Would it be possible for a long running select query on a standby to prevent the
standby from sending status updates? This seemed to occur and consequently the
master broke the connection:
### Master:
2012-10-31 00:56:44 UTC replicator 5057 192.168.1.20(45690) database
[[unknown]] LOG: terminating walsender process due to replication
timeout
### Slave:
2012-10-31 00:56:47 UTC 12834 database[] FATAL: could not send data
to WAL stream: server closed the connection unexpectedly
This probably means the server terminated abnormally
before or while processing the request.
2012-10-31 00:57:02 UTC johndoe 27599 192.168.1.30(46426) database[main]
LOG: duration: 65886.667 ms statement: select...
### Replication Config
replication_timeout = 5s
wal_receiver_status_interval = 2s
Thanks, Jesse Hathaway