Hei
I think we have found a bug in mysql replication.
We have some programs that are working against a master DB and that use
temporary tables. The hole system works without problems until we run a
flush master/slave. Then the replication in the slave DB stops working
because of this.
How to produce the crash:
- We need a system with replication working master<->slave.
Run f.ex:
MASTER: create temporary table testtmp(id text);
MASTER: insert into testtmp values ("testing");
Until here everything works fine. You can continue working with this tmp
table without problems. But if we run:
MASTER: flush tables with read lock;
MASTER: flush master;
SLAVE: flush slave;
MASTER: unlock tables;
MASTER; insert into testtmp values("testing2");
The replication in slave will stop working with a error message saying
that table testtmp does not exist.
This is a huge problem for us. We have many programs working against the
master DB, they run for weeks and the process of stopping/starting all of
them is not trivial, so we can not stop all of them to run a flush
master/slave when the binary update log file gets to big.
Is this a bug?, is it another way of doing what we want to do without
crashing replication? Do you need more information?
Thanks for your help
Sincerely
Rafael Martinez
---------------------------------------------------------------------
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
