Привет.
Производил плановый фейловер.
1. Остановил мастер
2. standby promote
3. На мастере создал recovery.conf
standby_mode = 'on'
primary_conninfo = 'user=postgres host=10.10.10.1 port=5432'
primary_slot_name = 'standby_slot'
recovery_target_timeline = 'latest'

в конфиге hot_standby = on

Стартую постгрес, и он уходит в бесконечное ожидание старта... в логах:
2019-10-18 12:06:07 MSK 3392 LOG: database system is shut down
2019-10-18 12:09:35 MSK 21875 LOG: database system was shut down at 2019-10-18 12:06:02 MSK
2019-10-18 12:09:35 MSK 21875 LOG: entering standby mode
2019-10-18 12:09:35 MSK 21875 LOG: invalid record length at 290/70000098: wanted 24, got 0
2019-10-18 12:09:35 MSK 21883 LOG: fetching timeline history file for timeline 2 from primary server
2019-10-18 12:09:36 MSK 21883 LOG: started streaming WAL from primary at 290/70000000 on timeline 1
2019-10-18 12:09:36 MSK 21883 LOG: replication terminated by primary server
2019-10-18 12:09:36 MSK 21883 DETAIL: End of WAL reached on timeline 1 at 290/70000098.
2019-10-18 12:09:36 MSK 21875 LOG: new target timeline is 2
2019-10-18 12:09:36 MSK 21883 LOG: restarted WAL streaming at 290/70000000 on timeline 2
2019-10-18 12:09:38 MSK monuser 10.10.10.10 postgres 21901 FATAL: the database system is starting up
2019-10-18 12:09:39 MSK 21875 LOG: redo starts at 290/70000098
2019-10-18 12:09:54 MSK monuser 10.10.10.10 user 21976 FATAL: the database system is starting up
..
..
ждал в течение часа, сделал ей повторный рестарт, без изменений.

2019-10-18 13:16:06 MSK 21842 LOG: received fast shutdown request
2019-10-18 13:16:06 MSK 21883 FATAL: terminating walreceiver process due to administrator command
2019-10-18 13:16:07 MSK monuser 10.10.10.10 user 9750 FATAL: the database system is shutting down
2019-10-18 13:16:08 MSK 21881 LOG: recovery restart point at 290/739D02B0
2019-10-18 13:16:08 MSK 21881 DETAIL: last completed transaction was at log time 2019-10-18 13:16:05.264617+03
2019-10-18 13:16:08 MSK 21881 LOG: shutting down
2019-10-18 13:16:08 MSK 21842 LOG: database system is shut down
2019-10-18 13:16:19 MSK 9808 LOG: database system was shut down in recovery at 2019-10-18 13:16:08 MSK
2019-10-18 13:16:19 MSK 9808 LOG: entering standby mode
2019-10-18 13:16:19 MSK 9808 LOG: redo starts at 290/739D02B0
2019-10-18 13:16:19 MSK monuser 10.10.10.10 user 9811 FATAL: the database system is starting up
2019-10-18 13:16:20 MSK 9808 LOG: invalid resource manager ID 183 at 290/7CE01F78
2019-10-18 13:16:20 MSK 9812 LOG: started streaming WAL from primary at 290/7C000000 on timeline 2
2019-10-18 13:16:20 MSK monuser 10.10.10.10 user 9815 FATAL: the database system is starting up
2019-10-18 13:16:22 MSK monuser 10.10.10.10 user 9819 FATAL: the database system is starting up


Далее, попробовал убрать из конфига hot_standby.
Без hot_standby сервис успешно запустился, однако хочется доступа на реплику. Возвращаю параметр - опять бесконечный: FATAL: the database system is starting up
Пришлось перезаливать реплику из бэкапа :(
Что это и как с этим бороться?
PostgreSQL 10.6 on x86_64-redhat-linux-gnu, compiled by gcc (GCC) 4.8.5 20150623 (Red Hat 4.8.5-16), 64-bit