[Samba] Re: TDB locking overhead and performance...

[Jeremy Allison]

On Sun, Apr 23, 2006 at 03:32:56PM +0100, Daniel J Blueman wrote:
> 
> Does anyone know if sysV IPC semaphores have any mechanism for recovering locks?
> 
> One mechanism to handle the process-crashing-holding-lock corner case
> is to have lock acquisition timeout after eg 10s, and check if the pid
> of the locker is still there - if not, the lock is dropped by 1.
> 
> Any comments on why this is a bad/unworkable idea with eg sysV semaphores?

sysV semaphores are unusable for our case as most kernels have
to pre-allocate a known number. We have to scale up without
such artificial restrictions, and fcntl locks (usually) do
this (I'm only thinking of one kernel with pre-allocation
semantics for fcntl locks, hi there HPUX.... :-).

Jeremy.