Binlog servers (BLS) won't move between co-masters

[ethan-riskiq]

Binlog servers (BLS) won't move between co-masters

[shlomi-noach]

Is that a MaxScale binlog server?

A MaxScale binlog server cannot be moved. It just can't; you need to kill it, reconfigure it, then start it again.

This is known to orchestrator which is why the Web UI won't let you do it. I will need to add a similar check to the CLI to let you know this can't be done (and this will avoid the runtime error you got).

I'd argue you don't want to move the binlog server, though. What you want to so is move slave a.
If using Pseudo-GTID, you can just

orchestrator -c relocate -i slave.a -d binlog02

Actually this may just work without Pseudo-GTID for you, because of the relative simplicity of the topology (one up via binlog server). Anyway if it doesn't (could be), you'd use multiple steps:

orchestrator -c relocate -i slave.a -d binlog01
orchestrator -c move-below -i slave.a -d master.2
orchestrator -c relocate -i slave.a -d binlog02

[shlomi-noach]

My assignment: to make sure CLI refuses to move a BLS

[ethan-riskiq]

okay...any suggestions for how to do this without bringing down the write master and having to execute a change master on each binlog host while it's down?

[shlomi-noach]

any suggestions for how to do this

Sorry, how to do what ?

The maintenance? Like I suggested, the binlog servers aren't your problem, are they? You wish to move the replicas around and I've given you the recipe to do that.

Otherwise I'm not sure what you are asking.

[shlomi-noach]

Closing this seeing this isn't an orchestrator issue. Feel free to continue discussion even as this issue is closed.