Re: Failed to contact xxxx-1085, ellpack

On Oct 3, 2005, at 9:16 AM, Sacha Mallais wrote:

> On Oct 3, 2005, at 5:44 AM, John Pollard wrote:
>
>
>> Since we moved to Tiger (client) for hosting WO apps, I sometimes  
>> have trouble with JavaMonitor not being able to report the status  
>> of applications.
>>
>> Rebooting sorts this out.
>>
>> The message on the application detail page in JavaMonitor in red  
>> letters is "Failed to contact plug-1085" and no statistics are shown.
>>
>> Yet the applications themselves all appear to be functioning  
>> correctly.
>>
>> Each time JavaMonitor fails to get the info it needs,  
>> webobjects.log has the following line added to it:
>>
>> [2005-10-03 13:23:39 BST] <WorkerThread7> <WOWorkerThread id=7  
>> socket=Socket[addr=/192.168.0.34,port=59254,localport=1085]>  
>> Exception while sending response: java.net.SocketException: Broken  
>> pipe
>>

This looks to be a message from wotaskd, probably when returning the  
configuration information to JavaMonitor.  It looks like something is  
occupying wotaskd so that it is responding too slowly to requests.   
The next time this happens, get some threaddumps from wotaskd (check  
beforehand to ensure that it is not configured to send this to /dev/ 
null, I suspect that is the default).  Take them several seconds / a  
minute apart and see if that shows anything.  Or you could get really  
ambitious and run wotaskd in a Java debugger and attach to it when  
this happens and see where the holdup is.

And while we are grasping at straws, add "- 
D_DeploymentDebugging=true" when starting wotaskd or JavaMonitor  
makes the tools very talkative.  They may say something interesting.


Chuck


>> Running lsof -i tcp:2001-2050 gives many lines with something  
>> called ellpack in:
>>
>> java      628 root  259u  IPv6 0x04d1d490      0t0  TCP  
>> plug:ellpack->plug:63305 (CLOSE_WAIT)
>> java      628 root  260u  IPv6 0x04d1d100      0t0  TCP  
>> plug:ellpack->plug:63318 (CLOSE_WAIT)
>> java      628 root  261u  IPv6 0x04d1c0f8      0t0  TCP  
>> plug:ellpack->plug:63333 (CLOSE_WAIT)
>> java      628 root  262u  IPv6 0x04d1bf30      0t0  TCP  
>> plug:ellpack->plug:63337 (CLOSE_WAIT)
>> java      628 root  263u  IPv6 0x04d1d9e8      0t0  TCP  
>> plug:ellpack->plug:63384 (CLOSE_WAIT)
>> ...
>>
>> and I am guessing from past experience that these are causing the  
>> problem.
>>
>> What is this all about and how can it be prevented? I can't find  
>> information on this ellpack thing.
>>
>
> ellpack is a wild goose chase: it refers to the service that  
> usually occurs on that port (port 2025 -- see /etc/services).
>
> As to what the actual problem is, I don't know: I'm seeing similar  
> things on our site.  What seems to be happening to me is one  
> instance goes south and after that point all communication with  
> wotaskd is lost.  That causes the "Failed to contact..." message.
>

-- 
Coming in 2006 - an introduction to web applications using WebObjects  
and Xcode     http://www.global-village.net/wointro

Practical WebObjects - for developers who want to increase their  
overall knowledge of WebObjects or who are trying to solve specific  
problems.    http://www.global-village.net/products/practical_webobjects