I am writing some DBA scripts with perl, one of them is the archive log file related. I have to implement this feature for more works. The first version is implemented by UNIX find command.
#
# Get the full archived log name by sequence
#
sub findLogBySequence
{
my ($logseq) = @_;
my $logfile = "";
if (length($logfile) == 0)
{
# Foloowing logic is for Oracle 9i (sid_%t_%s.arc)
$logfile = `find /data*/arch -follow -name *_${logseq}.* | head -1`;
chomp($logfile);
# Foloowing logic is for Oracle 10g (sid_%t_%s_%r.arc)
if (length($logfile) == 0)
{
$logfile = `find /data*/arch -follow -name *_${logseq}_* | head -1`;
chomp($logfile);
}
}
scalar $logfile;
}
Later I feel that the find command will run as many times as archive log count, it's not so effective enough, then I rewrite it with the following, it elimated lots of executions of find command.
my $logfmt = "???";
#
# Get the full archived log name by sequence
#
sub findLogBySequence
{
my ($logseq) = @_;
my $logfile = "";
#
# if archive log pattern is not null
# get the archive log file from pattern
#
if ($logfmt ne "???")
{
$logfile = $logfmt;
$logfile =~ s/XXXXXX/${logseq}/g;
if (! -e $logfile)
{
$logfile = "";
$logfmt = "???";
}
}
if (length($logfile) == 0)
{
# Foloowing logic is for Oracle 9i (sid_%t_%s.arc)
$logfile = `find /data*/arch -follow -name *_${logseq}.* | head -1`;
chomp($logfile);
# Foloowing logic is for Oracle 10g (sid_%t_%s_%r.arc)
if (length($logfile) == 0)
{
$logfile = `find /data*/arch -follow -name *_${logseq}_* | head -1`;
chomp($logfile);
}
}
# Store the archive log pattern
if (length($logfile) and -e $logfile and $logfmt eq "???")
{
$logfmt = $logfile;
$logfmt =~ s/_${logseq}\./_XXXXXX\./;
$logfmt =~ s/_${logseq}_/_XXXXXX_/;
}
scalar $logfile;
}
I will write more perl DBA scripts with it, for example, keep given gigabytes of archive logs.