top posting my self to clarify, since the title is in fact inverted: close might be more expensive the open. On 03/03/2011 08:28 AM, Rob Sargent wrote: > A developer here accidentally flooded a server with connection opens and > closes, essentially one per transaction during a multi-threaded data > migration process. > > We were curious if this suggests that connection clean up is more > expensive than creation thereby exhausting resources, or if perhaps the > server wasn't returning to the essentially backgrounded clean-up task > adroitly (enough)? > > Don't worry we are now using a connection pool for the migration, but > the situation was, um, er, entertaining. The server side said > eof-from-client, the client side said many variations of cannot-connect, > then they would work out their differences briefly and do the same dance > over again after sufficient connections. > > Cheers, > rjs > > > > -- Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
