2.0.4: Could not find master among the backend servers. Previous master's state : RUNNING SLAVE

[Mark P]

I have previously posted this issue and the fix was to upgrade.

After upgrading to 2.0.4 I'm still seeing in maxscale.log:

2017-03-23 17:21:31   error  : Could not find master among the backend servers. Previous master's state : RUNNING SLAVE

This is sometimes happening ~20 times a second... the database is still usable through maxscale.

There are no database events/issues/etc anywhere near this time.

I have turned on the debug log to hopefully help diagnosing this issue (attached).

I have also attached my configuration, and a dump of the "show xxxx" commands (run just after dumping the logs)

Im also seeing another issue:

Looking through maxscale.log I see :

2017-03-23 17:22:34   error  : Could not find master among the backend servers. Previous master's state : RUNNING SLAVE

2017-03-23 17:22:34   warning: [RW Split Router] Write query received from ro...@192.168.5.152. Could not find a valid master connection. Closing client connection.

2017-03-23 17:22:34   error  : Routing the query failed. Session will be closed.

2017-03-23 17:22:35   error  : Could not find master among the backend servers. Previous master's state : RUNNING SLAVE

This error seemed to correlate with a "Mysql::Error: Lost connection to MySQL server during query" I'm occasionally getting in my application (my application will make a series of sql calls then is suddenly fails with that error).

I'm hoping the issue is related as I have been getting these random "Mysql::Error: Lost connection to MySQL server during query" errors for many months.

Thanks for any help you can provide.

Mark