log4j.properties vs log4j.xml - log4j

I was asked to move from properties style configuration to xml style. The process itself is straightforward and it is not causing me any trouble, I was simply curious why I got asked this and I looked for the reason. I found this nice post Why chose XML over properties files for Log4J configuration? (more than 3 years old but still the most informative I could find), which was pointing pointing to this even older tutorial. They state
Properties can be defined by a properties file or by an XML file.
Log4j looks for a file named log4j.xml and then for a file named log4j.properties. Both must be placed in the src folder.
I checked it out and, contrary to the statement, found out that Log4j looks first for the properties file and in a second time for xml. In fact, if I keep both in the same folder, the output is formatted according to the properties file. So, am I misunderstood the statement? Is it simply wrong? The statement has been made for log4j 1.2.14 while I am on log4j 1.2.17, is it possible the log4j team changed the priority? I couldn't find any related documentation, but it will be a strange trend inversion since the xml looked to be the preferred way to configure log4j

Have a look at the JavaDoc. The documentation of the PropertyConfiguratorClass (log4j.properties) points out that
The PropertyConfigurator does not handle the advanced configuration
features supported by the DOMConfigurator such as support custom
ErrorHandlers, nested appenders such as the AsyncAppender, etc.
So the DOMConfigurator (log4j.xml) offers advanced options.
Beside that you can have (at least a simple) validity check of log4j.xml files using the provided log4j.dtd.
If youd do not make use of the advanced features it is needless to change from properties to xml files.
What you really should think about is to change from log4j to log4j 2 beta or even
slf4j. Development of log4j has stopped and the founder of it (#Ceki) invented slf4j.

As per FrVaBe's answer, the PropertyConfigurator class does not handle the advanced features that XML can support.
It's worth noting one of the more useful advanced options (only available log4j.xml) is the ability to filter on a specific log level or range of log levels using LevelMatchFilter and DenyAllFilter
A good example can be seen here.

Related

Equivalent of PropertyConfigurator.configureAndWatch() in log4j2?

I am updating some software from log4j1 to log4j2. It uses PropertyConfigurator.configureAndWatch(String configFilename, long delay), for which I do not see any corresponding class or method in log4j2.
How do I update this to log4j2? What's the most straightforward way?
Do I have to make my own implementation of configureAndWatch, in a class inheriting from ConfigurationFactory?
I found this answer in the log4j2 user's guide.
So the configureAndWatch time can be set in the configuration file via the monitorInterval attribute.
Unfortunately this doesn't solve my situation as the configureAndWatch call I am trying to update to log4j2 occurs within a conditional. So though the monitorInterval attribute can be set in the properties file, I still do not see how to set the value conditionally based on runtime variables.

Logstash Shipper configuration for Jira

I am running Jira and Confluence within my company. I would like the logfiles to be shipped to Kibana.
This is very easy to do but I do not want to rewrite the Grok filters. I cannot imagine that nobody has done this already.
Does anybody have an example of a logstash shipper configuration. Most of the logging like catalina.log is standard.
Please help me with examples
One would think that Java application logs only come in one form, but my experience is that there often are subtle differences. Sometimes the thread name is in square brackets and sometimes in parentheses, sometimes the thread name goes first and other times after the logger name, and so on. This gets more painful as you attempt to parse more than one type of log.
Instead of messing with various filters to join multiline messages and grok all the fields I strongly favor using the Log4j layout in github.com/logstash/log4j-jsonevent-layout to produce JSON-based logs that Logstash can read directly without any filters. Apart from not having to maintain filters you get all fields from each log event. Since I don't know what your catalina.log looks like I can't say what you'd be missing by parsing its contents instead of using the JSON layout.
The drawback is that it's a bit more work deployment-wise. You obviously have to deploy the layout jar file itself, but it has a couple of dependencies of its own (net.minidev:jsonsmart and commons-lang:commons-lang) that you need to make available too.

log4net configuration: Can I refer to the same layout in several appenders?

I want to send log messages to several files (i.e. different appenders) based upon some property of the message.
The problem is that each appender needs to specify quite a verbose layout (that contains a compication conversionPattern and a couple of converters). I have ended up duplicating this configuration in each appender. This works but is not ideal as it makes the config much longer than I would like as well as the pain of having to update 3 complicated bit of configuration when the layout changes.
I want to be able to define the layout once and have all my appenders refer back to that one definition (in the same way that several loggers can refer to the same appender). But perhaps there is a better way to achieve my goal of reducing duplication in the configuration?
My google-foo is weak and I could not find an answer. Can anyone here help?
TIA.
I am sorry but unless you are ready to code your own Layout class there is currently no way to skip on the copy-pasted configuration.
You can inherit from the LayoutSkeleton to get started, and either build your layout in code or use an alternate configuration file (I don't think that log4net would be kind to a dangling layout configuration in its config file)

RollingFileAppender: preserve the file extension without log4net patch

M question is related to this one as I have the same problem. How can I make the RollingFileAppender of log4net preserve the file extension without having to actually "patch" (i.e. create an interim build of the current trunk) log4net? How could I achieve the same thing by overriding the RollingFileAppender or creating any other extension with the least effort?
As mentioned in my comment, you're not required to "patch" or change anything in log4net. You cannot use the 1.2.10 release, though. It is a rather old build, but the PreserveLogFileNameExtension is in the current source. You only have to bring down the source and compile the dll yourself and you're good to go.
Update: I see your concern. IMO, the only way is to make a build and test it. To further your confidence you could review the current list of resolved (and unresolved) issues for 1.2.11. I would think that seeing they are not doing a major revision, changes are mainly bug fixes and additional features.

Why doesn't grails.views.default.codec default to "html"?

The Grails Config.groovy setting grails.views.default.codec specifies the default codec used to encode data within ${...} in Grails views.
This config setting can take any of the values none (no filtering required), html (to avoid XSS-attacks) and base64 (has no real-world use-case that I know of).
The Grails default is none (no filtering).
Questions:
Are there any compelling technical reasons not to use the safer option "html"?
When do you choose to go with the default option of "none" in your Grails projects?
A question on similar topic here.. I don't claim big expertise on this, but I imagine. Why it's not html by default is strange to me. I found GRAILS-2945, where this was proposed but ultimately rejected, without too much explanation. There is also some more information in GRAILS-1827, when the issue was first implemented.

Resources