You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@tomcat.apache.org by sysdba <sy...@ohiointernet.us> on 2005/04/15 15:48:40 UTC
Memory Leak Solved
We have struggled with a memory leak in 5.0.28/5.0.30 for months. There have
been many complaints about the necessity to restart Tomcat every couple days
due to Out of Memory errors, but no solutions that cured it. Well, the
suggestion to put the single line:
Introspector.flushCaches();
in the destroy method of a servlet in a redeployable app finally solves it. Our
Tomcat web server has now run for seven days without an OOM error. The amount
of time spent with the Optimizeit profiler trying to locate a nonexistent
memory leak in the application code cannot be overestimated. This discovery
should be bold red lettered in the docs.
Gary Harris
Re: RES: Memory Leak Solved
Posted by "Trond G. Ziarkowski" <tr...@gep-as.com>.
Hi,
here's the bugzilla:
http://issues.apache.org/bugzilla/show_bug.cgi?id=33368, and here's a
link to the archives where Robert Willie found this issue (read the
entire thread...):
http://www.mail-archive.com/tomcat-user@jakarta.apache.org/msg146409.html.
Trond
Paulo Alvim wrote:
>The "Introspector.flushCaches();" in the "contextDestroyed" wasn't enough
>for us...
>
>How should we use the 'swallowOutput'? We've tried this way without results:
>
><Context displayName="eCompany 3.0"
> docBase="/eprj"
> path="eprj.war"
> privileged="true" swallowOutput="false">
>(...)
>
>Trond,
>
>Do you have links to this "known issue"?
>
>-----Mensagem original-----
>De: Trond G. Ziarkowski [mailto:trond.ziarkowski@gep-as.com]
>Enviada em: sexta-feira, 15 de abril de 2005 11:08
>Para: Tomcat Users List
>Assunto: Re: Memory Leak Solved
>
>
>Hi,
>
>don't know if you are using it, but there's also a known issue with
>5.0.28 and 'swallowOutput' in the <Context> element. After I turned it
>off, I haven't gotten any out of memory errors...
>
>Trond
>
>sysdba wrote:
>
>
>
>>We have struggled with a memory leak in 5.0.28/5.0.30 for months. There
>>
>>
>have
>
>
>>been many complaints about the necessity to restart Tomcat every couple
>>
>>
>days
>
>
>>due to Out of Memory errors, but no solutions that cured it. Well, the
>>suggestion to put the single line:
>> Introspector.flushCaches();
>>in the destroy method of a servlet in a redeployable app finally solves it.
>>
>>
>Our
>
>
>>Tomcat web server has now run for seven days without an OOM error. The
>>
>>
>amount
>
>
>>of time spent with the Optimizeit profiler trying to locate a nonexistent
>>memory leak in the application code cannot be overestimated. This discovery
>>should be bold red lettered in the docs.
>>
>>Gary Harris
>>
>>
>>
>>
>
>
>---------------------------------------------------------------------
>To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
>For additional commands, e-mail: tomcat-user-help@jakarta.apache.org
>
>
>
>---------------------------------------------------------------------
>To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
>For additional commands, e-mail: tomcat-user-help@jakarta.apache.org
>
>
>
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: tomcat-user-help@jakarta.apache.org
RES: Memory Leak Solved
Posted by Paulo Alvim <al...@powerlogic.com.br>.
The "Introspector.flushCaches();" in the "contextDestroyed" wasn't enough
for us...
How should we use the 'swallowOutput'? We've tried this way without results:
<Context displayName="eCompany 3.0"
docBase="/eprj"
path="eprj.war"
privileged="true" swallowOutput="false">
(...)
Trond,
Do you have links to this "known issue"?
-----Mensagem original-----
De: Trond G. Ziarkowski [mailto:trond.ziarkowski@gep-as.com]
Enviada em: sexta-feira, 15 de abril de 2005 11:08
Para: Tomcat Users List
Assunto: Re: Memory Leak Solved
Hi,
don't know if you are using it, but there's also a known issue with
5.0.28 and 'swallowOutput' in the <Context> element. After I turned it
off, I haven't gotten any out of memory errors...
Trond
sysdba wrote:
>We have struggled with a memory leak in 5.0.28/5.0.30 for months. There
have
>been many complaints about the necessity to restart Tomcat every couple
days
>due to Out of Memory errors, but no solutions that cured it. Well, the
>suggestion to put the single line:
> Introspector.flushCaches();
>in the destroy method of a servlet in a redeployable app finally solves it.
Our
>Tomcat web server has now run for seven days without an OOM error. The
amount
>of time spent with the Optimizeit profiler trying to locate a nonexistent
>memory leak in the application code cannot be overestimated. This discovery
>should be bold red lettered in the docs.
>
>Gary Harris
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: tomcat-user-help@jakarta.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: tomcat-user-help@jakarta.apache.org
Re: Memory Leak Solved
Posted by "Trond G. Ziarkowski" <tr...@gep-as.com>.
Hi,
don't know if you are using it, but there's also a known issue with
5.0.28 and 'swallowOutput' in the <Context> element. After I turned it
off, I haven't gotten any out of memory errors...
Trond
sysdba wrote:
>We have struggled with a memory leak in 5.0.28/5.0.30 for months. There have
>been many complaints about the necessity to restart Tomcat every couple days
>due to Out of Memory errors, but no solutions that cured it. Well, the
>suggestion to put the single line:
> Introspector.flushCaches();
>in the destroy method of a servlet in a redeployable app finally solves it. Our
>Tomcat web server has now run for seven days without an OOM error. The amount
>of time spent with the Optimizeit profiler trying to locate a nonexistent
>memory leak in the application code cannot be overestimated. This discovery
>should be bold red lettered in the docs.
>
>Gary Harris
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: tomcat-user-help@jakarta.apache.org