Re: trouble backup with perl.

From: Goffredo Saffioti (goffredo.saffioti_at_famaconsulting.it)
Date: 10/23/03 

Date: Thu, 23 Oct 2003 17:07:59 +0200
To: Wiggins d'Anconia <wiggins@danconia.org>

Wiggins d'Anconia ha scritto:

> Goffredo Saffioti wrote:
>
>> Good Evening all .
>>
>> I'm experiencing to write a perl script for to make backup with afio
>> under perl .
>> But a have some problem when try to execute the system command with
>> the " @_ " this seem to be not the right way for exec this system
>> command "i'm not sure" .
>> but in past time i'm used this kind of syntax for another script that
>> make the quota for all users in a directory on the same way.
>> the backup scrpt is as follow:
>>
>>
>> #!/usr/bin/perl -w
>>
>
> use strict; # always
>
>>
>> #use IO::Socket;
>> use IPC::Open2;
>> use Term::ANSIColor;
>> use POSIX ":sys_wait_h";
>>
>
> These three modules are great, but are you sure you need them?
>
>>
>> my $status;
>
>
> Where does $status get used?
>
>> #$pid = fork;
>> #$begin = 1 if ( !$begin );
>> #$end = $i if ( !$end);
>>
>
> I am assuming these won't be coming back, if they are where do they
> come from?
>
>>
>> #$FULLSYSTEM_SOURCE_DIR = "/home/webhosting/utenti/";
>> $MOUNT_DIR = "/mnt/test";
>> $BACKUP_FILE = "Testbackup.cpio";
>> $FULL_LOG_FILE = "Testlog.log";
>>
>
> These have not been scoped.
>
>>
>> $filename = "/home/webhosting/utenti/stats.txt";
>>
>
> No scoping.
>
>>
>> @_=`ls -l`; #Read the direcory content
>
>
> perldoc -f opendir
> perldoc -f readdir
>
> Shelling out is a very inefficient, insecure, error prone way of doing
> this operation, it is much better handled with a dir handle, or with
> the File::Find module if you want recursion. While Perl likes to do
> lots of fancy stuff with @_ and $_, lets dispense with the cryptic
> stuff and use real variable names until we understand when and why we
> should use cryptic names...
>
>> while ($_ = shift @_) {
>
>
> I assume here you are attempting to loop over the list of files? See
> above documentation.
>
>> @strarr = split ' ';
>
>
> Why are you splitting on space?
> perldoc -f split
>
>> open(LOG, ">> $filename") || die "cannot open file";
>
>
> Include $! in your error messages so you know why it died, not just
> that it did. This is a very inefficient loop, if you are going to
> write to the same file every pass over the array then open the file
> once outside of the while loop, then just keep printing to it rather
> than re-opening it.
>
>
>> print LOG "$strarr[8]\n"; #and write the filenames on a file
>
>
> What does the *9th* element have in it? How do you know it is the 9th
> element? If this is supposed to be the filename check out
>
> perldoc File::Basename
>
>> close (LOG);
>
>
> Again if you open your filehandle outside of the while loop you don't
> have to close it every iteration, this should go outside the loop.
>
>> }
>>
>>
>
> At this point I am very confused because it looks like you are writing
> to a file that you then read and loop over to call a command on each
> of the elements, why the write just to read in the same thing? Why not
> just do your work all in one loop?
>
>>
>>
>> open (LOG1, "< /home/webhosting/utenti/stats.txt") || die "cannot
>> open file with LOG1";
>
>
> Again include $!...
>
>> @log1=(<LOG1>);
>
>
> Ok read the whole thing in, no scoping.
>
>> for ($i=1;$i<=((@log1)/4);$i++){ #my dir is too big so i devide the
>> content in parts more little
>>
>
> If you are concerned about the size there is no need to store the
> whole file to an array and then loop over the array. You do realize
> you are skipping the first elmenet of the array right? Array indices
> start at zero. This also isn't dividing your array into parts, it is
> looping from 1 to 4 so you will only get the 2nd through 5th elements
> of the array, and even if it was your initial read of the file stores
> the whole thing to memory so it is a moot point if this for loop did
> only read a quarter at a time....
>
>> open (LOG2, ">> /home/webhosting/utenti/statswork.txt") || die
>> "cannot open file with LOG1";
>
>
> Include $!... I assume this is a file to let you know how far you have
> gone in your work? Again open the file outside of the loop once, then
> print to it inside the loop.
>
>> print LOG2 "$log1[$i]"; #and write this part on a new file
>>
>
>> open (LOG3, "< /home/webhosting/utenti/statswork.txt") || die "cannot
>> open file with LOG1"; #so... now the trouble .
>>
>>
> Why are we reopening the file we just wrote inside the loop?
>
>
>> #this is the system command that i experience to exec but when i exec
>> the script i receive this : sh: -print: command not found
>> and the script terminate .
>> So how can i accomplish this trouble?
>>
>> @_= `/usr/bin/find $log1[$i] -print -depth -mount -xdev |
>> /usr/bin/afio -oulvAZ/ -T3k -s 2000m -L$MOUNT_DIR/$FULL_LOG_FILE
>> $MOUNT_DIR/$BACKUP_FILE`; /
>>
>
> Again use a descriptive variable name for your output capturing. As to
> why you are getting that particular error message I really can't say,
> what is in $log1[$i]? You should try cleaning up the program first,
> then print the commands that should be run rather than just calling
> them, then copy and paste one of them, see if it does what you want
> then add it to the looping constructs. What's the trailing backslash
> doing there? You should also be checking the return status of the
> command by examining $?, etc. perldoc -f system for more
>
>> close (LOG3);
>> close (LOG2);
>> close (LOG1);
>
>
> Assuming you still need to open these at all they should be closed
> outside the loop....
>
>> }
>>
>> Pleeese help I need backup .
>> Thanks in advance and sorry for my ugly english .
>>
>
> I am more worried about the Perl ;-). Depending on what your needs
> are there have to be thousands of backup programs available for free
> online. If all you want is a dynamic list of directories to pass to
> find why not just issue a sub command at the shell?
>
> I still also think this might be better handled by File::Find...
>
> http://danconia.org
>
>
> -- UNTESTED --
> #!/usr/bin/perl
> use strict;
> use warnings;
>
>
> my $MOUNT_DIR = "/mnt/test";
> my $BACKUP_FILE = "Testbackup.cpio";
> my $FULL_LOG_FILE = "Testlog.log";
>
>
> my $backup_dir = '/home/user';
> my $running_output = '/tmp/im_running.txt';
>
>
> my $OUTPUT;
> open($OUTPUT, ">>$running_output") or die "Can't open output file: $!";
>
>
> my $DIRHANDLE;
> opendir($DIRHANDLE, $backup_dir) or die "can’t opendir $backup_dir: $!";
> while (my $file = readdir($DIRHANDLE)) {
> next if (($file eq '.') || ($file eq '..'));
>
>
> unless (-d $file) {
> print $OUTPUT "Skipping non-directory: $file\n";
> next;
> }
> unless (-r $file) {
> print $OUTPUT "Skipping unreadable directory: $file\n";
> next;
> }
>
>
> print $OUTPUT "Backing up starting at $file...\n";
>
>
> my @find_output = `/usr/bin/find $file -print -depth -mount -xdev
> | /usr/bin/afio -oulvAZ/ -T3k -s 2000m -L$MOUNT_DIR/$FULL_LOG_FILE
> $MOUNT_DIR/$BACKUP_FILE`;
> # do some checking of $? here...
>
>
> print $OUTPUT "Back up of $file complete.\n";
> }
> closedir $DIRHANDLE;
>
>
> print $OUTPUT "I am outta here....\n";
>
>
> close($OUTPUT);
>
>
>
Ok your script work fine, but my problem still persist, because the work
directory is too big so i need to do backup on more little portion of
dir , i attempt to process i wish to process $backup_dir/2 or /4 how can
do it?
Thnx in advance .

Relevant Pages

  • Re: Scalable method for searching in relatively big files
    ... directly in the file A each line for the while loop, ... it would interfer with the search if I used the array @a for the while ... So it isn't a scalable solution in my opinion. ... Show us your script, and someone here will be able to help. ...
    (comp.lang.perl.misc)
  • Re: trouble backup with perl.
    ... > I'm experiencing to write a perl script for to make backup with afio ... > under perl. ... I assume here you are attempting to loop over the list of files? ... file to an array and then loop over the array. ...
    (perl.beginners)
  • RE: hashes and loops
    ... Karyn Williams wrote: ... > trying to create a script to remove from a file the items in ... then could create an array which has the filenames: ... Now you could loop off this like: ...
    (perl.beginners)
  • Re: newbie problem - file revision & size
    ... Use of included script samples are subject to the terms specified at ... I would *like* to set up an array of the ... > filenames I want info on, then cycle thru them until the list is finished. ...
    (microsoft.public.windowsxp.wmi)
  • Saving results within a loop
    ... a script that runs that iterates the function over different ... and then use the script you proposed replicate the estimation ... because an array of outcome of this complicated ... I use indexation of a loop to save the estimation results? ...
    (comp.soft-sys.matlab)