Bug #67138 mysqldump: Couldn't execute 'FLUSH /*!40101 LOCAL */ TABLES'

Description:
I use mysqldump with Ashay OBS to backup my database.
The database is running a master master configuration, and I use one for production and the other for fallback and runs the backup from it too.

I'm starting to get these errors when running the mysqldump. Not every day, but almost.

mysqldump: Couldn't execute 'FLUSH /*!40101 LOCAL */ TABLES': Got an
error writing communication packets (1160)

The consequence does not seem to be severe, but I dont like such messages.

Recently I upgraded the fallback server from 5.1.41-3 to 5.5.24-0, thus I suspect that this behaviour is related to 5.5.24-0

I have been googling for a long time but I haven't found any similar examples of this behaviour.

 

How to repeat:
Run this command
/usr/bin/mysqldump --master-data=2 --routines --single-transaction --quick --routines $*

Thank you for the bug report. Looks like duplicate of http://bugs.mysql.com/bug.php?id=66184 . Please check for duplicate/related behavior.

No feedback was provided for this bug for over a month, so it is
being suspended automatically. If you are able to provide the
information that was originally requested, please do so and change
the status of the bug back to "Open".

Indeed,there is a bug.But I have found workarround.
It should replace connection method and use port instead socket.

Regards.

Baruch Lvovsky.

duplicate of similar federated/table_cache bugs. the federated table loose connection to the remote server.  when table is removed from table_cache, it fails because connection is not valid.  

In fact you may have better luck in 5.6.23+ There have been a few fixes (not all fixed, but some), for example:

FLUSH TABLES on a FEDERATED table failed if the table had been idle longer than the wait_timeout time plus the TCP keepalive time. (Bug #17599258)