On Jan 5, 2004, at 21:49, Nathaniel Talbott wrote:

> First of all, thanks so much to all those that have helped with 
> improving the state of --enable-pthread support. That said, I'm 
> encountering another issue...
>
> Every once in a while my app will hang for period (seems to average 
> around 30 minutes), and then it will continue as though nothing 
> happened. It's hard to catch it in the act, but last time I did, I 
> quickly found out from ps that it's hanging in:
>
>   WCHAN
>   wait_for_tcp_memory
>
> Attaching to the process with GDB seems to unhang it as soon as I 
> continue it, but I grabbed the following backtrace first that 
> hopefully is indicative of the actual place that it's hanging:

My process has just hung again at this point, and this time attaching 
with GDB did not fix the problem. Thus I can confirm that the earlier 
backtrace I sent was indicative of the hang point.


> Unfortunately, I have no clue as to how to duplicate the problem in a 
> smaller example :-(. The hang is so infrequent (I've caught it three 
> times in three days of running; I need to go back to my logs and see 
> if it's happened more than that) that it's hard to know where to 
> begin. Any help would, of course, be appreciated.

I still haven't been able to boil down the problem to something smaller 
:-(. Any ideas, either on fixing the problem or even just on 
identifying it? While I can tolerate this behavior for a short time, I 
have to eliminate it for the long-term.

Thanks,


Nathaniel

<:((><