Change Details

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

Once the goal of moving sanitarium to multi-instance (T190704) is achieved, we will end up with 4 sanitarium hosts, 2 active in eqiad (db1124 and db1125) and 2 passive in codfw (db2094 and db2095). Each host replicate different sections on each DC. db1124: s1, s3, s5, s8 db2094: s1, s3, s5,s8 db1125: s2, s4, s6, s7 db1095: s2, s4, s6, s7 If we have to failover to codfw, we need to manually look for the equivalent positions (if we do not use GTID) on the binlogs, so we can star replication on the other DC. It would be a good idea to have a script that automatically provides the exact binlog/position coordinates that are equivalent on both hosts, so replication can started.