[MXS-1132] Does not accept forcing monitor flag to a server in some case Created: 2017-02-09  Updated: 2018-03-26  Resolved: 2018-03-26

Status: Closed
Project: MariaDB MaxScale
Component/s: Core
Affects Version/s: 2.0.4
Fix Version/s: N/A

Type: New Feature Priority: Major
Reporter: VAROQUI Stephane Assignee: Unassigned
Resolution: Incomplete Votes: 0
Labels: None

Issue Links:
Blocks
is blocked by MXS-978 Support for stale master in case of r... Closed

 Description   

Some external abitrator tool can not impose to maxscale it's own view of the world

The case a server become the last available server in a replication topology , we should be able to tell maxscale to use f the server never was a master before and lost all other infoon other nodes , for any type of router plugin , it should route read and write trx to that lonely imposed server.
This feature should be addressable by external client telnet or maxadmin mysql protocol.

A test case could be build using replication manager if we and manager decide to invetigate into a build and non regression architecture for replication-manager

Any workaround with existing feature is welcome and should be better documented if this is not a real issue



 Comments   
Comment by markus makela [ 2017-02-09 ]

The main use case for this would be when the monitor would pick one server but the user would want another server to be the master. This could be thought of as guided monitoring mode. It would allow for safer conflict resolution between the monitoring in MaxScale and the real state of the cluster.

Comment by Johan Wikman [ 2017-02-14 ]

Changed to feature.

Comment by Johan Wikman [ 2018-03-26 ]

Seems irrelevant now that failover-functionality is built in to MaxScale.

Generated at Thu Feb 08 04:04:28 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.