[ 
https://issues.apache.org/jira/browse/LOG4J2-3639?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17642631#comment-17642631
 ] 

Alex Lambert commented on LOG4J2-3639:
--------------------------------------

Absolutely, I'll get some measurements for this.

> RollingFileAppender: Add option to allow compression of logs as they are 
> written
> --------------------------------------------------------------------------------
>
>                 Key: LOG4J2-3639
>                 URL: https://issues.apache.org/jira/browse/LOG4J2-3639
>             Project: Log4j 2
>          Issue Type: New Feature
>          Components: Appenders
>            Reporter: Alex Lambert
>            Priority: Minor
>
> We have a forked implementation of this functionality that I’m hoping to get 
> merged in – I'm raising this issue now just in case there are any obvious 
> problems with taking the feature that should be addressed before that starts.
> h2. Summary
> The motivation is to add an option to allow compression of log content before 
> it is written to file, saving on I/O operations. There was some discussion 
> around this in [RollingFileAppender enhancement for file compression-Apache 
> Mail 
> Archives|https://lists.apache.org/thread/0wj82d168dcjozcd3fdnp6gk9mm1goy0]. A 
> couple of issues were raised:
>  * Leaving the file in a bad state after the application dies.
>  * Reading the file as it is written.
> Both of these are resolved by making sure that the file is in a valid state 
> after each write, described in the next section.
> h2. Proposed Change
> Add new config option to RollingFileAppender “compressOnWrite”. If true, 
> write to a GZIPOutputStream instead of a FileOutputStream. Skip the 
> compression action step on rollover if this config option is set. Close the 
> gzip stream between writes to ensure that the log file is always in a valid 
> state.
> We only have this functionality implemented for gzip compression right now.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to