Hi all,
Were running replication on our Prod. DB (3.23.42 on SuSE 7.0) since a
half year. Our way to rotate the binlog files was to stop all software
inserting data into the Prod. DB and then reset the master and the
replication server. This works fine but produces a downtime of aprox. 10
minutes every time we rotate the files.
In the last week were tryout a new procedure to rotate the binlog files.
First we locked all tables for writing, wait until the replication
server has executed all outstanding SQL Querys and then reseted the
master
and the replication server. This produces a downtime of less than 2
minutes only.
But after we restarted the replication, the master writes all updates
into one binlog file which grows up and up. If the binlog file reached
the filesystem limit of 2GB (on ext2) all updates written to the file
are lost and the replication stops without any error message.
The strange thing is that we set the max_binlog_size variable in the
my.cnf file to a limit of 30MB. The variable will be ignored if we use
the new procedure to rotate the binlog files but works fine with the old
one.
The final result now is that the DBs are out of sync and we have to
spend a downtime of aprox. 45 minutes to bring the DBs back in sync.
Does anybody knows the reason for that problem and have a fix? Is this
strange behavior a bug or a goodie?
I searched through the mailinglist and the web but could not find any
solution yet.
Thanx...
--
_
__| |_ Andre Stahl Texas Instruments Freising Wafer Fab
\_ _) Probe Systems Haggertystrasse 1 85356 Freising-Germany
\_(
---------------------------------------------------------------------
Before posting, please check:
http://www.mysql.com/manual.php (the manual)
http://lists.mysql.com/ (the list archive)
To request this thread, e-mail <[EMAIL PROTECTED]>
To unsubscribe, e-mail <[EMAIL PROTECTED]>
Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
