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
Branch
------
master
Details
-------
http://git.postgresql.org/pg/commitdiff/754baa21f723255272c24dc5f9ab456858e361e3
Modified Files
--------------
doc/src/sgml/config.sgml | 28 +++
src/backend/libpq/pqcomm.c | 240 ++++++++++++++++++-------
src/backend/port/unix_latch.c | 26 ++-
src/backend/port/win32/socket.c | 13 ++-
src/backend/port/win32_latch.c | 30 +++-
src/backend/replication/walsender.c | 143 ++++++++++-----
src/backend/utils/misc/guc.c | 10 +
src/backend/utils/misc/postgresql.conf.sample | 1 +
src/include/libpq/libpq.h | 3 +
src/include/replication/walsender.h | 1 +
src/include/storage/latch.h | 2 +-
11 files changed, 368 insertions(+), 129 deletions(-)