Concurrency Bug in memcached

Tomash Brechko tomash.brechko at
Wed Feb 6 07:13:02 UTC 2008

On Tue, Feb 05, 2008 at 14:34:36 -0800, Brian Aker wrote:
> From your argument I take it that you would go with number 2? I  
> thought it was sloppy to not catch the failure, which is why I added  
> the loop, but I agree with you that it should be fine to miss a clock  
> count.

I actually didn't get your description of the problem altogether,
sorry.  So I would go with "don't change anything" this time ;).
Maybe someone else will be able to comment...

   Tomash Brechko

More information about the memcached mailing list