Automatically terminate replication connections that are idle for more
Heikki Linnakangas <heikki.linnakangas@iki.fi>
Automatically terminate replication connections that are idle for more than replication_timeout (a new GUC) milliseconds. The TCP timeout is often too long, you want the master to notice a dead connection much sooner. People complained about that in 9.0 too, but with synchronous replication it's even more important to notice dead connections promptly. Fujii Masao and Heikki Linnakangas
Files
| Path | Change | +/− |
|---|---|---|
| doc/src/sgml/config.sgml | modified | +28 −0 |
| src/backend/libpq/pqcomm.c | modified | +171 −69 |
| src/backend/port/unix_latch.c | modified | +19 −7 |
| src/backend/port/win32_latch.c | modified | +24 −6 |
| src/backend/port/win32/socket.c | modified | +12 −1 |
| src/backend/replication/walsender.c | modified | +98 −45 |
| src/backend/utils/misc/guc.c | modified | +10 −0 |
| src/backend/utils/misc/postgresql.conf.sample | modified | +1 −0 |
| src/include/libpq/libpq.h | modified | +3 −0 |
| src/include/replication/walsender.h | modified | +1 −0 |
| src/include/storage/latch.h | modified | +1 −1 |