Hello everyone,
I am applying postgres PITR for quite a month. Not before long, it worked well but recently something wrong happens. The WALs have not been sending to the remote backup host. All the WALs are continuously generated and stored in the local PG directory (pg_xlog/).
I checked the PG's log and found out an error message: "postgres: archiver process failed on 0000000400000007000000FA"
It seems that the WAL is missing so PG could not continue the process. As a result, I manually copy the specific WAL from the remote backup host to the local PG directory. After that, I thought I had solved the problem because the message showed: "postgres: archiver process last was 000000090000000800000012".
And then, I executed the restore process stated in the PG manual step by step. It worked fine. The message in log showed "LOG: archive recovery complete", "LOG: received SIGHUP, reloading configuration files". I also saw the PG server is on. However, it showed again, "postgres: archiver process failed on 0000000400000007000000FA". And again, all the WALs were continuously generated and stored in the local PG directory.
Note:
When I copy the specific WAL from the remote backup host to the local PG directory, it temporarily solves the problem. The WALs successfully send to the remote backup host. However, when I perform restore, the same situation happens again after PG server is on.
I am pleased that somebody could help me.
Thank you very much.
Best regards,
Leo