On Sat, Dec 17, 2011 at 11:58 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> I think this argument is bogus: if this is a real issue, then no use of
> kill() anytime, by anyone, is safe. In practice I believe that Unix
> systems avoid recycling PIDs right away so as to offer some protection.
I'm not sure they do anything more sophisticated than cycling through
a sufficiently-large PID space, but whether it's that or something
else, I guess it must be adequate or they'd have enlarged the space...
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company