Michael Widenius <michael.widenius@gmail.com> writes:



This is because of errorneous merge. The original code:

    if (waitee_buf_ptr) {
            lock_report_waiters_to_mysql(waitee_buf_ptr,
                                         start_mysql_thd,
                                         victim_trx_id);

The bb-10.2-jan code:

    if (victim_trx && waitee_buf_ptr) {
            lock_report_waiters_to_mysql(waitee_buf_ptr,
                                         start_mysql_thd,
                                         victim_trx->id);

So if victim_trx is NULL the waits are not reported to parallel replication
at all, causing the stalls and/or hangs. victim_trx is NULL unless InnoDB
itself detects a deadlock.