Commit 8fade5fa authored by unknown's avatar unknown

Can't believe it: one more cset (04-27) which never found his way to the central

repository (incredible that I forgot to push, but why not).
So unfortunately the bugfix missed 3.23.57 and will be in .58 :(
Instead of looking like working (bug #198), replication between
a 3.23 slave and 4.0 master should frankly stop. Here we detect
4.0 masters in the 3.23 slave code when we see a strange Rotate
event, and in that case we print an error and stop.
4.0.13 and older masters will be "often" caught (see the patch); 4.0.14
and newer masters will always be immediately caught.


sql/slave.cc:
  Instead of looking like working (bug #198), replication between
  a 3.23 slave and 4.0 master should frankly stop. Here we detect
  4.0 masters in the 3.23 slave code when we see a strange Rotate
  event, and in that case we print an error and stop.
  4.0.13 and older masters will be "often" caught (see the patch); 4.0.14
  and newer masters will always be immediately caught.
parent a7792d57
......@@ -1216,6 +1216,28 @@ static int exec_event(THD* thd, NET* net, MASTER_INFO* mi, int event_len)
Rotate_log_event* rev = (Rotate_log_event*)ev;
int ident_len = rev->ident_len;
pthread_mutex_lock(&mi->lock);
/*
If the master is 4.0, he has an incompatible binlog format, which we
cannot read. We always can detect this in _fake_ Rotate events, where we
see zeros at the beginning of the expected binlog's filename (this is
because 4.0 _fake_ Rotate events have zeros in the 'log_pos' and 'flags'
parts of the event header).
Consider the following test as a sanity check, which must always pass,
and which has the good side effect of catching 4.0 masters.
Masters >= 4.0.14 will always send a fake Rotate event (even if the slave
asked for a position >4) so we are 100% sure to catch the problem.
*/
if (!ident_len || !(rev->new_log_ident[0]))
{
sql_print_error("Slave: could not parse Rotate event; it *might* be \
that your master's version is 4.0 or newer, which cannot be replicated by \
3.23 slaves (in that case you need to upgrade your slave to 4.0 or newer)");
delete ev;
pthread_cond_broadcast(&mi->cond);
pthread_mutex_unlock(&mi->lock);
return 1;
}
memcpy(mi->log_file_name, rev->new_log_ident,ident_len );
mi->log_file_name[ident_len] = 0;
mi->pos = 4; // skip magic number
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment