io: Do not use sys_write to write to client sockets

When sending messages to clients, ctdb checks for EAGAIN error code and
schedules next write in the subsequent event loop.  Using sys_write in
these places causes ctdb to loop hard till a client is able to read from
the socket.  With real time scheduling, ctdb daemon spins consuming 100%
of CPU trying to write to the client sockets.  This can be quite harmful
when running under VMs or machines with single CPU.

This regression was introduced when all read/write calls were replaced to
use sys_read/sys_write wrappers (c1558adeaa980fb4bd6177d36250ec8262e9b9fe).

The existing code backs off in case of EAGAIN failures and waits for an
event loop to process the write again.  This should give ctdb clients
a chance to get scheduled and to process the ctdb socket.

Signed-off-by: Amitay Isaacs <amitay@gmail.com>
Reviewed-by: Martin Schwenke <martin@meltin.net>
Autobuild-User(master): Martin Schwenke <martins@samba.org>
Autobuild-Date(master): Tue Feb 24 12:29:30 CET 2015 on sn-devel-104

(Imported from commit 04a061e4d19d5bdbd8179fb0fab8b0875eec243e)

diff --git a/common/ctdb_io.c b/common/ctdb_io.c
index 467ec9a0fa59172b50c718dd92d4010d36d5eea5..53486f43da63d568939c0c935ccac58e12ecf274 100644 (file)
--- a/common/ctdb_io.c
+++ b/common/ctdb_io.c
@@ -232,9 +232,9 @@ static void queue_io_write(struct ctdb_queue *queue)
                struct ctdb_queue_pkt *pkt = queue->out_queue;
                ssize_t n;
                if (queue->ctdb->flags & CTDB_FLAG_TORTURE) {
-                       n = sys_write(queue->fd, pkt->data, 1);
+                       n = write(queue->fd, pkt->data, 1);
                } else {
-                       n = sys_write(queue->fd, pkt->data, pkt->length);
+                       n = write(queue->fd, pkt->data, pkt->length);
                }
 
                if (n == -1 && errno != EAGAIN && errno != EWOULDBLOCK) {
@@ -310,7 +310,7 @@ int ctdb_queue_send(struct ctdb_queue *queue, uint8_t *data, uint32_t length)
           queue overhead. This relies on non-blocking sockets */
        if (queue->out_queue == NULL && queue->fd != -1 &&
            !(queue->ctdb->flags & CTDB_FLAG_TORTURE)) {
-               ssize_t n = sys_write(queue->fd, data, length2);
+               ssize_t n = write(queue->fd, data, length2);
                if (n == -1 && errno != EAGAIN && errno != EWOULDBLOCK) {
                        talloc_free(queue->fde);
                        queue->fde = NULL;