Обсуждение: [ADMIN] replication error

Hello.

I try setup replica from instructions located at https://wiki.postgresql.org/wiki/Streaming_Replication

I copied data to slave, after psql -c "SELECT pg_start_backup('label', true)"  and start slave -  i see in logs many message like

cp: cannot stat '/var/lib/postgresql/9.6/main/archive/000000010000000000000012': No such file or directory
2017-02-28 17:48:45 MSK [2333-3] LOG:  redo starts at 0/12000098
2017-02-28 17:48:45 MSK [2333-4] LOG:  consistent recovery state reached at 0/120000D0
2017-02-28 17:48:45 MSK [2333-5] LOG:  record with incorrect prev-link 10000/2470000 at 0/120000D0
2017-02-28 17:48:45 MSK [2332-1] LOG:  database system is ready to accept read only connections
2017-02-28 17:48:45 MSK [2339-1] LOG:  started streaming WAL from primary at 0/12000000 on timeline 1
on master node file 000000010000000000000012  located in pg_xlog directory and copy to archive diectory  after execute
psql -c "SELECT pg_stop_backup()" and after copy database data to slave.

What i doing wrong?

Anton Gorlov wrote:
> I try setup replica from instructions located at
> https://wiki.postgresql.org/wiki/Streaming_Replication
> 
> 
> I copied data to slave, after psql -c "SELECT pg_start_backup('label', true)"  and start slave -  i
> see in logs many message like
> 
> cp: cannot stat '/var/lib/postgresql/9.6/main/archive/000000010000000000000012': No such file or
> directory
> 2017-02-28 17:48:45 MSK [2333-3] LOG:  redo starts at 0/12000098
> 2017-02-28 17:48:45 MSK [2333-4] LOG:  consistent recovery state reached at 0/120000D0
> 2017-02-28 17:48:45 MSK [2333-5] LOG:  record with incorrect prev-link 10000/2470000 at 0/120000D0
> 2017-02-28 17:48:45 MSK [2332-1] LOG:  database system is ready to accept read only connections
> 2017-02-28 17:48:45 MSK [2339-1] LOG:  started streaming WAL from primary at 0/12000000 on timeline 1
> 
> 
> on master node file 000000010000000000000012  located in pg_xlog directory and copy to archive
> diectory  after execute
> psql -c "SELECT pg_stop_backup()" and after copy database data to slave.

Looks like everything is fine.

The standby recovered all WAL it could find in /var/lib/postgresql/9.6/main/archive/
and then connected to the primary for streaming replication.

000000010000000000000012 is not archived yet, so the standby cannot find it.

Test replication by changing something on the primary and checking if the
change gets replicated.

Yours,
Laurenz Albe