ctdb-client: If g_lock lock conflicts, try again sooner
authorAmitay Isaacs <amitay@gmail.com>
Tue, 19 Apr 2016 07:37:46 +0000 (17:37 +1000)
committerMartin Schwenke <martins@samba.org>
Tue, 5 Jul 2016 08:53:14 +0000 (10:53 +0200)
Instead of delaying for 1 second, try to get g_lock lock again after
1 milli-second.

Signed-off-by: Amitay Isaacs <amitay@gmail.com>
Reviewed-by: Martin Schwenke <martin@meltin.net>
ctdb/client/client_db.c

index 09d9f23ecac5386c8749297aa20748db279ba690..6a21e0b2dcab3a4db18e20edde2960891bf76e30 100644 (file)
@@ -1366,7 +1366,7 @@ static void ctdb_g_lock_lock_checked(struct tevent_req *subreq)
        if (val == 1) {
                /* server process exists, need to retry */
                subreq = tevent_wakeup_send(state, state->ev,
-                                           tevent_timeval_current_ofs(1,0));
+                                           tevent_timeval_current_ofs(0,1000));
                if (tevent_req_nomem(subreq, req)) {
                        return;
                }