[infinispan-dev] query repl timeout

[Marko Lukša]

Actually my tests showed that ispn's trace log lines were only written to the AS logs and did not reach capedwarf's handler. So they couldn't have been written to the ispn cache. So I guess our logging is working as it should.

Sent from my iPad

On 9. apr. 2013, at 20:12, Ales Justin <ales.justin at gmail.com> wrote:

> Yes, afaik, Dan got a ton of trace logs into CD for Ispn.
> But as you can see we explicitly exclude org.infinispan by default.
> 
> I also asked Marko to investigate this.
> 
> -Ales
> 
> On Apr 9, 2013, at 19:25, "James R. Perkins" <jperkins at redhat.com> wrote:
> 
>> I don't see anything obvious wrong with the logging deployer. Are you seeing log statements from the excluded categories appearing in the handlers defined in the WEB-INF/logging.properties file?
>> 
>> 
>> On 04/09/2013 12:51 AM, Ales Justin wrote:
>>> Hey Dan,
>>> 
>>> first, thanks for looking into this!
>>> 
>>>> I managed to start the app with 3 nodes on my laptop, and it inserted a flight in about 26.7 seconds with TRACE enabled for org.infinispan. However, when I counted the number of cache commands being executed and I got 55000 (8700 of which went remote), which seems way too much for a single insert. (The log file grew by more than 100 MB.)
>>> 
>>> Yeah, with trace you get insane number of log lines.
>>> 
>>> Although, this seems to show that again our logging doesn't really work as expected.
>>> * https://github.com/capedwarf/capedwarf-jboss-as/blob/master/extension/src/main/java/org/jboss/as/capedwarf/deployment/CapedwarfLoggingParseProcessor.java#L83
>>> As we explicitly exclude logs from the frameworks/libs we use internally in CapeDwarf.
>>> 
>>> @Luksa: we need to check this, again ... :-)
>>> 
>>> @James: do you see any obvious mistakes in our logging deployer?
>>> 
>>>> I think there may be a cycle whereas each operation on a cache generates a log message, which then triggers a change in the Lucene caches, which writes another log message, and so on.
>>> 
>>> That's why we excluded the internal frmwrks/libs in the first place.
>>> 
>>>> How does CapeDwarf capture the logs?
>>> 
>>> See the url above -- we define a new handler: org.jboss.capedwarf.shared.log.Logger
>>> 
>>>> I haven't seen any appender in the standard-capedwarf.xml configuration. How can I enable TRACE logging for org.infinispan without the logs being indexed by CapeDwarf?
>>> 
>>> As I said, it should already work that way -- but unfortunately it looks like it doesn't.
>>> 
>>> But, hopefully, this should work:
>>> * you disable org.infinispan explicitly in GAE log file
>>> --> https://developers.google.com/appengine/docs/java/runtime#Logging
>>> 
>>> Add WEB-INF/logging.properties file with JUL logging config / properties to helloorm2.war (aka our ROOT.war).
>>> 
>>> HTH
>>> Ping me on irc for any other issues.
>>> 
>>> -Ales
>> 
>> -- 
>> James R. Perkins
>> JBoss by Red Hat
> _______________________________________________
> infinispan-dev mailing list
> infinispan-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/infinispan-dev
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.jboss.org/pipermail/infinispan-dev/attachments/20130409/03971449/attachment-0001.html