Ok, I take this bug back (at least how I originally thought of it). Given a node that I kill -9'ed. On restart, the recover step is correctly setting the start position automatically:
130302 06:48:55 mysqld_safe Starting mysqld daemon with databases from /var/lib/mysql
130302 06:48:55 mysqld_safe WSREP: Running position recovery with --log_error=/tmp/tmp.mwI4iVaIvz
130302 06:49:01 mysqld_safe WSREP: Recovered position 8797f811-7f73-11e2-0800-8b513b3819c1:314673
130302 6:49:01 [Note] WSREP: wsrep_start_position var submitted: '8797f811-7f73-11e2-0800-8b513b3819c1:314673'
130302 6:49:01 [Note] WSREP: Found saved state: 8797f811-7f73-11e2-0800-8b513b3819c1:-1
130302 6:49:01 [Note] WSREP: Start replication130302 6:49:01 [Note] WSREP: Setting initial position to 8797f811-7f73-11e2-0800-8b513b3819c1:314673
Ok, I take this bug back (at least how I originally thought of it). Given a node that I kill -9'ed. On restart, the recover step is correctly setting the start position automatically:
130302 06:48:55 mysqld_safe Starting mysqld daemon with databases from /var/lib/mysql /tmp/tmp. mwI4iVaIvz 7f73-11e2- 0800-8b513b3819 c1:314673 position var submitted: '8797f811- 7f73-11e2- 0800-8b513b3819 c1:314673'
130302 06:48:55 mysqld_safe WSREP: Running position recovery with --log_error=
130302 06:49:01 mysqld_safe WSREP: Recovered position 8797f811-
130302 6:49:01 [Note] WSREP: wsrep_start_
130302 6:49:01 [Note] WSREP: Found saved state: 8797f811- 7f73-11e2- 0800-8b513b3819 c1:-1
130302 6:49:01 [Note] WSREP: Start replication130302 6:49:01 [Note] WSREP: Setting initial position to 8797f811- 7f73-11e2- 0800-8b513b3819 c1:314673