Open yami12376 opened 7 years ago
I got the same issue. If i connect servicemix with TCPAppender to logstash, got the same error on version 5.4.3. But if I do the same test with an older version of logstash (5.2.2) the log are capture by logstash
Issue still occurs on Logstash 5.5
I confirm what lours1 said - on logstash 5.2.2 the issue does not appear
We've got the same issue in JBoss. It is related to Limit objects accepted to be only log4j objects.
When MDC/NDC is touched (even if it is removed before the actual logging call), it gets serialized and triggers the error:
[2017-07-26T15:32:05,690][DEBUG][logstash.inputs.log4j ] Closing connection {:client=>"x.x.x.x:62734", :exception=>java.io.InvalidObjectException: Object type java.util.Hashtable is not allowed.}
The last working version of logstash-input-log4j
with MDC/NDC is 3.0.3
.
Issue still occurs on Logstash 7.8,the following is issue detail:
Closing connection {:client=>"xxx", :exception=>java.io.InvalidObjectException: Object type org.apache.log4j.spi.LocationInfo is not allowed.}
Hey, i have two servers of applications (using the same version of application), and only on one of them i see this issue. Application is stored in Weblogic. They may differ slightly by
configuration ear
- but i did not find there anything that would have meaning in this matter.not sure what is the cause ?
ELK 5.4
- elsewhere working fine Configuration on both of them is the same.What can be wrong in one of them ? I tailed the log - and both seems to send the same log - but one is ok - the second one is not ok.
I run logstash by these commands:
centralLogstash.conf
tcp_agent_4561.conf
Java configuration for both of application is the same:
Both of them share the same application.properties file regarding these properties.
Both of them send pretty much the same message (except for time)
changed some stuff to
X
Y
etc. just in case.I'm not sure how to find reason why this is happening - any help narrowing down the issue will be great.