- logback-user - mailman3.qos.ch

Serialization improvement
by Alf Kristian Støyle 24 Apr '13

24 Apr '13

Hi everyone. First off, my first mail on the mailinglist, thanks for logback, it is awesome :) We had a problem in our project that our serialized object graph was way too large. After digging, we found that a referenced ch.qos.logback.classic.Logger was the cause. A serialized object was about 27 kB, after making the referenced logger static, the serialized size was about 4,7kB ch.qos.logback.classic.Logger has several transient fields, but parent and childrenList are not. My guess is that when serializing a logger, you then implicitly serialize every referenced logger inn the logger tree. Am I correct? Since the readResolve method is only dependent on name, every field could be made transient, problem solved? Cheers, Alf

3 5

How to programmatically find an appender for a logger?
by Konstantinidis, Johannes 23 Apr '13

23 Apr '13

Hi all, I have a custom appender and sometimes I want to change its configuration at runtime. I would expect that the appenders are inherited in the same way as the log level. So when I get a Logger for "com.mycompany.foo" and ask for the log level, and I've specified the log level only for the root logger, then I get this log level. But when I ask the same logger for a specific appender (via myLogger.getAppender("myAppender")) and the appender is referenced by the root logger, then I get nothing. It seems that I have to know which logger references the appender. To me this doesn't seem right. Where is my mistake? If I could access the parent logger, then I might climb up the hierarchy and ask the parent loggers for the appender. But I also haven't found a method to get the parent logger. Regards, Johannes ________________________________ Firma: Capgemini Deutschland GmbH Gesch?ftsf?hrer: Dr. Michael Schulte (Sprecher) * Dr. Uwe Dumslaff * Josef Ranner Aufsichtsratsvorsitzender: Antonio Schnieder Amtsgericht Berlin-Charlottenburg, HRB 98814 This message contains information that may be privileged or confidential and is the property of the Capgemini Group. It is intended only for the person to whom it is addressed. If you are not the intended recipient, you are not authorized to read, print, retain, copy, disseminate, distribute, or use this message or any part thereof. If you receive this message in error, please notify the sender immediately and delete all copies of this message.

2 2

External property file
by Suriyanto 23 Apr '13

23 Apr '13

Hello, I am following the instruction in the Logback manual to use an external property file. <property file="log.properties" /> I have log.properties file in the same directory as my executable jar file, which are created as an uber jar file from Maven plugin. However, I keep getting the error: ERROR in ch.qos.logback.core.joran.action.PropertyAction - Could not find properties file [log.properties]. Am I putting the file in the correct directory? Thanks, Suriyanto

2 11

Failing appender cause nested logging calls in application
by Emerson Farrugia 22 Apr '13

22 Apr '13

Hi, I'm writing a logback appender that pushes errors to a remote API. It uses supporting classes like RestTemplate and HttpClient to talk to the API. I initially tried to register the appender through a logback XML configuration, but I was warned that loggers named after the supporting classes would be suppressed because they were created during the logback configuration phase. Since those loggers are needed in other parts of the system, and multi-step initialisation didn't seem feasible, I had to revise the implementation. The revised implementation programmatically registers the appender after the logback configuration phase, and it seems to work correctly. The issue is that I'm getting nested logging calls. For example, RestTemplate logs to SLF4J. If there is a failure in RestTemplate, it sends a logging event to SLF4J and therefore to logback, which goes to my appender, which will once again try to hit the RestTemplate. Some mechanism seems to be protecting against an infinite loop, since only four messages are logged, but I'm wondering if there's something I can do to prevent my appender from attempting to even try the nested call. MDC magic perhaps, assuming the same thread is making the nested call? Any feedback would be appreciated. Thanks, Emerson

2 1

Is grouped (deferred) log statements from same threads possible using logback?
by janolaveide＠gmail.com 22 Apr '13

22 Apr '13

I have read about the Mapped Diagnostic Context, and although useful, the log statements from different threads still appear interspersed in the output. What I would like is to have all log statements from a thread outout consecutively when that thread is done (and I know what and when "done" is) For example : Thread A starts at time T-Astart, performs logging at various points, and is done at T-Adone. Before A is done, at time T-Bstart (T-Adone < T-Bstart < T-Adone) another thread B starts processing, does the same kind of logging and finishes at T-Bdone. Now, I would like all logging statements from thread A to appear before the logging statements from Thread B, and not interspersed as the MDC examples I have seen. Is this kind of deferred, grouped logging possible with logback ? -- View this message in context: http://logback.10977.n7.nabble.com/Is-grouped-deferred-log-statements-from-… Sent from the Users mailing list archive at Nabble.com.

3 3

Oldest archive files not removed when using SizeAndTimeBasedFNATP when rolling over and there are multiple archive files for the oldest interval.
by Jan-Olav Eide 22 Apr '13

22 Apr '13

Hello, using the following programmatic configuration there seems to be a problem with how the maxHistory argument is implemented. I roll over multiple times withing the time interval (every minute here), creating the following files : (maxHistory = 3) trace.log (active file) trace-2013-03-13-16.14.0.log.zip trace-2013-03-13-16.13.0.log.zip trace-2013-03-13-16.12.2.log.zip trace-2013-03-13-16.12.1.log.zip trace-2013-03-13-16.12.0.log.zip The nex time rollover occurs, the 16.13.0.log.zip file is removed, and not the oldest ones (12.{1,2,3} as I would have expected. Is this intended behaviour ? Or is it a problem with my configuration ? public static Logger createTimeAndSizeRollingLogger(String fileBase, String maxSize, int maxHistory, String resolutionPattern) { LoggerContext loggerContext = (LoggerContext) LoggerFactory.getILoggerFactory(); RollingFileAppender<ILoggingEvent> appender = new RollingFileAppender<ILoggingEvent>(); appender.setContext(loggerContext); appender.setFile(fileBase + ".log"); TimeBasedRollingPolicy<ILoggingEvent> timePolicy = new TimeBasedRollingPolicy<ILoggingEvent>(); timePolicy.setFileNamePattern(fileBase + "_%d{" + resolutionPattern + "}-%i.log.zip"); timePolicy.setContext(loggerContext); timePolicy.setMaxHistory(maxHistory); timePolicy.setParent(appender); appender.setRollingPolicy(timePolicy); SizeAndTimeBasedFNATP<ILoggingEvent> rollingPolicy = new SizeAndTimeBasedFNATP<ILoggingEvent>(); rollingPolicy.setMaxFileSize(maxSize); rollingPolicy.setTimeBasedRollingPolicy(timePolicy); rollingPolicy.setContext(loggerContext); timePolicy.setTimeBasedFileNamingAndTriggeringPolicy(rollingPolicy); timePolicy.start(); rollingPolicy.start(); PatternLayoutEncoder encoder = new PatternLayoutEncoder(); encoder.setContext(loggerContext); encoder.setPattern("%msg%n"); encoder.start(); appender.setEncoder(encoder); appender.start(); Logger logger = loggerContext.getLogger(fileBase); logger.setLevel(Level.TRACE); logger.addAppender(appender); StatusPrinter.print(loggerContext); return logger; }

3 5

Is FixedWindowRollingPolicy useful to anyone?
by ceki 22 Apr '13

22 Apr '13

Hello all, While fixing an issue related to FixedWindowRollingPolicy [1], the question of whether FixedWindowRollingPolicy still served a purpose came up. Indeed, since it is possible to archive log files by both date and size [2], why is FixedWindowRollingPolicy needed, if at all? Archiving in FixedWindowRollingPolicy involves as many file renames as "MaxIndex". File renaming can be expensive on certain file systems. In order to protect users from shooting themselves in the foot, FixedWindowRollingPolicy caps MaxIndex to 12. To cope with this issue, David Roussel suggested a different sliding window policy [1], where index for archived files would be always incremented by one and the file with the oldest (lowest) index would be deleted on each rollover. This would avoid the file renaming problem altogether. Again, is such a policy useful given that we already have SizeAndTimeBasedFNATP [2]? Your comments are welcome, especially if you are from the operational side of things, [1] http://jira.qos.ch/browse/LOGBACK-266 [2] http://logback.qos.ch/manual/appenders.html#SizeAndTimeBasedFNATP -- Ceki 65% of statistics are made up on the spot

2 1

How to send one event to more than one appender using the SiftingAppender?
by Konstantinidis, Johannes 19 Apr '13

19 Apr '13

Hi all, I have the following situation: I have log statements which are marked in any way (f.i. the MDC contains a special entry "id=47") and in addition there are log statements without a mark. Now I want to split the marked statements into files using the SiftingAppender, for example, resulting in files like "foo-47.log, foo-0815.log". But in addition I want to add every unmarked statement to all files. The result should be something like foo-47.log "... 01.01.2013 12:00:00.000 [47] This is a log entry for ID 47 01.01.2013 12:00:00.002 [] This is a log entry for all IDs ..." foo-0815.log "... 01.01.2013 12:00:00.001 [0815] This is a log entry for ID 0815 01.01.2013 12:00:00.002 [] This is a log entry for all IDs ..." So the point is to select one appender using the discriminating value or all appenders in case of the default value. As I haven't found a standard solution my idea is to subclass the SiftingAppender and - add methods to register/deregister values which are used in case of the discriminator returning the default value - add handling for more than one appender Are there any concerns about that? F.i, is the logging event changed in a way that it can't be reused? There may be some problems, because the SiftingAppenderBase contains some package-private code and the code for retrieving the discriminator value and finding the appender is within one method, referencing package-private variables. Any ideas, comments, pitfalls? Thanks in advance Johannes ________________________________ Firma: Capgemini Deutschland GmbH Gesch?ftsf?hrer: Dr. Michael Schulte (Sprecher) * Dr. Uwe Dumslaff * Josef Ranner Aufsichtsratsvorsitzender: Antonio Schnieder Amtsgericht Berlin-Charlottenburg, HRB 98814 This message contains information that may be privileged or confidential and is the property of the Capgemini Group. It is intended only for the person to whom it is addressed. If you are not the intended recipient, you are not authorized to read, print, retain, copy, disseminate, distribute, or use this message or any part thereof. If you receive this message in error, please notify the sender immediately and delete all copies of this message.

3 3

Why are encoder-less appenders not consistently allowed?
by Stuart Rossiter 19 Apr '13

19 Apr '13

I have a need to dynamically adjust appenders at run-time to use a custom Layout. What I really want is to have encoder-less appenders set up (via the config file) and then add the appropriate LayoutWrappingEncoder programmatically. This works fine for a basic FileAppender. *However*, this fails for a ConsoleAppender, or (interestingly) for a FileAppender created dynamically by a SiftingAppender. (You get errors that an encoder has to be set and that the init method can't be invoked.) Is there a reason for this inconsistency? I suspect it might just be that there is inconsistent handling internally of when an appender is started(??). (I assume the init method that is being referred to in the error messages is start().) (The context is that this is logging from a simulation. I want all log messages to be prefixed with the simulation time, which is something that is only available within the application, and differs per logged message. I guess I could use a SimTime MDC key which is updated before each message logged, but that's a bit clumsy. Doing it generically inside a Layout, independent of the rest of the simulation code, is much better.) Stuart -- ________________________________ Stuart Rossiter

2 4

Re: [logback-user] browse application log
by diroussel 17 Apr '13

17 Apr '13

Saty, Your post didn't make it onto the mailing list, as you are not subscribed. It only made it into nabble. I don't know of any, but it sounds like the kind of thing someone would have written before. If you can't find one you could write your own appender which keeps the most recent log messages in a buffer, and exposes them to a servlet. Of you could just server up your log folder with a file servlet. David > Does logback has anything to browse application log at a URL for web > applications. I read about logback access log but i am interested to > browse actual application log in somewhat similar way or download logs, > any pointer to related documentation would be helpful. -- View this message in context: http://logback.10977.n7.nabble.com/browse-application-log-tp11901p11969.html Sent from the Users mailing list archive at Nabble.com.

2 1