portals-jetspeed-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Ate Douma (JIRA)" <jetspeed-...@portals.apache.org>
Subject [jira] [Updated] (JS2-1183) Tomcat 6.0.24+ logs SEVERE errors on the console during shutdown from WebappClassLoader clearReferencesThreads and clearThreadLocalMap
Date Wed, 11 May 2011 10:20:47 GMT

     [ https://issues.apache.org/jira/browse/JS2-1183?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Ate Douma updated JS2-1183:
---------------------------

             Priority: Minor  (was: Major)
    Affects Version/s:     (was: 2.2.1)
        Fix Version/s:     (was: 2.2.2)
             Assignee:     (was: Ate Douma)

Basic background for what is happening can be read here: http://wiki.apache.org/tomcat/MemoryLeakProtection

As can be found by a simple google search, this "issue" is hampering everyone big scale. And
IMO causes far too much noise (by Tomcat) than needed.
Primary "cause" of all this is, that the Tomcat thread pooling keeps hanging on to ThreadLocals
and doesn't renew/clear them.
There is work ongoing to improve this handing by Tomcat (see for instance: https://issues.apache.org/bugzilla/show_bug.cgi?id=49159),
but even on preliminary testing on latest Tomcat 7.0.12 shows its not "solved" AFAICT.

Anyway, common advise by most is these "SEVERE" errors can simply be ignored, except if they
can be easily fixed.
For the Jetspeed codebase, it might be possible to add additional request filtering to do
ThreadLocal "cleanup" ourselves (cumbersome as it is).
However it won't help us to get rid of all these errors as many are coming from external/third
party libraries which we have no runtime control over.

The easy and/or important causes already have been taken care of earlier by Randy, what remains
(from Jetspeed codebase) is imo non-critical.
So bottom line, my advise is to just follow the common sense and indeed ignore/accept these
messages as is.
Hopefully a future version of Tomcat will be better capable of cleaning up their own act in
this regard...

I'm going to downgrade this issue to Minor priority and postpone to a later, undefined, version
for renewed evaluation.

If anyone else thinks this should be re-prioritized again, be my guest to pick it up.

> Tomcat 6.0.24+ logs SEVERE errors on the console during shutdown from WebappClassLoader
clearReferencesThreads and clearThreadLocalMap 
> ---------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: JS2-1183
>                 URL: https://issues.apache.org/jira/browse/JS2-1183
>             Project: Jetspeed 2
>          Issue Type: Task
>          Components: Container
>            Reporter: Ate Douma
>            Priority: Minor
>         Attachments: catalina.out
>
>
> Since Tomcat 6.0.24 and higher Tomcat now actively verifies potential "shutdown" problems
concerning application created thread, ThreadLocal instances etc.
> If it encounters such issues, it reports this as SEVERE errors, which by default end
up in the console output or catalina.out.
> While these reporting "issues" indeed need looking into, the Console logging is rather
annoying.
> For the Jetspeed 2.2.1 installer which will bundle Tomcat 6.0.26, I'll add an override
of the Tomcat logging.properties to redirect these errors to the localhost FileHandler instead
(logs/localhost.{date}.log).
> For Jetspeed 2.2.2 the reported "issues" warrant further investigation to see if these
need to be solved or not.

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

---------------------------------------------------------------------
To unsubscribe, e-mail: jetspeed-dev-unsubscribe@portals.apache.org
For additional commands, e-mail: jetspeed-dev-help@portals.apache.org


Mime
View raw message