similar to the embedded Flume appender using Berkeley DB as the "file channel"
so that I don't have to bring in all the Flume dependencies and the socket connection to Flume
can be done asynchronously.  In effect, I would be using Berkeley DB as the queue between
the client and the asynchronous portion of the Appender.
>>>> 
>>>> I'd like to know if I am opening myself up for problems in doing this so
if anyone has experience with using it I'd like to know about them.
>>>> 
>>>> Ralph
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: log4j-dev-unsubscribe@logging.apache.org
>>>> For additional commands, e-mail: log4j-dev-help@logging.apache.org
>>> 
>>> 
>>> 
>>> -- 
>>> E-Mail: garydgregory@gmail.com | ggregory@apache.org 
>>> JUnit in Action, 2nd Ed: http://bit.ly/ECvg0
>>> Spring Batch in Action: http://bit.ly/bqpbCK
>>> Blog: http://garygregory.wordpress.com 
>>> Home: http://garygregory.com/
>>> Tweet! http://twitter.com/GaryGregory

Mime
View raw message