[Wrapper-user] Re: Other Processes affecting Servicewrapper

[Leif M. <le...@ta...>]

Arun,
What version of the Wrapper are you using? Newer versions of the
Wrapper are able to detect that the CPU is at 100%. Rather than timing
out and restarting the JVM, you will get a message like the following:

jvm 1 | JVM Process has not received any CPU time for 11798 seconds.
Extending timeouts.
wrapper | Wrapper Process has not received any CPU time for 11800
seconds. Extending timeouts.

The case above was when I had suspended my system for a few hours.
The Wrapper was able to recover without any problems. The warning message
is just that, a warning. If another process takes up 100% of the CPU you
may see one or both of the above messages. But hopefully it would be for
a period of time much shorter than the 3 hours above. :-)

Were you seeing timeouts and JVM restarts or had you just noticed the
note in Troubleshooting guide? I'll take a look at that entry and bring
it up
to date.

Cheers,
Leif

Ryan Shaw wrote:

>On Thu, 2003-05-08 at 22:34, Arun Ramachandra wrote:
>
>  
>
>>I have seen in the Troubleshooting section of your website the 
>>following line:
>>
>>"the Wrapper will think that the JVM is hung when it isn't if 
>>another process is taking 100% of the CPU for longer than 30 
>>seconds. This will result in an entry like the following in your 
>>log file, and the JVM being restarted:"
>>
>>My question is why is my process being penalised if the 
>>other process is taking up 100% of cpu time?
>>Secondly is there some way to avoid this?
>>    
>>
>
>If another process is taking 100% of CPU, then there is no
>CPU left for the JVM. The wrapper has no way of knowing that 
>the reason the JVM has not pinged it is that the JVM is being 
>starved of processor power. So, it makes the assumption that 
>the JVM has hung.
>
>The only way to avoid this is to not let other processes 
>dominate your server for long periods. If you are running
>on unix you could use "nice" to achieve this.
>
>  
>