Diskless replication: rio fdset target new supports buffering.

To perform a socket write() for each RDB rio API write call was
extremely unefficient, so now rio has minimal buffering capabilities.
Writes are accumulated into a buffer and only when a given limit is
reacehd are actually wrote to the N slaves FDs.

Trivia: rio lacked support for buffering since our targets were:

1) Memory buffers.
2) C standard I/O.

Both were buffered already.

1 files changed, 3 insertions, 0 deletions

diff --git a/src/rdb.c b/src/rdb.c
index 45beae14d..c6a1ec691 100644
--- a/src/rdb.c
+++ b/src/rdb.c
@@ -1425,6 +1425,9 @@ int rdbSaveToSlavesSockets(void) {
         redisSetProcTitle("redis-rdb-to-slaves");
 
         retval = rdbSaveRioWithEOFMark(&slave_sockets,NULL);
+        if (retval == REDIS_OK && rioFlush(&slave_sockets) == 0)
+            retval = REDIS_ERR;
+
         if (retval == REDIS_OK) {
             size_t private_dirty = zmalloc_get_private_dirty();