Hi,
I set up an warm standby server which fetches WAL logs from a remote
server. It has been working very well. But today, I was editing the
restore script(which is set as the restore_command) with vi. Just right
after I saved the script, I noticed suddenly the standby server
terminated. I checked log and only found these:
scp: /dbbk/postgres/logarch/0000000100000079000000B3.gz: No such file or
directory #this is in my script and it unzip the WAL log if it exists
2011-09-20 23:56:50 CST FATAL: could not restore file
"0000000100000079000000B3" from archive: return code 65280
2011-09-20 23:56:51 CST LOG: startup process (PID 2153) exited with
exit code 1
2011-09-20 23:56:51 CST LOG: terminating any other active server processes
I believe the script doesn't exit with 0 for any error(and it has been
working smoothly without any modification after server restart). After I
started the server, it just restored passing the file without problem
and it can live with file "No such file" warning because the master
server hasn't generated the log yet. I have no idea why the restore
failed. But even the restore fails, why it lead to a server crash(looks
like so)?