if a lock wait child died/finished, we could have released the lockwait handle and...

The waiters reference the locakwait handle in order to remove itself from the li
nked list which caused a SEGV.

We dont actually need to remove ourselves from this list here since
if the parent freeze_handle holding the list is freed, then all waiters are rele
ased as well, and the only place we actually need to relink the waiter is in ctd
b_freeze_lock_handler, where we want to respond back to the clients and release
the waiters  but we still want to keep the freeze_handle hanging around.

diff --git a/server/ctdb_freeze.c b/server/ctdb_freeze.c
index 69e70abc44a071ad358896322de439a93af831b2..36d033d49ba29b4f2f86832f9596b67d9e291c60 100644 (file)
--- a/server/ctdb_freeze.c
+++ b/server/ctdb_freeze.c
@@ -151,9 +151,12 @@ static void ctdb_freeze_lock_handler(struct event_context *ev, struct fd_event *
        h->ctdb->freeze_mode[h->priority] = CTDB_FREEZE_FROZEN;
 
        /* notify the waiters */
-       while ((w = h->ctdb->freeze_handles[h->priority]->waiters)) {
+       if (h != h->ctdb->freeze_handles[h->priority]) {
+               DEBUG(DEBUG_ERR,("lockwait finished but h is not linked\n"));
+       }
+       while ((w = h->waiters)) {
                w->status = status;
-               DLIST_REMOVE(h->ctdb->freeze_handles[h->priority]->waiters, w);
+               DLIST_REMOVE(h->waiters, w);
                talloc_free(w);
        }
 }
@@ -241,7 +244,6 @@ static struct ctdb_freeze_handle *ctdb_freeze_lock(struct ctdb_context *ctdb, ui
  */
 static int ctdb_freeze_waiter_destructor(struct ctdb_freeze_waiter *w)
 {
-       DLIST_REMOVE(w->ctdb->freeze_handles[w->priority]->waiters, w);
        ctdb_request_control_reply(w->ctdb, w->c, NULL, w->status, NULL);
        return 0;
 }
@@ -267,7 +269,7 @@ int ctdb_start_freeze(struct ctdb_context *ctdb, uint32_t priority)
        }
 
        /* if there isn't a freeze lock child then create one */
-       if (!ctdb->freeze_handles[priority]) {
+       if (ctdb->freeze_handles[priority] == NULL) {
                ctdb->freeze_handles[priority] = ctdb_freeze_lock(ctdb, priority);
                CTDB_NO_MEMORY(ctdb, ctdb->freeze_handles[priority]);
                ctdb->freeze_mode[priority] = CTDB_FREEZE_PENDING;
@@ -309,6 +311,11 @@ int32_t ctdb_control_freeze(struct ctdb_context *ctdb, struct ctdb_req_control *
        }
 
        /* add ourselves to list of waiters */
+       if (ctdb->freeze_handles[priority] == NULL) {
+               DEBUG(DEBUG_ERR,("No freeze lock handle when adding a waiter\n"));
+               return -1;
+       }
+
        w = talloc(ctdb->freeze_handles[priority], struct ctdb_freeze_waiter);
        CTDB_NO_MEMORY(ctdb, w);
        w->ctdb     = ctdb;
@@ -376,8 +383,10 @@ static void thaw_priority(struct ctdb_context *ctdb, uint32_t priority)
        system("mkdir -p test.db.saved; /usr/bin/rsync --delete -a test.db/ test.db.saved/$$ 2>&1 > /dev/null");
 #endif
 
-       talloc_free(ctdb->freeze_handles[priority]);
-       ctdb->freeze_handles[priority] = NULL;
+       if (ctdb->freeze_handles[priority] != NULL) {
+               talloc_free(ctdb->freeze_handles[priority]);
+               ctdb->freeze_handles[priority] = NULL;
+       }
 }
 
 /*

diff --git a/server/ctdb_monitor.c b/server/ctdb_monitor.c
index 5e1f7ad12c7f1ce5eaf18ffb3be3e2f630cc9e69..056d8311150bdcb7f0fe17bbe8ec7b7a94e5faa7 100644 (file)
--- a/server/ctdb_monitor.c
+++ b/server/ctdb_monitor.c
@@ -235,7 +235,7 @@ static void ctdb_check_health(struct event_context *ev, struct timed_event *te,
                        DEBUG(DEBUG_ERR,("Skip monitoring during recovery\n"));
                }
                for (i=1; i<=NUM_DB_PRIORITIES; i++) {
-                       if (ctdb->freeze_handles[i] != 0) {
+                       if (ctdb->freeze_handles[i] != NULL) {
                                DEBUG(DEBUG_ERR,("Skip monitoring since databases are frozen\n"));
                                skip_monitoring = 1;
                                break;