[Solved by workaround]: wo adaptor produces "broken pipe"s

Hi Will,

I was reading your post on ominlist and found that we're having exactly  
the same problem  
(http://www.omnigroup.com/mailman/archive/webobjects-admin/2004- 
December/003213.html) about the broken pipe. You talk about a  
workaround, is there any chance you can share that workaround?. I can't  
find what is causing the broken pipe. I'm using JavaPlaybackManager to  
load test my woapp and as soon as start to add load (or number of  
virtual users hitting the woapp) I start to see the broken pipe in the  
woapp-log but the fetch size is not too big to wait so long that the  
adaptor exceeds the receive time. I've changed:

"Load balancing and adaptor settings"
Retries 3
Dormant 3
Send timeout 60
Receive timeout 60
Connect timeout 60

and "Application settings"
Maximum adaptor threads 1024
Listen queue size 512

and the broken pipe still there. I have only one instance deployed to  
do the loadtest and I also reviewed some indexes in my DB.

Anybody care for a broken pipe?... thanks list!

-Carlos