"Ed L." <pgsql@xxxxxxxxxxxxx> writes: > Also seeing lots of postmaster zombies (190 and growing)... Yeah, that and the lack of service for new connections would both be expected if the postmaster is stuck. And the autovac worker start failures, too. There's only one bug here. > While new connections are hanging, top shows postmaster using > 100% of cpu. Oh, for some reason I thought it was sitting idle. That sounds more like an infinite loop. Try reattaching to the postmaster, confirm the stack trace, and then see how many times you can do "fin" before it doesn't return control. That will tell us which level of subroutine is looping. regards, tom lane -- Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general