April 2010 - logback-dev - mailman3.qos.ch

[JIRA] Created: (LBCLASSIC-183) Tomcat reports SEVERE ThreadLocal issues upon shutdown
by Anthony Whitford (JIRA) 01 Feb '11

01 Feb '11

Tomcat reports SEVERE ThreadLocal issues upon shutdown ------------------------------------------------------ Key: LBCLASSIC-183 URL: http://jira.qos.ch/browse/LBCLASSIC-183 Project: logback-classic Issue Type: Bug Affects Versions: 0.9.18 Environment: Tomcat 6.0.24, JDK 6, Windows or Solaris Reporter: Anthony Whitford Assignee: Logback dev list I have a webapp that uses Logback, deployed on a Tomcat instance. Upon shutdown, Tomcat reports this and doesn't completely shut down: SEVERE: A web application created a ThreadLocal with key of type [org.slf4j.impl.CopyOnInheritThreadLocal] (value [org.slf4j.impl.CopyOnInheritThreadLocal@1bb35 b]) and a value of type [null] (value [null]) but failed to remove it when the web application was stopped. To prevent a memory leak, the ThreadLocal has been forcibly removed. Feb 3, 2010 10:49:22 PM org.apache.catalina.loader.WebappClassLoader clearThreadLocalMap SEVERE: A web application created a ThreadLocal with key of type [null] (value [ch.qos.logback.core.UnsynchronizedAppenderBase$1@10e2558]) and a value of type [java.lang.Boolean] (value [false]) but failed to remove it when the web application was stopped. To prevent a memory leak, the ThreadLocal has been forcibly removed. Is LogBack leaking resources? -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://jira.qos.ch/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

7 12

[JIRA] Created: (LBSITE-38) Reference the configuration chapter from onJoran and vice versa
by Ceki Gulcu (JIRA) 11 Jan '11

11 Jan '11

Reference the configuration chapter from onJoran and vice versa --------------------------------------------------------------- Key: LBSITE-38 URL: http://jira.qos.ch/browse/LBSITE-38 Project: logback-site Issue Type: Task Components: Documentation Reporter: Ceki Gulcu Assignee: Logback dev list The onJoran chapter provides low level details about joran and the configuration chapter shows usage of joran. each chapter should reference the other for completeness. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://jira.qos.ch/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

2 1

[JIRA] Created: (LBCLASSIC-113) Nullpointer
by Mathias Bogaert (JIRA) 23 Dec '10

23 Dec '10

Nullpointer ----------- Key: LBCLASSIC-113 URL: http://jira.qos.ch/browse/LBCLASSIC-113 Project: logback-classic Issue Type: Bug Components: Other Reporter: Mathias Bogaert Assignee: Logback dev list Priority: Critical java.lang.NullPointerException at ch.qos.logback.classic.Logger.callTurboFilters(Logger.java:769) [logback-classic-0.9.15.jar:na] at ch.qos.logback.classic.Logger.isTraceEnabled(Logger.java:641) [logback-classic-0.9.15.jar:na] at ch.qos.logback.classic.Logger.isTraceEnabled(Logger.java:637) [logback-classic-0.9.15.jar:na] at org.hibernate.event.def.DefaultLoadEventListener.proxyOrLoad(DefaultLoadEventListener.java:198) [hibernate-core-3.3.0.SP1.jar:3.3.0.SP1] at org.hibernate.event.def.DefaultLoadEventListener.onLoad(DefaultLoadEventListener.java:126) [hibernate-core-3.3.0.SP1.jar:3.3.0.SP1] at org.hibernate.impl.SessionImpl.fireLoad(SessionImpl.java:905) [hibernate-core-3.3.0.SP1.jar:3.3.0.SP1] Hibernate calls if ( log.isTraceEnabled() ). -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://jira.qos.ch/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

3 3

[JIRA] Created: (LBACCESS-10) Servlet input stream gets corrupted
by Augusto (JIRA) 16 Nov '10

16 Nov '10

Servlet input stream gets corrupted ----------------------------------- Key: LBACCESS-10 URL: http://jira.qos.ch/browse/LBACCESS-10 Project: logback-access Issue Type: Bug Components: Other Affects Versions: 0.9.17 Environment: Windows 7 - java 1.6.0_16-b01 - Hightide 6.1H10 Windows 7 - java 1.6.0_16-b01 - Tomcat 6.0.16 Solaris 5.10 - java 1.6.0_16 - Hightide 6.1H10 Reporter: Augusto Assignee: Logback dev list The request.inputStream() is wrapped in a TeeServletInputStreamm but this class fails to pass all the data in the original input stream to the servlets. In my tests is passes properly between 4 and 8kb of data and later it just return zeros instead of the characters. I'm attaching a very simple webapp that writes the input stream of a POST request into a file to demonstrate the bug. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://jira.qos.ch/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

2 4

[JIRA] Created: (LBCLASSIC-130) Multiple resource warning is not appropriate for logback-test.xml
by Jon Vincent (JIRA) 20 Oct '10

20 Oct '10

Multiple resource warning is not appropriate for logback-test.xml ----------------------------------------------------------------- Key: LBCLASSIC-130 URL: http://jira.qos.ch/browse/LBCLASSIC-130 Project: logback-classic Issue Type: Improvement Affects Versions: 0.9.14, 0.9.15, 0.9.16 Reporter: Jon Vincent Assignee: Logback dev list Priority: Minor When LBCLASSIC-85 was fixed, it also applied to logback-test.xml. Unfortunately, some IDEs (such as Eclipse) include the test classpath for all modules in a multi-module project -- which means that it's nearly impossible to avoid having multiple logback-test.xml files on the classpath if one project depends on another when running tests in Eclipse. Every test in the project will therefore emit this warning when run from the IDE. Is it possible to have the ContextInitializer only emit this warning for logback.xml, and not logback-test.xml? This would avoid a spurious/unactionable warning for developers using Eclipse. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://jira.qos.ch/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

5 14

[JIRA] Created: (LBCORE-114) Lifecycle.stop is not called for policy objects
by Anders Wallgren (JIRA) 14 Oct '10

14 Oct '10

Lifecycle.stop is not called for policy objects ----------------------------------------------- Key: LBCORE-114 URL: http://jira.qos.ch/browse/LBCORE-114 Project: logback-core Issue Type: Bug Affects Versions: 0.9.17 Reporter: Anders Wallgren Assignee: Logback dev list TriggeringPolicy implements Lifecycle, and start is called during configuration. However, when the context is shut down, stop is not called. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://jira.qos.ch/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

2 1

[JIRA] Created: (LBCORE-117) Supply an appender that collects N events on all levels and dumps them to the output when an ERROR is logged
by Aaron Digulla (JIRA) 04 Oct '10

04 Oct '10

Supply an appender that collects N events on all levels and dumps them to the output when an ERROR is logged ------------------------------------------------------------------------------------------------------------ Key: LBCORE-117 URL: http://jira.qos.ch/browse/LBCORE-117 Project: logback-core Issue Type: New Feature Components: Appender Reporter: Aaron Digulla Assignee: Logback dev list For many applications, it's not necessary to log at the levels DEBUG and TRACE all the time. In fact, we usually only do this to be able to debug ERROR events. It would be great if there was an appender that would only collect a configurable number of events internally until a message of level ERROR or above is logged. At this time, the appender should dump it's internal state to the log and reset the internal state. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://jira.qos.ch/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

2 1

[JIRA] Created: (LBCORE-131) SizeAndTimeBasedFNATP will overwrite previous logging files when File property is set at RollingFileAppender
by tomliliu (JIRA) 30 Jun '10

30 Jun '10

SizeAndTimeBasedFNATP will overwrite previous logging files when File property is set at RollingFileAppender ------------------------------------------------------------------------------------------------------------ Key: LBCORE-131 URL: http://jira.qos.ch/browse/LBCORE-131 Project: logback-core Issue Type: Bug Components: Rolling Affects Versions: 0.9.18 Reporter: tomliliu Assignee: Logback dev list Here's the failure case: Configuration: <appender name="appender" class="ch.qos.logback.core.rolling.RollingFileAppender"> <File>c:/var/tmp/base.log</File> <rollingPolicy class="ch.qos.logback.core.rolling.TimeBasedRollingPolicy"> <FileNamePattern> c:/var/tmp/%d{yyyy-MM-dd_HH}.%i.log </FileNamePattern> <TimeBasedFileNamingAndTriggeringPolicy class="ch.qos.logback.core.rolling.SizeAndTimeBasedFNATP"> <MaxFileSize>1KB</MaxFileSize> </TimeBasedFileNamingAndTriggeringPolicy> </rollingPolicy> <layout class="ch.qos.logback.classic.PatternLayout"> <Pattern> %d{HH:mm:ss.SSS} [%thread] %-5level %logger{36} - %msg%n </Pattern> </layout> </appender> Steps to reproduce: 1. Start the application 2. base.log was created. 3. yyyy-MM-dd_HH.0.log was created on rollover 4. rebounce the application 5. log to base.log 6. yyyy-MM-dd_HH.0.log was created on rollover, overwrite the previous yyyy-MM-dd_HH.0.log. Logging messages are lost. It looks like the root cause is that SizeAndtimeBasedFNATP does not update the currentPeriodsCounter when parentsRawFileProperty is set. if (tbrp.getParentsRawFileProperty() == null) { String regex = tbrp.fileNamePattern.toRegex(dateInCurrentPeriod); String stemRegex = FileFilterUtil.afterLastSlash(regex); computeCurrentPeriodsHighestCounterValue(stemRegex); } SizeAndTimeBasedFNATP should update currentPeriodsCounter regardless whether parentsRawFileProperty is set or not. Here's a fix works for me. SizeAndTimeBasedFNATP.java @Override public void start() { // we depend on certain fields having been initialized // in super.start() super.start(); archiveRemover = new SizeAndTimeBasedArchiveRemover(tbrp.fileNamePattern, rc); archiveRemover.setContext(context); // we need to get the correct value of currentPeriodsCounter. // usually the value is 0, unless the appender or the application // is stopped and restarted within the same period //if (tbrp.getParentsRawFileProperty() == null) { String regex = tbrp.fileNamePattern.toRegex(dateInCurrentPeriod); String stemRegex = FileFilterUtil.afterLastSlash(regex); computeCurrentPeriodsHighestCounterValue(stemRegex); //} started = true; } void computeCurrentPeriodsHighestCounterValue(final String stemRegex) { File file = new File(getCurrentPeriodsFileNameWithoutCompressionSuffix()); File parentDir = file.getParentFile(); File[] matchingFileArray = FileFilterUtil .filesInFolderMatchingStemRegex(parentDir, stemRegex); if (matchingFileArray == null || matchingFileArray.length == 0) { return; } FileFilterUtil.reverseSortFileArrayByName(matchingFileArray); currentPeriodsCounter = FileFilterUtil.extractCounter(matchingFileArray[0], stemRegex); //If parentsRawFileProperty is set, we should increment currentPeriodsCounter by one in order to avoid overwrite the last archive file. if(tbrp.getParentsRawFileProperty() != null) { currentPeriodsCounter ++; } } Thanks, Tom -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://jira.qos.ch/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

2 1

[JIRA] Created: (LBCLASSIC-167) RollingFileAppender does not auto create subdirectories in some circumstances
by Rick Janda (JIRA) 08 May '10

08 May '10

RollingFileAppender does not auto create subdirectories in some circumstances ----------------------------------------------------------------------------- Key: LBCLASSIC-167 URL: http://jira.qos.ch/browse/LBCLASSIC-167 Project: logback-classic Issue Type: Bug Components: appender Affects Versions: 0.917 Reporter: Rick Janda Assignee: Logback dev list If the FileNamePattern of the rollingPolicy contains another directory than the File element of the appender, than the directory in the FileNamePattern is not created automatically. The OnConsoleErrorListener writes out the following error in this case: -ERROR in c.q.l.core.rolling.helper.Compress - Error occurred while compressing [/opt/baa-logserver/logs/logback.log33794048136194907.tmp] into [/opt/baa-logserver/logs/archive/logback_2009-10-26.log.zip]. java.io.FileNotFoundException: /opt/baa-logserver/logs/archive/logback_2009-10-26.log.zip (No such file or directory) at java.io.FileNotFoundException: /opt/baa-logserver/logs/archive/logback_2009-10-26.log.zip (No such file or directory) The appender is configured as followed (files are rolled into a archive directory on daily basis): <appender name="LOGBACK" class="ch.qos.logback.core.rolling.RollingFileAppender"> <File>${LOG_DIR}/logback.log</File> <rollingPolicy class="ch.qos.logback.core.rolling.TimeBasedRollingPolicy"> <FileNamePattern>${LOG_DIR}/archive/logback_%d{yyyy-MM-dd}.log.zip</FileNamePattern> <MaxHistory>${MAX_HISTORY_DAYS}</MaxHistory> </rollingPolicy> <layout class="ch.qos.logback.classic.PatternLayout"> <Pattern>%date{dd-MM-yyyy HH:mm:ss.SSS} [%contextName] $%mdc{refId}$ %-5level %logger{0} - %msg%n%xEx{full}</Pattern> </layout> </appender> -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://jira.qos.ch/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

3 3

[JIRA] Created: (LBCORE-151) RenameUtil doesn't create create a non-existing destination directory
by Torsten Juergeleit (JIRA) 08 May '10

08 May '10

RenameUtil doesn't create create a non-existing destination directory --------------------------------------------------------------------- Key: LBCORE-151 URL: http://jira.qos.ch/browse/LBCORE-151 Project: logback-core Issue Type: Bug Components: Rolling Affects Versions: 0.9.20 Reporter: Torsten Juergeleit Assignee: Logback dev list RenameUtil doesn't handle moving files into non-existing directories in a robust manner. There's no check if the destination directory is available and writable. If File.renameTo() fails then it's assumed that this is due to a locked from file. The other reason of a non-existing or non-writable destination directory is not checked (corresponding unit test is attached). Here it makes sense to create a non-existing directory tree via File.mkdirs(). Otherwise the requirement of an existing destination directory should be clearly stated in the documentation of RollingFileAppender. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://jira.qos.ch/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

2 5