Hi all,
I recently hit a problem about sync replication of postgres. When I check the state of the replication, I got this:
postgres=# select * from pg_stat_replication ;
pid | usesysid | usename | application_name | client_addr | client_hostname | client_port | backend_start | backend_xmin | state | sent_location | write_location | flush_location | replay_location | sync_priority | sync_state
------+----------+----------+---------------------+---------------+---------------------------------------------+-------------+-------------------------------+--------------+-----------+---------------+----------------+----------------+-----------------+---------------+------------
4674 | 10 | postgres | 10.87.107.182:20133 | 10.87.107.182 | ip-10-87-107-182.us-west-2.compute.internal | 42408 | 2018-08-20 23:51:22.742259+00 | | streaming | 0/1C83A00 | 0/1C83A00 | 0/1C83B50 | 0/1C83B50 | 1 | sync
As you can see, the flush_location is ahead of write_location. I am wondering how this could happen?
Thanks,
Simon
