Stephane VAROQUI
Stephane VAROQUI
https://github.com/github/orchestrator/blob/master/docs/using-the-web-api.md
Solution would be to set replication delay to 9999 when detecting a replication stopped on slave
Use etcd client to store the active replication manager, failover, switchover should implement key retrieval before acting to detect a possible split brain scenarion only the active replication manager should...
We could define a rolling_sql_script that will be execute on all node the rolling way
A client have a master with some old replication thread running by starting replication the master get the read only falg , This task would prevent read-only set if the...