Simon Riggs wrote:
> + elog(WARNING, "could not set notify for archiver to read log file
> %u, segment %u",
Reason? (disk full, network down, leap year?)
I think elog() calls don't get translated. You should always use
ereport. Tom would know more about the distinction.
> + ereport(LOG,
> + (errmsg("recovery.conf found...starting archive recovery")));
comma
> + elog(LOG, "redo: cannot restore \"%s\" from archive",
> restoreXlog);
Reason?
> + ereport(ERROR,
> + (errcode_for_file_access(),
> + errmsg("could not write archive_status file
> \"%s\" ",
> + tmppath)));
Reason?
> + elog(LOG, "redo: moving last restored xlog to %s", tmppath);
Probably a user doesn't know what "xlog" is. Also, I don't like the
prefix: style you use. Just tell what happened.
> + elog(LOG, "redo: rename failed");
We don't write "foo failed", but "could not do foo". And again: reason?
> + elog(LOG, "redo: archive chain ends; using local copy of \"%s\"",
> restoreXlog);
> + ereport(LOG,
> ! (errmsg("too many transaction log files, removing \"%s\"",
> xlde->d_name)));
How/where is the limit defined?
> + ereport(WARNING,
> + (errcode_for_file_access(),
> + errmsg("chkpt: cannot find archive_status file: %s ",
> + rlogpath)));
There is enough space that you can write "checkpoint". Or actually
don't write anything. What is the archive_status file?
> + elog(WARNING, "chkpt: archiving not yet started for log
> file %s",
> + xlog);
What does that tell me?
> DEBUG MESSAGES
Debug messages should have a DEBUG severity.
>
> + elog(LOG, "backend: written %s", rlogpath );
> +
>
> + elog(LOG, "postmaster: WAKEN_ARCHIVER received, sending SIGUSR1
> to archiver");
>
> + elog(LOG, "chkpt: archiving done for log file %s",
> + xlog);
>
> + elog(LOG, "redo: system(%s)", xlogRestoreCmd);
---------------------------(end of broadcast)---------------------------
TIP 9: the planner will ignore your desire to choose an index scan if your
joining column's datatypes do not match
