[mod_python] ValueError: Failed to acquire global mutex lock

On Sun, 23 Nov 2003, Julian Ciccale wrote:

> Hi and thanks,

>   I got that problem solved but now it seems it's happening something
> weird. I compiled the sources from CVS (enabling the debugging lines on
> the locking calls). On a first request i get these lines:
>
> [Sun Nov 23 12:17:51 2003] [warn] _global_lock at index 30 from pid 21429
> [Sun Nov 23 12:17:51 2003] [warn] _global_lock DONE at index 30 from pid 21429
>
> , which seems to lock mutex at ix 30. (and remains locked even when the
> request has completed). Then when a second request comes i get this:
>
> [Sun Nov 23 12:18:02 2003] [warn] _global_lock at index 30 from pid 21425
>
> . Obviously that process remains locked forever, waiting on a lock forever.

How are you using your sessions? Do you by any chance assign it to a
global variable - the session releases the lock when you call it's ulock()
method explicitely, or when it goes out of scope, so assigning it to a
global would make it stay there and locked even after the request is done.

Grisha