mysqldump causes slave to stop

[Steven Boger]

Hello,

I am dumping a single innodb DB using mysqldump.  Everytime I do it, it
crashes the slave thread.

Below are the commands I'm using to dump [I also have tried
--delete-master-logs command in the actual data dump too (second line)
with the same problem.]  I am willing to hear advice on the best place to
put this.

Note, I am cycling the bin-logs every fine minutes to provide a type -of
point-in-time backup using 'mysqladmin flush-logs'.  Thats why the binlog
numbers are so high.  I have the flush command NOT run if mysqldump is
running.

Important side-note: I have removed --quick from these as it was causing
another innodb db on the same server to *appear* to LOCKUP during the
backup of THIS db... ideas?

Also. during this backup, all threads attached to the db go into "Waiting
for release of readlock" - Is this normal?  Master is 4.0.14, slave is
4.0.20... Slave was 4.0.14 and I still saw the problem.



<MASTER DUMP COMMANDS>

# safety first!!! make a daily backup of the XXXXXX schema!
  mysqldump --extended-insert --single-transaction --no-data --add-locks
--compress --skip-lock-tables --delete-master-logs XXXXXX | bzip2 -c >
/mnt/data/backups/`date +%Y%m%d`/XXXXXX.SCHEMA.sql.bz2

# at one minute after midnight, fire off the backup
  mysqldump --single-transaction --add-locks --compress --skip-lock-tables
--extended-insert --master-data XXXXXX | bzip2 -c >
/mnt/data/backups/`date +%Y%m%d`/XXXXXX.dump.sql.bz2

 </MASTER DUMP COMMANDS>



<SLAVE ERROR LOG>

040605  0:09:20  Error reading packet from server: Could not open log file
(server_errno=1236)
040605  0:09:20  Got fatal error 1236: 'Could not open log file' from
master when reading data from binary log
040605  0:09:20  Slave I/O thread exiting, read up to log
'atlblg01-bin.262', position 714675

</SLAVE ERROR LOG>