Re: additivity attribute does not work for user-defined logger but works for root logger.
Thanks. I like your idea with the constructor but I call also the code that I
do not have an acces and I cannot modify, so I needed to look for something
different.
I have decided to use apache-log4j-extras-1.0.jar and use
filters(org.apache.log4j.filter.ExpressionFilter).
In each specific class I give an unique name for the thread associated with
the tool.
Then my log4j.xml looks like:
log4j:configuration debug=true
xmlns:log4j=http://jakarta.apache.org/log4j/;
appender name=addQuestionSequenceOrder
class=org.apache.log4j.RollingFileAppender
layout class=org.apache.log4j.PatternLayout
/layout
filter class=org.apache.log4j.filter.ExpressionFilter
/
/filter
filter class=org.apache.log4j.filter.DenyAllFilter/
/appender
appender name=taskRedirection
class=org.apache.log4j.RollingFileAppender
layout class=org.apache.log4j.PatternLayout
/layout
filter class=org.apache.log4j.filter.ExpressionFilter
/filter
filter class=org.apache.log4j.filter.DenyAllFilter/
/appender
root
level value=debug/
appender-ref ref=addQuestionSequenceOrder/
appender-ref ref=taskRedirection/
/root
/log4j:configuration
I hope it would help anybody else.
NCP instead of THREAD could be used as well, but it required additional
changes to the code.
M.
Jacob Kjome wrote:
Camer38 wrote:
I have two Java classes that used log4j.
curam.tools.upgradehelper.ieg.QuestionSequenceOrderConverter and
curam.tools.upgradehelper.taskredirection.TaskRedirectionConverter
Both classes from above uses different class:
curam.tools.upgradehelper.util.SQLStatements
where a logger in each of the class with defined in following pattern:
private static final Logger log = Logger.getLogger(NameOfTheClass.class);
Next I have defined log4j.xml. I want to have a full log file in two
separate files for both classes.
curam.tools.upgradehelper.ieg.QuestionSequenceOrderConverter in
QuestionSequenceOrderConverterTool.log
curam.tools.upgradehelper.taskredirection.TaskRedirectionConverter in
TaskRedirectionConverterTool.log
I did a following:
I have not defined root at all.
My assumption was that if I run one of the class
curam.tools.upgradehelper.ieg.QuestionSequenceOrderConverter I would got
all
the logs from that class and all the classes used internally by
curam.tools.upgradehelper.ieg.QuestionSequenceOrderConverter in the
addQuestionSequenceOrderTool.log.
That's simply not the way Log4j works. Logger naming is hierarchical, and
though
this does not have to be packaged-based, that's the traditional usage.
Instead I got an log4j warning:
log4j:WARN No appenders could be found for logger
(curam.tools.upgradehelper.util.SQLStatements).
...because curam.tools.upgradehelper.util.SQLStatements is not part of the
naming
hierarchy of either of the defined loggers. This is why you usually
define the
ROOT logger as a catch-all.
What is the meaning of the attribute additivity for each logger?
Why don't I get all the logs in one file?
During my testing I have defined a root logger as well, but it does not
do
what I want.
All the expected lines are present in both files but the files are
populated
with the log from the same class.
root
level value=info/
appender-ref ref=taskRedirection/
appender-ref ref=addQuestionSequenceOrder/
/root
As I said previously, using package names is the traditional usage, but
it's not
the only way to name loggers. You can use any arbitrary naming scheme.
Of
course, the problem you are running into here is that you want one utility
class's
logging to go to one of two separate files depending on which class used
the
utility. That's only possible if you design the utility class to change
it's
logging name each time it is used. You'll have do add a constructor that
takes a
logger name, such as...
public SQLStatements(String loggerName) {...}
And then each time you use this helper class, pass in the logger name to
use. For
instance...
public class QuestionSequenceOrderConverter {
public void doSomething() {
SQLStatements ss = new SQLStatements(this.getClass().getName());
}
}
The other way to do this is to use something like SLF4J/Logback's Markers
to
associate distinct loggers in a user-defined way. Log4j 1.3 started with
this,
but it was abandoned. I believe there's work in progress for something
like this
in the Log4j extras or companions, though i can't remember what it's
called, offhand?
I use log4j-1.2.15.jar
Could anybody help me with defining a correct log4j.xml file?
You're looking for additivity=false, not
RE: additivity attribute does not work for user-defined logger but works for root logger.
Now things getting clearer. Log4j maintains a logger hierarchy which is determined by the dot withing the logger names, similar to the package hierarchy in java.Example: Config: you declare a logger foo.bar and a root logger, both have one (the same) appender attached. Code: class foo.bar.XY instantiates the logger foo.bar.XY and issues a log call to it. - Log4j will look for a logger foo.bar.XY which is not declared. After that it climbs up the hierarchy path. Thus it looks for logger foo.bar which is found. The appenders of this logger will be served. Now the additivity flag comes into play. If it is set to true (the default), log4j will go on searching the next higher logger in hierarchy foo. No logger will be found. At the end it reaches the root logger and serves the appenders of the root logger. If the additivity flag of foo.bar would be set to false, log4j would stop looking further up for more loggers. With additivity=true you will end up to have the same log statement twice in your appender, with additivity=false only the appenders of the logger foo.bar were served. Heri -Original Message- From: Camer38 [mailto:[EMAIL PROTECTED] Sent: Thursday, July 03, 2008 6:22 PM To: log4j-user@logging.apache.org Subject: Re: additivity attribute does not work for user-defined logger but works for root logger. Thanks for your answer and your time. Why the root logger defined: root level value=debug/ appender-ref ref=taskRedirection/ appender-ref ref=addQuestionSequenceOrder/ /root knows that all the logs from ALL the classes should go the the file specified by appenders:taskRedirection, addQuestionSequenceOrder and a logger: logger name=curam.tools.upgradehelper.ieg.QuestionSequenceOrderConverter additivity=true level value=debug/ appender-ref ref=addQuestionSequenceOrder/ /logger knows that ONLY logs from that class curam.tools.upgradehelper.ieg.QuestionSequenceOrderConverter has to be served? Is no't realy any way to duplcate the work of the root logger in logger specified by the name curam.tools.upgradehelper.ieg.QuestionSequenceOrderConverter? If not, why this parameter additivity=true is present at all? It does nothing. MR. Thorbjørn Ravn Andersen wrote: Camer38 skrev den 03-07-2008 15:25: My assumption was that if I run one of the class curam.tools.upgradehelper.ieg.QuestionSequenceOrderConverter I would got all the logs from that class and all the classes used internally by curam.tools.upgradehelper.ieg.QuestionSequenceOrderConverter in the addQuestionSequenceOrderTool.log. No. The logger frameworks do not do magic to deduce where logging should go. If you want for classes called to use a specific logger, you must pass it as an argument like all values you want to use. Instead I got an log4j warning: log4j:WARN No appenders could be found for logger (curam.tools.upgradehelper.util.SQLStatements). You are most likely using a logger variable in SQLStatements which has been defined at the top of that class referring to SQLStatments.class. This does not allow log4j to distinguish between the two cases you have. I have the same problem in a project where it is two threads logging in the same file. I have not yet found a good solution. During my testing I have defined a root logger as well, but it does not do what I want. All the expected lines are present in both files but the files are populated with the log from the same class. root level value=info/ appender-ref ref=taskRedirection/ appender-ref ref=addQuestionSequenceOrder/ /root This say that you want all messages at info level or higher to go to both the taskRedirection appender AND the addQuestionSequenceOrder appender. -- Thorbjørn - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] -- View this message in context: http://www.nabble.com/%22additivity%22-attribute-does-not-work -for-user-defined-logger-but-works-for-root-logger.-tp18259092 p18263109.html Sent from the Log4j - Users mailing list archive at Nabble.com. - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
RE: additivity attribute does not work for user-defined logger but works for root logger.
Thank you very much for the explanation. It makes perfect sense for me. But in my case I hit slightly different problem. The class foo.bar.XY calls lets say test.bulk.ZY and the class com.create.CR calls test.bulk.ZY as well. I need two full log files for foo.bar.XY and test.bulk.ZY with all the logs information from test.bulk.ZY. How to do that with two loggers defined for each of the foo.bar.XY and test.bulk.ZY? Bender Heri wrote: Now things getting clearer. Log4j maintains a logger hierarchy which is determined by the dot withing the logger names, similar to the package hierarchy in java.Example: Config: you declare a logger foo.bar and a root logger, both have one (the same) appender attached. Code: class foo.bar.XY instantiates the logger foo.bar.XY and issues a log call to it. - Log4j will look for a logger foo.bar.XY which is not declared. After that it climbs up the hierarchy path. Thus it looks for logger foo.bar which is found. The appenders of this logger will be served. Now the additivity flag comes into play. If it is set to true (the default), log4j will go on searching the next higher logger in hierarchy foo. No logger will be found. At the end it reaches the root logger and serves the appenders of the root logger. If the additivity flag of foo.bar would be set to false, log4j would stop looking further up for more loggers. With additivity=true you will end up to have the same log statement twice in your appender, with additivity=false only the appenders of the logger foo.bar were served. Heri -Original Message- From: Camer38 [mailto:[EMAIL PROTECTED] Sent: Thursday, July 03, 2008 6:22 PM To: log4j-user@logging.apache.org Subject: Re: additivity attribute does not work for user-defined logger but works for root logger. Thanks for your answer and your time. Why the root logger defined: root level value=debug/ appender-ref ref=taskRedirection/ appender-ref ref=addQuestionSequenceOrder/ /root knows that all the logs from ALL the classes should go the the file specified by appenders:taskRedirection, addQuestionSequenceOrder and a logger: logger name=curam.tools.upgradehelper.ieg.QuestionSequenceOrderConverter additivity=true level value=debug/ appender-ref ref=addQuestionSequenceOrder/ /logger knows that ONLY logs from that class curam.tools.upgradehelper.ieg.QuestionSequenceOrderConverter has to be served? Is no't realy any way to duplcate the work of the root logger in logger specified by the name curam.tools.upgradehelper.ieg.QuestionSequenceOrderConverter? If not, why this parameter additivity=true is present at all? It does nothing. MR. Thorbjørn Ravn Andersen wrote: Camer38 skrev den 03-07-2008 15:25: My assumption was that if I run one of the class curam.tools.upgradehelper.ieg.QuestionSequenceOrderConverter I would got all the logs from that class and all the classes used internally by curam.tools.upgradehelper.ieg.QuestionSequenceOrderConverter in the addQuestionSequenceOrderTool.log. No. The logger frameworks do not do magic to deduce where logging should go. If you want for classes called to use a specific logger, you must pass it as an argument like all values you want to use. Instead I got an log4j warning: log4j:WARN No appenders could be found for logger (curam.tools.upgradehelper.util.SQLStatements). You are most likely using a logger variable in SQLStatements which has been defined at the top of that class referring to SQLStatments.class. This does not allow log4j to distinguish between the two cases you have. I have the same problem in a project where it is two threads logging in the same file. I have not yet found a good solution. During my testing I have defined a root logger as well, but it does not do what I want. All the expected lines are present in both files but the files are populated with the log from the same class. root level value=info/ appender-ref ref=taskRedirection/ appender-ref ref=addQuestionSequenceOrder/ /root This say that you want all messages at info level or higher to go to both the taskRedirection appender AND the addQuestionSequenceOrder appender. -- Thorbjørn - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] -- View this message in context: http://www.nabble.com/%22additivity%22-attribute-does-not-work -for-user-defined-logger-but-works-for-root-logger.-tp18259092 p18263109.html Sent from the Log4j - Users mailing list archive at Nabble.com. - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional
RE: additivity attribute does not work for user-defined logger but works for root logger.
Questions see inline. Heri -Original Message- From: Camer38 [mailto:[EMAIL PROTECTED] Sent: Monday, July 07, 2008 3:24 PM To: log4j-user@logging.apache.org Subject: RE: additivity attribute does not work for user-defined logger but works for root logger. But in my case I hit slightly different problem. The class foo.bar.XY calls lets say test.bulk.ZY and the class com.create.CR calls test.bulk.ZY as well. Do this two classes (foo.bar.XY and com.create.CR) run in different threads? Or is it the same thread? - Influences possible solutions! I need two full log files for foo.bar.XY and test.bulk.ZY with all the logs information from test.bulk.ZY. Does not make any sense for me. Is it a typo? (test.bulk.ZY in first line should be com.create.CR?). If so, then next question: Do you want the logs of test.bulk.ZY which were called by foo.bar.XY in the log file of foo.bar and the logs of test.bulk.ZY which were called by com.create.CR in the log file of com.create.CR? Otherwise you have to explain more detailled which log calls you want to appear in which appender. - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
RE: additivity attribute does not work for user-defined logger but works for root logger.
Yes, it is a typo. Should be: I need two full log files for foo.bar.XY and com.create.CR with all the logs information from test.bulk.ZY Yes the foo.bar.XY and com.create.CR are run in different threads. I want two logs files - one for foo.bar.XY and the second one for com.create.CR. All of them should have all the calls made in test.bulk.ZY (treat it as utility class) Do you want the logs of test.bulk.ZY which were called by foo.bar.XY in the log file of foo.bar.XY and the logs of test.bulk.ZY which were called by com.create.CR in the log file of com.create.CR? Yes. the foo.bar.XY and com.create.CR has some logging messages as well that needs to go to the right log file. Thanks for helping me with it. M. Bender Heri wrote: Questions see inline. Heri -Original Message- From: Camer38 [mailto:[EMAIL PROTECTED] Sent: Monday, July 07, 2008 3:24 PM To: log4j-user@logging.apache.org Subject: RE: additivity attribute does not work for user-defined logger but works for root logger. But in my case I hit slightly different problem. The class foo.bar.XY calls lets say test.bulk.ZY and the class com.create.CR calls test.bulk.ZY as well. Do this two classes (foo.bar.XY and com.create.CR) run in different threads? Or is it the same thread? - Influences possible solutions! I need two full log files for foo.bar.XY and test.bulk.ZY with all the logs information from test.bulk.ZY. Does not make any sense for me. Is it a typo? (test.bulk.ZY in first line should be com.create.CR?). If so, then next question: Do you want the logs of test.bulk.ZY which were called by foo.bar.XY in the log file of foo.bar and the logs of test.bulk.ZY which were called by com.create.CR in the log file of com.create.CR? Otherwise you have to explain more detailled which log calls you want to appear in which appender. - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] -- View this message in context: http://www.nabble.com/%22additivity%22-attribute-does-not-work-for-user-defined-logger-but-works-for-root-logger.-tp18259092p18319243.html Sent from the Log4j - Users mailing list archive at Nabble.com. - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
BufferingForwardingAppender
In the Java implementation of Log4j is there something similar to the BufferingForwardingAppender? We are trying to setup our Java projects in a similar fashion to our .Net projects. I have searched around and haven't seen anything. Thanks, Tom Henricksen Consultant Advanced Technologies Group, Inc.
RE: additivity attribute does not work for user-defined logger but works for root logger.
OK, in the following I denote foo.bar.XY as AppA and com.create.CR as
AppB and the test.bulk.ZY as UtilityC.
Possible solutions:
1. If the UtilityC is instantiated separatly by AppA and AppB (two
instances) you are perfectly done by configuring each instance with the
desired logger:
Config:
appender A
appender B
Logger AppA
with appender A
Logger AppB
with appender B
Logger AppA.UtilityC
with appender A
Logger AppB.UtilityC
with appender B
Code:
Use the suggestion of jake Kjome from today morning. The
constructor of SQLStatements wich receives the classname of the calling
class must then expand the logger name: receivedClassName + . +
this.getClass.getSimpleName(). With this logger name you retrieve the
logger from the repository. If AppA had called the constructor, you get
the Logger AppA.UtilityC, and so on.
2. Supply the desired logger on each call to the UtilityC class:
public static void UtilityC.doSomething( Logger aLoggerToUse );
{...}
Drawback: The location info in the log statements would be the
location of the call to doSomething(). You can avoid this by using in
the UtilityC the generic call of Logger log():
aLoggerToUse.log ( this.getClass.getName(), Level.DEBUG, logMsg,
aThrowable )
3. Since the AppA and AppB run in different threads you could use a MDC
value within UtilityC in order to distinguish the loggers. Maybe in
combination with a own Repository-Selector which relies on the actual
values in MDC. Search the archive, this was discussed a lot, and I
supplied whole examples.
I hope you got the hint in which direction you should think. Otherwise
ask again.
Heri
-Original Message-
From: Camer38 [mailto:[EMAIL PROTECTED]
Sent: Monday, July 07, 2008 5:23 PM
To: log4j-user@logging.apache.org
Subject: RE: additivity attribute does not work for
user-defined logger but works for root logger.
Yes, it is a typo.
Should be:
I need two full log files for foo.bar.XY and com.create.CR
with all the logs information from test.bulk.ZY
Yes the foo.bar.XY and com.create.CR are run in different threads.
I want two logs files - one for foo.bar.XY and the second one
for com.create.CR.
All of them should have all the calls made in test.bulk.ZY
(treat it as utility class)
Do you want the logs of test.bulk.ZY which were called by
foo.bar.XY in the log file of foo.bar.XY and the logs of
test.bulk.ZY which were called by com.create.CR in the log
file of com.create.CR?
Yes. the foo.bar.XY and com.create.CR has some logging
messages as well that needs to go to the right log file.
Thanks for helping me with it.
M.
Bender Heri wrote:
Questions see inline.
Heri
-Original Message-
From: Camer38 [mailto:[EMAIL PROTECTED]
Sent: Monday, July 07, 2008 3:24 PM
To: log4j-user@logging.apache.org
Subject: RE: additivity attribute does not work for user-defined
logger but works for root logger.
But in my case I hit slightly different problem.
The class foo.bar.XY calls lets say test.bulk.ZY and the class
com.create.CR calls test.bulk.ZY as well.
Do this two classes (foo.bar.XY and com.create.CR) run in different
threads? Or is it the same thread? - Influences possible solutions!
I need two full log files for foo.bar.XY and test.bulk.ZY with all
the logs information from test.bulk.ZY.
Does not make any sense for me. Is it a typo?
(test.bulk.ZY in first
line should be com.create.CR?). If so, then next question: Do you
want the logs of test.bulk.ZY which were called by
foo.bar.XY in the
log file of foo.bar and the logs of test.bulk.ZY which were
called by
com.create.CR in the log file of com.create.CR?
Otherwise you have to explain more detailled which log
calls you want
to appear in which appender.
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
--
View this message in context:
http://www.nabble.com/%22additivity%22-attribute-does-not-work
-for-user-defined-logger-but-works-for-root-logger.-tp18259092
p18319243.html
Sent from the Log4j - Users mailing list archive at Nabble.com.
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
RE: precision of log4j
Thanks Micheal, I do get better resolution timing now J I am also trying to time how long it takes a particular message to travel from identical machines A to B. That is propagation time = receive time – sent time. I cannot use System.nanoTime(); in this case and the resolution of System.currentTimeMillis() (i.e. 15.625ms) is unacceptable: On Machine A: long sent_time = System.currentTimeMillis(); out.write (msg); On Machine B: in.read(msg); long receive_time = System.currentTimeMillis(); Despite synchronizing the time on the two machines using NTP, I am getting not only poor resolution but also negative propagation values which are ridiculous. Any idea about how I can get better resolution is this scenario? Regards, Sana --- On Tue, 6/24/08, Michael Erskine [EMAIL PROTECTED] wrote: From: Michael Erskine [EMAIL PROTECTED] Subject: RE: precision of log4j To: Log4J Users List log4j-user@logging.apache.org Date: Tuesday, June 24, 2008, 4:16 PM From: sana qadir [mailto:[EMAIL PROTECTED] Sent: 24 June 2008 05:40 Subject: precision of log4j I am using log4j to help me record the execution times of several functions (1 to n) using code similar to the following snippet: logger.debug(Timestamp before...); function_n(); logger.debug(Timestamp after...); OK, Log4J isn't necessarily going to do what you want: if you want high resolution timings you can use System.nanoTime and to make the timings accurate, don't corrupt them by logging while you're timing... long t1 = System.nanoTime(); function_n(); long t2 = System.nanoTime(); logger.debug(duration in nanos: +(t2 -t1)); Regards, Michael Erskine. - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Replace a file in DailyRollingFileAppender
Hi All, Is there any way we can replace a file attached to a particular appender using java code? I am using DailyRollingFileAppender for my purpose and want to log to separate files depending upon certain configurations made in the database. Can we dynamically change the files an appender is writing to? If so, how? Please help. An early response shall be highly appreciated. Thanks and Regards, Rajat Aggarwal
