Unable to stop a jail

[Robert Watson]

On Fri, 1 Dec 2006, Steven Hartland wrote:

>> In essence, this would move to having two reference counts on the prison: a 
>> "strong" reference that has to do with having process members, and a "weak" 
>> reference that has to do with ucreds pointing at the prison.
>
> The proposal sounds like a good idea but I'm sure there's an argument that 
> would say thats just hiding the real underlieing issue?

Well, there are two things going on here:

(1) Jails that last a long time due to being referenced by data structures
     that last a long time.  I.e., time-wait TCP connections.

(2) Leaks in credentials or jails resulting in jails that never go away.

What I describe is intended to address the former issue, which is one that 
exists for a reason.  The latter issues are clearly bugs and just need to be 
fixed.

Robert N M Watson
Computer Laboratory
University of Cambridge