We were bitten by the following bug a few times, when our server tried
to reestablish connections under bad network conditions:
if connection is closed while trying to get response to SSL setup packet
(i.e. conn->status is CONNECTION_SSL_STARTUP), we get a busy loop, as
line 1035 in 8.0.0.beta2:
if (pqWaitTimed(1, 0, conn, finish_time) {
tells that there is data to read (returns 0) while actually it is error
(POLLERR & POLLHUP) and not POLLIN returned from poll() and
nread = recv(conn->sock, &SSLok, 1, 0);
on line 1449 sets nread to 0, as there is in fact nothing to read from a
closed socket
after that the check on line 1462:
if (nread == 0)
/* caller failed to wait for data */
return PGRES_POLLING_READING;
resumes the busy loop
in all other places recv()'ing is done throught pqsecure_read, which
checks for such condition.
I'm not familiar enough with code to suggest a proper fix (and I
quickfix our app by disabling SSL (PGSSLMODE=disable) - the default was
prefer ;), but this may come back to haunt others
--------------
Hannu
---------------------------(end of broadcast)---------------------------
TIP 4: Don't 'kill -9' the postmaster
