]
Gregor Rosenauer commented on WFCORE-2998:
------------------------------------------
thanks for the quick response to this issue!
[~jamezp] please note that the more serious issue is that a wrong order of levels *breaks*
deployment and the entire app server just exits with an error code 1.
There is nothing in the logs, it just dies.
This should not happen.
Instead, Wildfly should report a configuration error and switch to default, logging all
levels.
levelRange in filter-spec is documented wrong and makes Wildfly
startup fail silently
-------------------------------------------------------------------------------------
Key: WFCORE-2998
URL:
https://issues.jboss.org/browse/WFCORE-2998
Project: WildFly Core
Issue Type: Bug
Components: Logging
Affects Versions: 2.2.1.Final
Reporter: Gregor Rosenauer
Assignee: James Perkins
Labels: check, configuration, logging, startup
Linux 4.8.0-54-generic #57-Ubuntu SMP Wed May 24 10:21:44 UTC 2017 x86_64 x86_64 x86_64
GNU/Linux
Oracle JDK 8 build 1.8.0_131-b11
running in a docker container using Docker 1.12.6, build 78d1802
Trying to configure e.g. a console-appender with a logging filter as described in
[Wildfly 10 Logging
Configuration|https://docs.jboss.org/author/display/WFLY10/Logging+Config...]
The server will startup and happily parse the configuration, but then exit without any
error:
{{2017-06-22 15:12:28,185 DEBG fd 7 closed, stopped monitoring <POutputDispatcher at
140706226583888 for <Subprocess at 140706228894320 with name wildfly in state
RUNNING> (stdout)>
2017-06-22 15:12:28,186 DEBG fd 9 closed, stopped monitoring <POutputDispatcher at
140706226584392 for <Subprocess at 140706228894320 with name wildfly in state
RUNNING> (stderr)>}}
The culprit is the *order* of the levels in the range expression, which is the wrong way
around.
The correct order (which makes more sense anyway) is e.g. {{[DEBUG,WARN)}} not
{{[WARN,DEBUG)}}