XWiki 13.8 and 13.9: memory leaks and forcibly unregistered drivers

Help / Discuss
1

Hi! I keep workingto put in place a brand new XWiki instance running in CentOS Linux release 8.4.2105. Some weeks ago I installed XWiki 13.8 and recently updated to 13.9. My current configuration is

XWiki 13.9 Demo Flavor 1.1

apache-tomcat-9.0.54
MySQL Server version: 8.0.26 Source distribution
mysql-connector-java-8.0.27.jar

CentOS Linux release 8.4.2105

[root@e220i059h057w rjr]# java -version
openjdk version “1.8.0_312”
OpenJDK Runtime Environment (build 1.8.0_312-b07)
OpenJDK 64-Bit Server VM (build 25.312-b07, mixed mode)

When I stop Tomcat, I get the following WARNINGS:

[root@e220i059h057w rjr]# sudo cat /opt/apache-tomcat-current/logs/catalina.out|grep WARNING|grep 19:17
02-Nov-2021 19:17:53.263 WARNING [main] org.apache.catalina.loader.WebappClassLoaderBase.clearReferencesJdbc The web application [xwiki] registered the JDBC driver [org.apache.calcite.jdbc.Driver] but failed to unregister it when the web application was stopped. To prevent a memory leak, the JDBC Driver has been forcibly unregistered.
02-Nov-2021 19:17:53.263 WARNING [main] org.apache.catalina.loader.WebappClassLoaderBase.clearReferencesJdbc The web application [xwiki] registered the JDBC driver [org.apache.solr.handler.sql.CalciteSolrDriver] but failed to unregister it when the web application was stopped. To prevent a memory leak, the JDBC Driver has been forcibly unregistered.
02-Nov-2021 19:17:53.264 WARNING [main] org.apache.catalina.loader.WebappClassLoaderBase.clearReferencesJdbc The web application [xwiki] registered the JDBC driver [com.mysql.cj.jdbc.Driver] but failed to unregister it when the web application was stopped. To prevent a memory leak, the JDBC Driver has been forcibly unregistered.
02-Nov-2021 19:17:53.265 WARNING [main] org.apache.catalina.loader.WebappClassLoaderBase.clearReferencesThreads The web application [xwiki] appears to have started a thread named [RxSchedulerPurge-1] but has failed to stop it. This is very likely to create a memory leak. Stack trace of thread:
02-Nov-2021 19:17:53.266 WARNING [main] org.apache.catalina.loader.WebappClassLoaderBase.clearReferencesThreads The web application [xwiki] appears to have started a thread named [RxCachedWorkerPoolEvictor-1] but has failed to stop it. This is very likely to create a memory leak. Stack trace of thread:
02-Nov-2021 19:17:53.267 WARNING [main] org.apache.catalina.loader.WebappClassLoaderBase.clearReferencesThreads The web application [xwiki] appears to have started a thread named [searcherExecutor-15-thread-1] but has failed to stop it. This is very likely to create a memory leak. Stack trace of thread:
02-Nov-2021 19:17:53.268 WARNING [main] org.apache.catalina.loader.WebappClassLoaderBase.clearReferencesThreads The web application [xwiki] appears to have started a thread named [searcherExecutor-21-thread-1] but has failed to stop it. This is very likely to create a memory leak. Stack trace of thread:
02-Nov-2021 19:17:53.269 WARNING [main] org.apache.catalina.loader.WebappClassLoaderBase.clearReferencesThreads The web application [xwiki] appears to have started a thread named [searcherExecutor-27-thread-1] but has failed to stop it. This is very likely to create a memory leak. Stack trace of thread:
02-Nov-2021 19:17:53.270 WARNING [main] org.apache.catalina.loader.WebappClassLoaderBase.clearReferencesThreads The web application [xwiki] appears to have started a thread named [mysql-cj-abandoned-connection-cleanup] but has failed to stop it. This is very likely to create a memory leak. Stack trace of thread:
[root@e220i059h057w rjr]#

I’m planning to move to a Data Source to the container level, looking for a better integration between the container, the JDBC driver, MySQL and XWiki.

There are also three threads named searchExectur-nn. that seems related with Solr. I found at least one Jira could be related issue:

https://jira.xwiki.org/browse/XWIKI-18387

And finally two other threads, RxSchedulerPurge-1 and RxCachedWorkerPoolEvictor-1 about which I can not find any information in the forum or Jira.

Please, have you any idea of the source of the warning? Must I be worried about any of them? Until now, I’ve just created a few pages, and nothing weird seems to happen. Still, I would like to explain the WARNINGS and their relevance before going ahead with any further configuration or creating content.

Thanks!

2

Depends what you mean by “stop” exactly. If you completely stop Tomcat as in the Tomcat JVM does not run anymore in ps aux then it’s not a problem, those warning are just here to indicate that if you restart the application alone without completely stopping Java (which is not recommended) then those unstopped threads might become ghosts.

3

Thanks. Thus, it won’t be an issue as I’m using a test server, and I can completely stop the container without affecting other applications. But I would need to explain why these WARNINGs happen and how we can solve them to get XWiki working in a production server connected to the Internet.

Most of the WARNINGS seams related to the JDBC driver, MySQL and Solr. Please, does it make sense to you to move to a Data Source at the container level? Could this contribute to stopping all the threads correctly?

4

There’s absolutely no problem in production if you stop tomcat and start it again when you want to start/stop xwiki. What you should avoid is to undeploy and redeploy the xwiki webapp while leaving tomcat running.

See Loading...

Other related issues:

1 Like