ambari-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jonathan Hurley" <jhur...@hortonworks.com>
Subject Re: Review Request 37264: EclipseLink stdout Log Statements Fill Up Disk Space
Date Mon, 10 Aug 2015 12:47:00 GMT

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/37264/
-----------------------------------------------------------

(Updated Aug. 10, 2015, 8:46 a.m.)


Review request for Ambari, Alejandro Fernandez, Nate Cole, Sumit Mohanty, and Sid Wagle.


Bugs: AMBARI-12685
    https://issues.apache.org/jira/browse/AMBARI-12685


Repository: ambari


Description
-------

EclipseLink's default logger logs to standard output. Since Ambari redirects stdout to a file,
this means that all EclipseLink statements are captured in {{ambari-server.out}}.

On some systems with latency problems, EclipseLink will log retry attempts. Since this log
is not rolling, it can fill up the disk.

There is no log4j bridge for EclipseLink. This Jira will track the creation of an slf4j bridge
that will log all SQL statements to a separate, rolling logger.


Diffs (updated)
-----

  ambari-server/conf/unix/log4j.properties f619c6c 
  ambari-server/conf/windows/log4j.properties 607cdf9 
  ambari-server/src/main/conf/log4j.properties 6d53ca3 
  ambari-server/src/main/java/org/apache/ambari/server/logging/EclipseLinkLogger.java PRE-CREATION

  ambari-server/src/main/resources/META-INF/persistence.xml deb30bb 

Diff: https://reviews.apache.org/r/37264/diff/


Testing
-------

Verified new logger file created and is rolling. Examples of bridged log statements:

[EL Debug]: 2015-08-08 22:17:34.305--ServerSession(1653567754)--Connection(2013352132)--Thread(Thread[qtp-ambari-agent-57,5,main])--
SELECT definition_id, cluster_id, component_name, definition_name, description, ENABLED, HASH,
ignore_host, label, schedule_interval, scope, service_name, alert_source, source_type FROM
alert_definition WHERE (((service_name = ?) AND (component_name = ?)) AND (cluster_id = ?))
	bind => [YARN, RESOURCEMANAGER, 2]
[EL Debug]: 2015-08-08 22:17:34.306--ServerSession(1653567754)--Connection(761406242)--Thread(Thread[qtp-ambari-agent-57,5,main])--
SELECT definition_id, cluster_id, component_name, definition_name, description, ENABLED, HASH,
ignore_host, label, schedule_interval, scope, service_name, alert_source, source_type FROM
alert_definition WHERE (((service_name = ?) AND (component_name = ?)) AND (cluster_id = ?))
	bind => [ZOOKEEPER, ZOOKEEPER_SERVER, 2]
[EL Debug]: 2015-08-08 22:17:34.307--ServerSession(1653567754)--Connection(941036240)--Thread(Thread[qtp-ambari-agent-57,5,main])--
SELECT definition_id, cluster_id, component_name, definition_name, description, ENABLED, HASH,
ignore_host, label, schedule_interval, scope, service_name, alert_source, source_type FROM
alert_definition WHERE ((((service_name IN (?,?,?,?)) AND (scope = ?)) AND (cluster_id = ?))
AND (component_name IS NULL))
	bind => [ZOOKEEPER, MAPREDUCE2, HDFS, YARN, SERVICE, 2]


Thanks,

Jonathan Hurley


Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message