On Mon, 17 Mar 2003, Robert L. Millner wrote:
> So, I've been going around the hallways lately, pointing at people and
> saying "Hey! Futex buddy!"
>
> Doing distribution testing when the rpm command stalls on the futex
> syscall every 1/3 to 1/4 times you use it where a reboot is apparently
> needed to make rpm work again at all is really, really irritating.
It could be waiting for a (possibly stale) lock. Make sure no rpm related
processes are running, then delete /var/lib/rpm/__db* .
> ...
> [ The symptoms of needing a reboot to clear it appears to me to be a bug
> with the kernel's handling of thread lock contention but I haven't
> really dug into it yet and could be completely wrong. ]
That doesn't prove it is a kernel bug, because rebooting also clears rpm's
lock files.
Michael Young
--
Phoebe-list mailing list
[EMAIL PROTECTED]
https://listman.redhat.com/mailman/listinfo/phoebe-list
