image-2018-02-13-16-34-19-270.png
[MXS-1667] Fixing M/M replication between datacenters when Master changes in multi-DC environment Created: 2018-02-13 Updated: 2018-08-31 Resolved: 2018-08-31 |
|
| Status: | Closed |
| Project: | MariaDB MaxScale |
| Component/s: | mariadbmon |
| Affects Version/s: | None |
| Fix Version/s: | 2.3.0 |
| Type: | New Feature | Priority: | Major |
| Reporter: | Dipti Joshi (Inactive) | Assignee: | Esa Korhonen |
| Resolution: | Won't Do | Votes: | 0 |
| Labels: | None | ||
| Attachments: |
image-2018-02-13-16-34-19-270.png
|
| Sprint: | MXS-SPRINT-58, MXS-SPRINT-59, MXS-SPRINT-60, MXS-SPRINT-61, MXS-SPRINT-62, MXS-SPRINT-63, MXS-SPRINT-64, MXS-SPRINT-65 |
| Description |
|
Fixing M/M replication between datacenters when Master changes in either one for following configuration Add an automation option so that when MaxScale node on mate-DC does switchover or failover, local master's external master is re-homed to the new master in mate-DC. This requires either (A) communication between the two MaxScales, or alternatively (B) both MaxScales need to monitor both clusters. B seems like an easier option. Both MaxScales could have two monitors, one for the local and one for the external cluster. The local monitor module queries the external monitor for the master server regularly, or the external monitor notifies the local monitor if its master changes. |
