I had this same problem. This isn't fool proof or recomended but what
i ended up doing was calling mysqlbinlog over and over again
incrementing --position (or --offset i can't remember) until it picked
up a good record again and kept going. Chances are you will lose
queries but it does work.
On Mon, 28 Jun 2004 13:51:00 -0700, Bill Earl <bearl@stripped> wrote:
> Hello everyone,
> We're having a problem with replication. The servers are all running 4.0.18 under
> SuSE Linux 9.1 Professional.
> I noticed that the two slaves were both showing a status of I/O thread not running.
> In the mysqld.log file for both slaves I found the following entry:
> 040628 12:59:08 Error reading packet from server: bogus data in log event
> 040628 12:59:08 Got fatal error 1236: 'bogus data in log event' from master when
> reading data from binary log
> 040628 12:59:08 Slave I/O thread exiting, read up to log 'daredevil1-bin.001',
> position 427847536
> On the master server, when I run mysqlbinlog on the daredevil1-bin.001 log, I get the
> following error:
> ERROR: Error in Log_event::read_log_event(): 'Event too small', data_len: 0,
> event_type: 0
> I've tried using the -f option with mysqlbinlog, but I can't read past this point in
> the log. However, the master server is running along just fine, writing new info into
> this log file just fine.
> Is there any way to get the slaves past this bad record, short of manually resyncing
> everything again?
> Thanks for your help!
> Bill Earl
> Network Admin