[debian-mysql] Fwd: Mylvmbackup appears to break Mysql replication on Jessie

[Brent Clark]

Good day

In February, I sent the below email.

After sending, we stopped testing and waited for Jessie to be released.

With the release of Jessie, it would appear the problem still persists.

Fast forward to today. We tested again.

Basically what we did was
a) build two Jessie servers
b) install and configure MySQL
c) configure replication

We then installed sysbench on each server and let it 'soak test' for a 
period of for days (each server instance writing to its own database).
At no stage did replication break.

On the fifth day, we installed mylvmbackup, we started incurring 
intermittent replication breakages.

I would like to encourage you to please conduct the same tests (i.e. 
setup mysql and on each machine setup mylvmbackup).

If you can assist, it would be very much appreciated.

Kind Regards
Brent Clark

P.s. If you don't have the facilities to test, please don't hesitate to 
let me know,and I can provide you with two Jessie VMs to test and view on.

-------- Forwarded Message --------
Subject: 	Mylvmbackup appears to break Mysql replication on Jessie
Date: 	Sat, 28 Feb 2015 20:32:25 +0200
From: 	Brent Clark <brentgclarklist at gmail.com>
To: 	pkg-mysql-maint at lists.alioth.debian.org



To whom it may concern.

Where I work, we have a client that runs Debian Jessie, with MySQL
replication (I wont go into details about why he is running Jessie,
given that it is still testing).
What we have noticed, is that every so often Mysql replication breaks,
and we have, via a process of elimination, nailed it down to Mylvmbackup.
But bare in mind, this is not everytime Mylvmbackup run.

This is what we get from MySQL
root at all-sql01 ~ # cat /var/log/mysql/error.log
150228  9:20:02 [ERROR] Error reading packet from server: could not find
next log; the first event '' at 4, the last event read from
'/var/log/mysql/mysql-bin.000155' at 231, the last byte read from
'/var/log/mysql/mysql-bin.000155' at 231. ( server_errno=1236)
150228  9:20:02 [ERROR] Slave I/O: Got fatal error 1236 from master when
reading data from binary log: 'could not find next log; the first event
'' at 4, the last event read from '/var/log/mysql/mysql-bin.000155' at
231, the last byte read from '/var/log/mysql/mysql-bin.000155' at 231.',
Error_code: 1236
150228  9:20:02 [Note] Slave I/O thread exiting, read up to log
'mysql-bin.000157', position 4
150228  9:39:41 [Note] Error reading relay log event: slave SQL thread
was killed
150228  9:39:41 [Note] Slave SQL thread initialized, starting
replication in log 'FIRST' at position 0, relay log
'/var/lib/mysql/slave-relay.000001' position: 4
150228  9:39:41 [Note] Slave I/O thread: connected to master
'username at host:3306',replication started in log 'FIRST' at position 4

We notice, after mylvmbackup runs, Mysql increments its binlog, but then
sometimes the slave does not realise this, and then replication breaks.

Come Monday, we are going to have a conference call with the client to
rebuild with Wheezy. I just wanted to send this email, to share our
experiences, and let you know what we are seeing.

I also wanted to say, for Mylvmbackup, I tried the following two
settings in the config.

recoveryopts=--skip-networking --skip-grant --bootstrap --skip-syslog
--skip-slave-start
extra_flush_tables=1

If you have any questions, or would like me to run or do something to
help you test, please don't hesitate in contacting me.

Kind Regards, and thank you for all the work you do.
Brent Clark




-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.alioth.debian.org/pipermail/pkg-mysql-maint/attachments/20150529/4c8893f1/attachment.html>