read only queries on hot standby server fail due to acquiring transaction id

Question

we just backed up a primary onto a hot standby server successfully. however, when we try to query to hot standby, we get the following error:

ERROR: cannot assign TransactionIds during recovery

I enabled verbose mode and see the error occurs at varsup.c:72

/* safety check, we should never get this far in a HS standby */ 
if (RecoveryInProgress())
   elog(ERROR, "cannot assign TransactionIds during recovery");

the strange thing is, if we use the psql client and issue multiple read queries, the first one will fail, but all subsequent ones will succeed.

pipeline=#    SELECT pg_is_in_recovery();
ERROR:  cannot assign TransactionIds during recovery
pipeline=#    SELECT pg_is_in_recovery();
 pg_is_in_recovery 
-------------------
t
(1 row)

using postgres 9.5.3

any ideas would be greatly appreciated it.

btw: we have added: hot_standby = on to pipelinedb.conf (equivalent to postgresql.conf)

and we have added:

primary_slot_name = 'replicator_slot'
primary_conninfo = 'user=username host=ip_address port=5432'
recovery_target_timeline = 'latest'

to recovery.conf

Answer 1

Use

pg_ctl promote

to exit recovery mode and resume normal (read/write) operation.

From the postgresql documentation:

In promote mode, the standby server that is running in the specified 
data directory is commanded to exit recovery and begin read-write 
operations.