Bug: 147157

Fix Description: There was already a flag to disable the tombstone reap
thread, but there was no way to set it from outside the replica code.
I added some functions to set the flag, and to query it to show it
in the status entry.  When a total update request comes in, the flag
is set to disable, and when the total update ends for whatever reason
(success, error, connection terminated), the flag is set back to
enable.
In addition, the tombstone reap thread was being started after 1 hour -
this is too long and needed to be shorter.  We already have the
tombstone reap interval, so I changed it so that the reap thread
starts after this interval, then runs after every interval elapses.

1 files changed, 3 insertions, 0 deletions

diff --git a/ldap/servers/plugins/replication/repl_connext.c b/ldap/servers/plugins/replication/repl_connext.c
index 8b0c0551..9bda7ca4 100644
--- a/ldap/servers/plugins/replication/repl_connext.c
+++ b/ldap/servers/plugins/replication/repl_connext.c
@@ -75,6 +75,9 @@ void consumer_connection_extension_destructor (void *ext, void *object, void *pa
 						"of replicated area.\n");
 				}
 				slapi_pblock_destroy(pb);
+
+				/* allow reaping again */
+				replica_set_tombstone_reap_stop(r, PR_FALSE);
 			}
 			replica_relinquish_exclusive_access(r, connid, -1);
             object_release ((Object*)connext->replica_acquired);