[
https://issues.apache.org/jira/browse/HADOOP-10413?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Arpit Agarwal updated HADOOP-10413:
-----------------------------------
Description:
Log statements do not include process IDs and thread IDs which makes debugging
hard when the output of multiple requests is interleaved. It's even worse when
looking at the output of test runs because the logs from multiple daemons are
interleaved in the same file.
Log4j does not provide a builtin mechanism for this, so we'd likely have to
write some extra code. One possible solution is to initialize the IDs in the
[MDC|https://logging.apache.org/log4j/1.2/apidocs/org/apache/log4j/MDC.html]
and extract via {{ConversionPattern}} as [described
here|http://stackoverflow.com/a/12202124].
was:
Log statements do not include process IDs and thread IDs which makes debugging
hard when the output of multiple requests is interleaved. It's even worse when
looking at the output of test runs because the logs from multiple daemons are
interleaved in the same file.
Log4j does not provide a builtin mechanism for this, so we'd likely have to
write some extra code. One possible solution is to initialize the IDs in the
[MDC|https://logging.apache.org/log4j/1.2/apidocs/org/apache/log4j/MDC.html]
and extract by updating the {{ConversionPattern}} as [described
here|http://stackoverflow.com/a/12202124].
> Log statements must include pid and tid information
> ---------------------------------------------------
>
> Key: HADOOP-10413
> URL: https://issues.apache.org/jira/browse/HADOOP-10413
> Project: Hadoop Common
> Issue Type: Improvement
> Affects Versions: 3.0.0, 2.3.0
> Reporter: Arpit Agarwal
>
> Log statements do not include process IDs and thread IDs which makes
> debugging hard when the output of multiple requests is interleaved. It's even
> worse when looking at the output of test runs because the logs from multiple
> daemons are interleaved in the same file.
> Log4j does not provide a builtin mechanism for this, so we'd likely have to
> write some extra code. One possible solution is to initialize the IDs in the
> [MDC|https://logging.apache.org/log4j/1.2/apidocs/org/apache/log4j/MDC.html]
> and extract via {{ConversionPattern}} as [described
> here|http://stackoverflow.com/a/12202124].
--
This message was sent by Atlassian JIRA
(v6.2#6252)
