messaging: Fix a memleak with clustering

We have to properly throw away unexpected messages that came in via ctdb

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

diff --git a/source3/lib/msg_channel.c b/source3/lib/msg_channel.c
index 8e23fd44c4512137340002ccc419713e29de5900..6be5e2ef1c6d5dd61bb98328375025d424521a24 100644 (file)
--- a/source3/lib/msg_channel.c
+++ b/source3/lib/msg_channel.c
@@ -310,6 +310,9 @@ static void msg_read_got_ctdb(struct tevent_req *subreq)
        /*
         * Got some unexpected msg type, wait for the next one
         */
+
+       TALLOC_FREE(state->rec);
+
        subreq = ctdb_msg_read_send(state, state->ev,
                                    state->channel->ctdb_channel);
        if (tevent_req_nomem(subreq, req)) {