[Sysadmins] Падает Postgresql 12.1 с логической репликацией.

Yuri Khachaturyan yukh на yukh.ru
Ср Янв 29 15:02:20 MSK 2020


Добрый день!

Долгое время работали 2 сервера с логической репликацией master - slave
одной БД под управлением Postgresql10.

Вчера обновил оба сервера до 12.1 и началось:

2020-01-29 14:56:52.575 MSK [15370] LOG:  background worker "logical
replication worker" (PID 22239) was terminated by signal 11: Segmentation
fault
2020-01-29 14:56:52.575 MSK [15370] LOG:  terminating any other active
server processes
2020-01-29 14:56:52.576 MSK [24708] WARNING:  terminating connection
because of crash of another server process
2020-01-29 14:56:52.576 MSK [24708] DETAIL:  The postmaster has commanded
this server process to roll back the current transaction and exit, because
another server process exited abnormally and possibly corrupted shared
memory.
2020-01-29 14:56:52.576 MSK [24708] HINT:  In a moment you should be able
to reconnect to the database and repeat your command.
2020-01-29 14:56:52.576 MSK [15375] WARNING:  terminating connection
because of crash of another server process
2020-01-29 14:56:52.576 MSK [15375] DETAIL:  The postmaster has commanded
this server process to roll back the current transaction and exit, because
another server process exited abnormally and possibly corrupted shared
memory.
2020-01-29 14:56:52.576 MSK [15375] HINT:  In a moment you should be able
to reconnect to the database and repeat your command.
2020-01-29 14:56:52.633 MSK [15370] LOG:  all server processes terminated;
reinitializing
2020-01-29 14:56:52.895 MSK [30309] LOG:  database system was interrupted;
last known up at 2020-01-29 14:54:49 MSK
2020-01-29 14:56:53.408 MSK [30309] LOG:  recovered replication state of
node 1 to 2/40C8CA80
2020-01-29 14:56:53.408 MSK [30309] LOG:  database system was not properly
shut down; automatic recovery in progress
2020-01-29 14:56:53.410 MSK [30309] LOG:  redo starts at 6/DCD2A208
2020-01-29 14:56:53.415 MSK [30309] LOG:  redo done at 6/DCEBB168
2020-01-29 14:56:53.422 MSK [30312] FATAL:  the database system is in
recovery mode
2020-01-29 14:56:53.491 MSK [15370] LOG:  database system is ready to
accept connections
2020-01-29 14:56:53.529 MSK [30319] LOG:  logical replication apply worker
for subscription "wm0" has started


Что это может быть и куда копать?
На период теста, на slave нет никакой нагрузки, на него просто идет
репликация одной единственной БД.



-- 
С уважением,
Хачатурян Юрий (yukh на yukh.ru)
----------- следующая часть -----------
Вложение в формате HTML было удалено...
URL: <http://lists.altlinux.org/pipermail/sysadmins/attachments/20200129/5b76ba02/attachment.html>


Подробная информация о списке рассылки Sysadmins