Add a new variable VerifyRecoveryLock which can be used to disable the test that...

diff --git a/include/ctdb_private.h b/include/ctdb_private.h
index df194bcb7eb052e08b7037a89893ee1e1d106b15..5121cd96d70ecd17947c691cec90d30237e2c64c 100644 (file)
--- a/include/ctdb_private.h
+++ b/include/ctdb_private.h
@@ -118,6 +118,7 @@ struct ctdb_tunable {
        uint32_t recd_ping_failcount;
        uint32_t log_latency_ms;
        uint32_t recovery_drop_all_ips;
+       uint32_t verify_recovery_lock;
 };
 
 /*

diff --git a/server/ctdb_recover.c b/server/ctdb_recover.c
index 153f698c7c5e5a385c752943c2c62e9bbd69244e..c2807b4d65e0c5520705d004d0ad0561f3c7ac89 100644 (file)
--- a/server/ctdb_recover.c
+++ b/server/ctdb_recover.c
@@ -505,7 +505,7 @@ static void ctdb_set_recmode_timeout(struct event_context *ev, struct timed_even
           caused by the cluster filesystem being very slow to
           arbitrate locks immediately after a node failure.       
         */
-       DEBUG(DEBUG_NOTICE,(__location__ " set_recmode timeout - allowing recmode set\n"));
+       DEBUG(DEBUG_ERR,(__location__ " set_recmode child process hung/timedout CFS slow to grant locks? (allowing recmode set anyway)\n"));
        state->ctdb->recovery_mode = state->recmode;
        ctdb_request_control_reply(state->ctdb, state->c, NULL, 0, NULL);
        talloc_free(state);
@@ -628,11 +628,17 @@ int32_t ctdb_control_set_recmode(struct ctdb_context *ctdb,
        state = talloc(ctdb, struct ctdb_set_recmode_state);
        CTDB_NO_MEMORY(ctdb, state);
 
+
+       if (ctdb->tunable.verify_recovery_lock == 0) {
+               /* dont need to verify the reclock file */
+               ctdb->recovery_mode = recmode;
+               return 0;
+       }
+
        /* For the rest of what needs to be done, we need to do this in
           a child process since 
           1, the call to ctdb_recovery_lock() can block if the cluster
              filesystem is in the process of recovery.
-          2, running of the script may take a while.
        */
        ret = pipe(state->fd);
        if (ret != 0) {
@@ -653,7 +659,7 @@ int32_t ctdb_control_set_recmode(struct ctdb_context *ctdb,
                char cc = 0;
                close(state->fd[0]);
 
-               /* we should not be able to get the lock on the nodes list, 
+               /* we should not be able to get the lock on the reclock file, 
                  as it should  be held by the recovery master 
                */
                if (ctdb_recovery_lock(ctdb, false)) {
@@ -665,6 +671,7 @@ int32_t ctdb_control_set_recmode(struct ctdb_context *ctdb,
                /* make sure we die when our parent dies */
                while (kill(parent, 0) == 0 || errno != ESRCH) {
                        sleep(5);
+                       write(state->fd[1], &cc, 1);
                }
                _exit(0);
        }
@@ -672,7 +679,7 @@ int32_t ctdb_control_set_recmode(struct ctdb_context *ctdb,
 
        talloc_set_destructor(state, set_recmode_destructor);
 
-       state->te = event_add_timed(ctdb->ev, state, timeval_current_ofs(3, 0),
+       state->te = event_add_timed(ctdb->ev, state, timeval_current_ofs(15, 0),
                                    ctdb_set_recmode_timeout, state);
 
        state->fde = event_add_fd(ctdb->ev, state, state->fd[0],

diff --git a/server/ctdb_tunables.c b/server/ctdb_tunables.c
index fe6a4b412d4ee97459f4048ff35b44758756bf0d..821236f93704c10df050d2ba92ed069ac5ec08e5 100644 (file)
--- a/server/ctdb_tunables.c
+++ b/server/ctdb_tunables.c
@@ -54,6 +54,7 @@ static const struct {
        { "RecdFailCount",        3,  offsetof(struct ctdb_tunable, recd_ping_failcount) },
        { "LogLatencyMs",         0,  offsetof(struct ctdb_tunable, log_latency_ms) },
        { "RecoveryDropAllIPs",  60,  offsetof(struct ctdb_tunable, recovery_drop_all_ips) },
+       { "VerifyRecoveryLock",   1,  offsetof(struct ctdb_tunable, verify_recovery_lock) },
 };
 
 /*