On Tue, Sep 3, 2019 at 9:57 AM Matthias Apitz <guru@unixarea.de> wrote:
> 2019-09-03 09:18:46.024 CEST [25388] LOG: Logdatei »00000001000000010000008F« aus Archiv wiederhergestellt
> ***> cp: der Aufruf von stat für '/data/postgresql11/wal_archive/000000010000000100000090' ist nicht möglich: Datei
oderVerzeichnis nicht gefunden
It's not an error, it is searching the next wal that after 0x8F is
0x90, but the log is not there, so the recovery has finished. The
cluster does not know where to stop because your recovery.conf did not
specified, so it tries to do as much wal processing as it can and
stops when it finds no more.
> 2019-09-03 09:18:47.838 CEST [25388] LOG: Redo fertig bei 1/8F000140
Redo is done.
> ***> cp: der Aufruf von stat für '/data/postgresql11/wal_archive/00000002.history' ist nicht möglich: Datei oder
Verzeichnisnicht gefunden
having finisched the wals it is searching for a branch on timeline 2
> 2019-09-03 09:18:47.900 CEST [25388] LOG: gewählte neue Zeitleisten-ID: 2
and it switches on timeline 2 autonomously
Pretty much that should be what you see in the logs.
> Btw: Is there away to run the server for German UTF-8, but with English
> messages in the log?
Set lc_messages in the postgresql.conf to something in english, I
think 'en_US.UTF-8' could do the trick, but haven't tried it.
Luca
