On Fri, Mar 28, 2008 at 10:30 AM, Gurjeet Singh <
singh.gurjeet@gmail.com> wrote:
I am still looking for comments on the correctness of this script and above mentioned procedure for running it on an 8.2.x release.
Well, I came across a serious bug in the script. Here's the corrected version of the perl script for anyone who might have picked up the script from here:
(Am still looking for any feedback on the correctness of the script; especially in the light of the recent possible bugs fixed on the server side, related to recovery)
<script>my @text = `pg_controldata .`; # here . represents the PGDATA, since the server is executing here.my $line;my $time_line_id; my $redo_log_id;my $redo_rec_off;my $wal_seg_bytes;my $redo_log_seg;foreach $line ( @text ){ $line = mychomp( $line ); if( $line =~ m/Latest checkpoint's TimeLineID:\s*(([0-9])+)/ ) { # decimal number $time_line_id = 0 + $1; } if( $line =~ m/Latest checkpoint's REDO location:\s*(([0-9]|[A-F])+)\/(([0-9]|[A-F])+)/ ) {
# hexadecimal numbers
$redo_log_id = hex( $1 );
$redo_rec_off = hex( $3 );
}
if( $line =~ m/Bytes per WAL segment:\s*([0-9]+)/ )
{
# decimal number
$wal_seg_bytes = $1;
}
}
$redo_log_seg = sprintf( "%d", $redo_rec_off / $wal_seg_bytes );
print "" . sprintf( "%08X%08X%08X", $time_line_id, $redo_log_id, $redo_log_seg ) . "\n";
# Wrapper around Perl's chomp function
sub mychomp
{
my ( $tmp ) = @_;
chomp( $tmp );
return $tmp;
}
</script>
--
gurjeet[.singh]@EnterpriseDB.com
singh.gurjeet@{ gmail | hotmail | indiatimes | yahoo }.com
EnterpriseDB
http://www.enterprisedb.comMail sent from my BlackLaptop device
