> > What do you think of commit e177294f5094 ? I think we should > > revert that. doLock() already handles this. > > That patch drops the timeout down a lot when we can't lock, and we > can't see the lock owner (presumably because it's exiting/exited). > I guess this is triggered by zombies?
There's a minimal delay between doLock() checking if the lock owner exists and running get_process_info() if it fails. An exited owner is handled in doLock() by NOT failing, so we don't run get_process_info then, and do not sleep() at all. -- Zdenek _______________________________________________ Yum-devel mailing list Yum-devel@lists.baseurl.org http://lists.baseurl.org/mailman/listinfo/yum-devel
