convert in hpux commands, script written in linux commands

[Deepu Chakravarty]

I have one perl script written in linux commands. Can someone convert it so that it can run in hpux 11.11? Script is as follows:

#!/usr/bin/perl -w
## Usage:
## perl osw_vmstat_parse.pl
## where is the name of an individual OSW vmstat archive file
## The utility will produce an output file in the form of:
## osw_vm_yymmdd_hhmm.txt
## where yymmdd_hhmm is month, day, year, hours taken from the time portion of
## input archive file.
##
##
use Time::Local;
## build hash table of month name to string
%Month = (
"Jan" => 0,
"Feb" => 1,
"Mar" => 2,
"Apr" => 3,
"May" => 4,
"Jun" => 5,
"Jul" => 6,
"Aug" => 7,
"Sep" => 8,
"Oct" => 9,
"Nov" => 10,
"Dec" => 11
);
%MetaColumns = ();
%ReverseColumns = ();
# initialize stuff here
$version = "1.0";
$carlage = 0; ## handy counter
$definedMeta = "FALSE";
$lineCount = 0; ## to store overall "good" lines
if (! $ARGV[0])
{
print "\n";
print "osw_vmstat_parse.pl Version ", $version, " Usage: \n";
print "perl osw_vmstat_parse.pl \n";
print "where is a vmstat archived file produced by OSWatcher.\n";
exit 1;
}
$fileName = $ARGV[0];
# will be in this form: hostname.oracle.com_vmstat_07.11.09.1300.dat
# chop up name, grab date
($host, $dummy, $date_part) = split("_", $fileName);
$year = substr($date_part, 0, 2);
$month = substr($date_part, 3, 2);
$day = substr($date_part, 6, 2);
$hours = substr($date_part, 9, 4);
## osw_vm_yymmdd_hhmm.txt
## while we have the values, build output filename
$fileOut = "osw_vm_" . $year . $month . $day . $hours . ".txt";
open(INFILE, $fileName) or die "File ", $fileName, " cannot be opened.\n";
## create output filename from input filename
##
while ($line = )
{
# first, chop into array of tokens
@thisline = split(" ", $line);
# grab version of OSW
if ($thisline[1] eq "OSW" )
{
($platform, $dummy, $version, $host) = @thisline; $platform="";
}
# parse out date/time stamp
if ($thisline[0] eq "zzz" )
{
#zzz ***Mon Oct 29 10:01:24 EDT 2007
($junk, $wkday, $month, $mday, $time, $zone, $year) = split(" ", $line);
$junk = $zone; ## just to eliminate spurious warnings
$wkday = substr($wkday, 3);
($hours, $min, $sec) = split(":", $time);
$timeseconds = timelocal($sec, $min, $hours, $mday, $Month{$month}, $year);
}
if ($thisline[0] eq "procs") { }; # do nothing, header thing
if ($thisline[0] eq "r" && $definedMeta eq "FALSE") # col hdrs
{
$colCount = 0;
## save off a "list" of columns, in original order
@columnNames = @thisline;
foreach $column (@thisline)
{
# first the obverse
$MetaColumns{$column} = $colCount;
# then the reverse
$ReverseColumns{$colCount} = $column;
$colCount++;
}
$definedMeta = "TRUE";
}
## now eat first two lines of vmstat output
if ($thisline[0] =~ /^\d+$/ && $thisline[1] =~ /^\d+$/ )
{
$carlage++;
if ($carlage == 3) # ...and the number of the counting shall be three.
{
$bucketNumber = 0;
foreach $bucket (@thisline)
{
$bucketName = $ReverseColumns{$bucketNumber};
push @$bucketName, $bucket;
$bucketNumber++;
}
push @timestamp, $timeseconds;
$carlage = 0; ## reset to a value suitable for carl
$lineCount++; ## bump "good" lines, so we know later
}
}
}
## knock out a text file of stuff
## osw_vm_yymmdd_hhmm.txt
open (TXTOUT, "> $fileOut") or die "Cannot open ", $fileOut, "for writing\n";
print TXTOUT "timestamp ";
foreach $col (@columnNames)
{
print TXTOUT $col, " ";
}
print TXTOUT "\n";
$lines = 0;
while ($lines < $lineCount)
{
print TXTOUT $timestamp[$lines], " ";
foreach $col (@columnNames)
{
print TXTOUT $$col[$lines], " ";
}
print TXTOUT "\n"; ## never forget the linefeed!
$lines++; ## and bump da pointer
}
close TXTOUT;

[Patrick Wallek]

I'm not sure what you mean by "written in linux commands." At a first glance I don't see any OS level commands in the perl script.

Perl is very portable and the same script should run on Linux and HP-UX just fine.

Have you tried running this script on HP-UX? If so, did you get any errors? If you did get errors it would help immensely if you posted those errors.

[Yogeeraj_1]

hi,


Since this is perl, it will work similar to the linux environment.

obvious changes will concern line 1:
#!/usr/bin/perl -w

where you will have to refer to the appropriate location of perl on your hp-ux 11.11 host. (do "which perl" to get the exact path)

However, calls to hp-ux commands may require individual analysis. E.g. vmstat output can be different on Linux compared to the HP-UX environment.

I would suggest that you break the script in smaller modules and try them individually, one by one.

good luck!

kind regards
yogeeraj

[Deepu Chakravarty]

Hi,

It is giving following error:

$ perl osw_vmstat.pl
Unrecognized character \xCD at osw_vmstat.pl line 18.

'which perl' gives following path:

$ which perl
/usr/bin/perl

[Hein van den Heuvel]

When I tossed (cut & paste) the text above into an HPUX file the semicolons (;) turned into dots (.).
Looks like that source is in UNICODE.
Paste it into a notepad, save as ANSI TEXT.
Open the TEXT file, cut & paste from there.
Ah... the seimicolons stay.
And the program works... or at least prints the help text when no args ar provided.

fwiw... is is not the pretties perl / parsin I've seen. Hard to read with constructs like:
"if ($thisline[0] eq "procs") { }; # do nothing, header thing"

I guess it serves as documentation, but why?

Cheers,
Hein.

[VK2COT]

Hello,

Hein is right, your Perl script was probably written on a Windows machine and used
the encoding 'utf-8'.

Certain editors (in particular on win32) may
add a UTF-8 Byte Order Marker (BOM:

http://www.unicode.org/faq/utf_bom.html#BOM)

at the beginning of the file. It creates a
situation where BOM appears past the
beginning of the file...

To test it further:

# hexdump -C osw_vmstat_parse.pl (Linux)

or

# od -t a osw_vmstat_parse.pl (HP-UX)

Your script is not wrong (albeit not
the best coding technique). Encoding is.

Cheers,

VK2COT