Tom Evans [tom@23palmer.net] wrote:
>
> Now that we're all ramping up Alpha support :-)
>
> Does anyone have any linuxthreads experience. I've been trying to track
> down
> the thread "zombie" issue on my own, but I'm no libc developer.
>
> I'm somewhat lost in my debugging - I keep getting the impression that
> certain
> exiting pthreads never get their resources freed by the thread manager
> (obvious, huh).
> However, some threads do seem to do the "right" thing and have their
> resources freed,
> so there isn't a 1-to-1 correlation between threads threads that have exited
> and zombies.
>
> I'm wondering if somehow a thread that implicitly exits (via a return) isn't
> hitting the
> "pthread_exit" code whereas one that explicitly calls it does.
>
> Has anyone else been looking at this?
I have not been looking at it, but I have certainly noticed it. I have
thousands of defunct processes from clamav-milter and mysqld...
So this is a recent libc problem? It should be possible to track it
down by looking through the relevant libc changes. The problem appeared
(for me) only a few months ago.
--
Cheers,
Bob McElrath [Univ. of California at Davis, Department of Physics]
"One of the best ways to get yourself a reputation as a dangerous citizen
these days is to go about repeating the very phrases which our founding
fathers used in the great struggle for independence." --Charles A. Beard
Attachment:
signature.asc
Description: Digital signature