If you are connected to corosync and registered for object notifications then corosync is asked to shutdown the IPC server will get stuck. This is because the pipe is closed and the refcount is increased. This leaves ipcs with a connection that it can't destroy.
Solution: 1) if a write to the pipe fails (pipe closed) decrement the refcounter. 2) fix the object_track_stop() - it was not working as the functions did not match up. (this caused the late callbacks). 3) in ipcs call exit_fn() then stats_destroy_connection() so that the service engine can have time to call object_track_stop() before the object gets destroyed. Signed-off-by: Angus Salkeld <[email protected]> --- exec/coroipcs.c | 2 +- services/confdb.c | 3 ++- 2 files changed, 3 insertions(+), 2 deletions(-) diff --git a/exec/coroipcs.c b/exec/coroipcs.c index 29655ba..dbafb18 100644 --- a/exec/coroipcs.c +++ b/exec/coroipcs.c @@ -527,8 +527,8 @@ static inline int conn_info_destroy (struct conn_info *conn_info) * Retry library exit function if busy */ if (conn_info->state == CONN_STATE_THREAD_DESTROYED) { - api->stats_destroy_connection (conn_info->stats_handle); res = api->exit_fn_get (conn_info->service) (conn_info); + api->stats_destroy_connection (conn_info->stats_handle); if (res == -1) { api->serialize_unlock (); return (0); diff --git a/services/confdb.c b/services/confdb.c index 3187718..b57a041 100644 --- a/services/confdb.c +++ b/services/confdb.c @@ -348,7 +348,7 @@ static int confdb_lib_exit_fn (void *conn) api->object_track_stop(confdb_notify_lib_of_key_change, confdb_notify_lib_of_new_object, confdb_notify_lib_of_destroyed_object, - NULL, + confdb_notify_lib_of_reload, conn); return (0); } @@ -857,6 +857,7 @@ retry_write: if (written == sizeof(struct confdb_ipc_message_holder)) { return 0; } else { + api->ipc_refcnt_dec(conn); return -1; } } -- 1.7.4 _______________________________________________ Openais mailing list [email protected] https://lists.linux-foundation.org/mailman/listinfo/openais
