Loading ...
Sorry, an error occurred while loading the content.

Re: Trying to get File and Directory info off of external server quickly

Expand Messages
  • Philip M. Gollucci
    ... I agree with that whole hardly. (the p6m7g8.net cluster is doing that) As long as you don t reboot its great though :)
    Message 1 of 16 , Aug 1, 2005
    • 0 Attachment
      Gedanken wrote:
      > One quick caveat - properly unmount volumes when rebooting and such. Due
      > to one of those things that somehow grew beyond its original intent, we
      > had a network of about 15 machines all mounting each other. NFS
      > chokes when a mount it expects to be there, isnt. it takes it several
      > minutes to give up. One machine rebooting in such a big spiderweb cluster
      > can cause massive problems without proper attention to cleanly unmounting
      > and remounting shares. And 'cascading' is usually such a lovely word...
      I agree with that whole hardly. (the p6m7g8.net cluster is doing that)
      As long as you don't reboot its great though :)
    • Philippe M. Chiasson
      ... Wouldn t HTTP HEAD request achieve this fairly nicely ? I am not sure you have described the actual problem you are trying to solve. Why is that
      Message 2 of 16 , Aug 1, 2005
      • 0 Attachment
        Boysenberry Payne wrote:
        > Hello All,
        >
        > I've got a two server platform one a static server for files and
        > runs the mysql server
        > and the other runs mod_perl. I'm trying to figure out the fastest way
        > to get info on directories
        > and files from the static server to the mod_perl server. Right now I'm
        > using Net::FTP which
        > is really slow, especially when they're a lot of files. Unfortunately,
        > I need to check the file info
        > quite frequently. I was wondering if anyone knew what was the fast way
        > to get this info, LDAP,
        > SSH, etc?

        Wouldn't HTTP HEAD request achieve this fairly nicely ?

        I am not sure you have described the actual problem you are trying to solve.
        Why is that information needed and how is it being used?

        --
        Philippe M. Chiasson m/gozer\@(apache|cpan|ectoplasm)\.org/ GPG KeyID : 88C3A5A5
        http://gozer.ectoplasm.org/ F9BF E0C2 480E 7680 1AE5 3631 CB32 A107 88C3A5A5
      • Boysenberry Payne
        The server system is hosted with a third party, so I was hoping I could use mod_perl as a solution without resorting to involving them in the solution if
        Message 3 of 16 , Aug 1, 2005
        • 0 Attachment
          The server system is hosted with a third party, so I was hoping I could
          use
          mod_perl as a solution without resorting to involving them in the
          solution if
          possible. If NFS ends up being the best solution I will ask them
          if they could set it up.

          Thanks,
          Boysenberry

          This message contains information that is confidential
          and proprietary to Humaniteque and / or its affiliates.
          It is intended only for the recipient named and for
          the express purpose(s) described therein.
          Any other use is prohibited.

          http://www.habitatlife.com
          The World's Best Site Builder
          On Aug 1, 2005, at 4:28 PM, Gedanken wrote:

          > On Mon, 1 Aug 2005, Philip M. Gollucci wrote:
          >
          > There is a running joke in my office that, no matter what the problem
          > is,
          > I simply blame NFS before hearing any details. I am correct a
          > surprising
          > amount of the time =)
          >
          > One quick caveat - properly unmount volumes when rebooting and such.
          > Due
          > to one of those things that somehow grew beyond its original intent, we
          > had a network of about 15 machines all mounting each other. NFS
          > chokes when a mount it expects to be there, isnt. it takes it several
          > minutes to give up. One machine rebooting in such a big spiderweb
          > cluster
          > can cause massive problems without proper attention to cleanly
          > unmounting
          > and remounting shares. And 'cascading' is usually such a lovely
          > word...
          >
          > gedanken
          >
          > You might try an NFS mount between the too.
          >
          > mount_nfs -L server:/path /local/path
          > (FreeBSD)
          >
          >
          > --
          > gedanken
          >
          >
        • Philip M. Gollucci
          ... Specifically what information do you need from this file server ? directories, path names, file names, mod_times, access_time times, size..... etc ?
          Message 4 of 16 , Aug 1, 2005
          • 0 Attachment
            Boysenberry Payne wrote:
            > The server system is hosted with a third party, so I was hoping I could use
            > mod_perl as a solution without resorting to involving them in the
            > solution if
            > possible. If NFS ends up being the best solution I will ask them
            > if they could set it up.
            Specifically what information do you need from this file server ?

            directories, path names, file names, mod_times, access_time times,
            size..... etc ?
          • Boysenberry Payne
            I m not sure if HEAD would work. Basically, I m trying to read a directory s files. After I confirm a file exists and doesn t have zero size I check that it
            Message 5 of 16 , Aug 1, 2005
            • 0 Attachment
              I'm not sure if HEAD would work.
              Basically, I'm trying to read a directory's files.
              After I confirm a file exists and doesn't have zero
              size I check that it has the appropriate extension
              for the directory then I add the directory address,
              file name and extension to a table in our database.

              It used to be easy when I was using php on a single
              server system (which is what I'm migrating from.) Now
              that I am using a two server system it's a little tricker.

              Thanks,
              Boysenberry

              This message contains information that is confidential
              and proprietary to Humaniteque and / or its affiliates.
              It is intended only for the recipient named and for
              the express purpose(s) described therein.
              Any other use is prohibited.

              http://www.habitatlife.com
              The World's Best Site Builder
              On Aug 1, 2005, at 5:12 PM, Philippe M. Chiasson wrote:

              > Boysenberry Payne wrote:
              >> Hello All,
              >>
              >> I've got a two server platform one a static server for files and
              >> runs the mysql server
              >> and the other runs mod_perl. I'm trying to figure out the fastest way
              >> to get info on directories
              >> and files from the static server to the mod_perl server. Right now
              >> I'm
              >> using Net::FTP which
              >> is really slow, especially when they're a lot of files.
              >> Unfortunately,
              >> I need to check the file info
              >> quite frequently. I was wondering if anyone knew what was the fast
              >> way
              >> to get this info, LDAP,
              >> SSH, etc?
              >
              > Wouldn't HTTP HEAD request achieve this fairly nicely ?
              >
              > I am not sure you have described the actual problem you are trying to
              > solve.
              > Why is that information needed and how is it being used?
              >
              > --
              > Philippe M. Chiasson m/gozer\@(apache|cpan|ectoplasm)\.org/ GPG KeyID
              > : 88C3A5A5
              > http://gozer.ectoplasm.org/ F9BF E0C2 480E 7680 1AE5 3631 CB32
              > A107 88C3A5A5
            • Philip M. Gollucci
              ... We actually do something very similar to this involving pictures being uploaded from a digital camera to eventually be published on a website. Cronjob1:
              Message 6 of 16 , Aug 1, 2005
              • 0 Attachment
                Boysenberry Payne wrote:
                > I'm not sure if HEAD would work.
                > Basically, I'm trying to read a directory's files.
                > After I confirm a file exists and doesn't have zero
                > size I check that it has the appropriate extension
                > for the directory then I add the directory address,
                > file name and extension to a table in our database.
                We actually do something very similar to this involving pictures being
                uploaded from a digital camera to eventually be published on a website.

                Cronjob1:
                Poll destination directory and move the files to a temp location
                The destination directory is where the camera puts them.

                Cronjob2:
                Poll temp directory and move image into permenent
                location andinsert a row into our "images" table.

                Its split only so that if some part breaks the uploading from camera's
                does not and people can continue to upload from camera's. Digital
                camera's [at least the ones the government uses :)] upload with the same
                non-unique file names for each upload, so we have to process each batch
                rather quickly.

                I didn't write this, but I can say in 3 years its only crashed once and
                makes us millions.

                [snipped for breviety of course]
                Cronjob1:

                use Net::FTP;
                my $ftp = Net::FTP->new($PEERADDR, Debug => 0, Timeout => 30)
                || die "Connect to server failed\n";
                $ftp->login($USERNAME, $PASSWORD)
                || die "Cannot login to FTP server\n";
                $ftp->binary();

                my @files = $ftp->ls('-R');
                foreach my $file (@files) {
                unless some critera
                $ftp->get("$dir/$file", $localFilename);
                }
                $ftp->quit();
              • Boysenberry Payne
                I ve already got it working using Net::FTP. The problem is it runs slow using FTP. Here is an example of what I m trying to do: my $h = $ftp- {handle};
                Message 7 of 16 , Aug 1, 2005
                • 0 Attachment
                  I've already got it working using Net::FTP. The problem is it runs
                  slow using FTP. Here is an example of what I'm trying to do:

                  my $h = $ftp->{handle};
                  foreach my $directory ( @directories ) {
                  $h->cwd( $directory ) or die "can't change to directory: $directory
                  $!";
                  my $dir_ls = $h->ls;
                  foreach my $file_name ( @$dir_ls ) {
                  unless ( substr( $file_name, 0, 1 ) eq "." ) {
                  my $dir_nfo = $h->dir( $directory . $file_name );
                  $_ = $dir_nfo->[ 0 ];
                  s/(\s)+/ /g;
                  my @file_nfo = split / /, $_;
                  my $file_size = $file_nfo[ 4 ];
                  if( $file_size != 0 ) {
                  add to database
                  }
                  }
                  }
                  }
                  $h->quit;

                  I tried using ftp->size( $directory . $file_name );
                  But it seems to only return a size for small files,
                  at least on my OSX box.

                  Thanks,
                  Boysenberry

                  This message contains information that is confidential
                  and proprietary to Humaniteque and / or its affiliates.
                  It is intended only for the recipient named and for
                  the express purpose(s) described therein.
                  Any other use is prohibited.

                  http://www.habitatlife.com
                  The World's Best Site Builder
                  On Aug 1, 2005, at 5:54 PM, Philip M. Gollucci wrote:

                  > Boysenberry Payne wrote:
                  >> I'm not sure if HEAD would work.
                  >> Basically, I'm trying to read a directory's files.
                  >> After I confirm a file exists and doesn't have zero
                  >> size I check that it has the appropriate extension
                  >> for the directory then I add the directory address,
                  >> file name and extension to a table in our database.
                  > We actually do something very similar to this involving pictures being
                  > uploaded from a digital camera to eventually be published on a
                  > website.
                  >
                  > Cronjob1:
                  > Poll destination directory and move the files to a temp location
                  > The destination directory is where the camera puts them.
                  >
                  > Cronjob2:
                  > Poll temp directory and move image into permenent
                  > location andinsert a row into our "images" table.
                  >
                  > Its split only so that if some part breaks the uploading from camera's
                  > does not and people can continue to upload from camera's. Digital
                  > camera's [at least the ones the government uses :)] upload with the
                  > same non-unique file names for each upload, so we have to process each
                  > batch rather quickly.
                  >
                  > I didn't write this, but I can say in 3 years its only crashed once
                  > and makes us millions.
                  >
                  > [snipped for breviety of course]
                  > Cronjob1:
                  >
                  > use Net::FTP;
                  > my $ftp = Net::FTP->new($PEERADDR, Debug => 0, Timeout => 30)
                  > || die "Connect to server failed\n";
                  > $ftp->login($USERNAME, $PASSWORD)
                  > || die "Cannot login to FTP server\n";
                  > $ftp->binary();
                  >
                  > my @files = $ftp->ls('-R');
                  > foreach my $file (@files) {
                  > unless some critera
                  > $ftp->get("$dir/$file", $localFilename);
                  > }
                  > $ftp->quit();
                  >
                  >
                • Randy Kobes
                  ... Can you get someone on the remote server to do a cd top_level_directory ls -lR ls-lR # or find -fls find-ls gzip ls-lR # or gzip find-ls
                  Message 8 of 16 , Aug 1, 2005
                  • 0 Attachment
                    On Mon, 1 Aug 2005, Boysenberry Payne wrote:

                    > I'm not sure if HEAD would work.
                    > Basically, I'm trying to read a directory's files.
                    > After I confirm a file exists and doesn't have zero
                    > size I check that it has the appropriate extension
                    > for the directory then I add the directory address,
                    > file name and extension to a table in our database.

                    Can you get someone on the remote server to do a
                    cd top_level_directory
                    ls -lR > ls-lR # or find -fls find-ls
                    gzip ls-lR # or gzip find-ls
                    periodically, and then you can grab and parse ls-lR.gz or
                    find-ls.gz?

                    --
                    best regards,
                    randy kobes
                  • Philip M. Gollucci
                    ... Whats this ftp- {handle} stuff ? Shouldn t it just be $ftp- xxx Thats not in perldoc Net::FTP. Are you using a relatively new version of it ? We ve got
                    Message 9 of 16 , Aug 1, 2005
                    • 0 Attachment
                      Boysenberry Payne wrote:
                      > my $h = $ftp->{handle};
                      > foreach my $directory ( @directories ) {
                      > $h->cwd( $directory ) or die "can't change to directory: $directory
                      > $!";
                      > my $dir_ls = $h->ls;
                      > foreach my $file_name ( @$dir_ls ) {
                      > unless ( substr( $file_name, 0, 1 ) eq "." ) {
                      > my $dir_nfo = $h->dir( $directory . $file_name );
                      > $_ = $dir_nfo->[ 0 ];
                      > s/(\s)+/ /g;
                      > my @file_nfo = split / /, $_;
                      > my $file_size = $file_nfo[ 4 ];
                      > if( $file_size != 0 ) {
                      > add to database
                      > }
                      > }
                      > }
                      > }
                      > $h->quit;
                      Whats this ftp->{handle} stuff ?
                      Shouldn't it just be $ftp->xxx
                      Thats not in perldoc Net::FTP.

                      Are you using a relatively new version of it ?
                      We've got Net::FTP 2.75,
                      Linux wickedwitch 2.6.12.3 #2 SMP Mon Jul 18 17:14:55 EDT 2005 i686 i686
                      i386 GNU/Linux

                      What is the file size check for .. why do you have files of size 0 ?

                      I think it might be faster to do

                      next unless $file_name =~ /^\./o;

                      instead of
                      > unless ( substr( $file_name, 0, 1 ) eq "." ) {

                      It might be your database connection... Do you prepare the handle
                      outside of the loop? Is the database connect/disconnect outside of the
                      loop? What are you inserting into the database ? If you're inserting a
                      BLOB of the image/file data it could be the bandwidth transfer now thats
                      its not a local socket anymore.

                      Luck

                      --
                      END
                      -----------------------------------------------------------------------------
                      Philip M. Gollucci
                      Senior Developer - Liquidity Services Inc.
                      Phone: 202.558.6268 (Direct)
                      Cell: 301.254.5198
                      E-Mail: pgollucci@...
                      Web: http://www.liquidityservicesinc.com
                      http://www.liquidation.com
                      http://www.uksurplus.com
                      http://www.govliquidation.com
                      http://www.gowholesale.com
                    • Boysenberry Payne
                      $ftp- {handle} = Net::FTP- new( $ftp- {host}, Passive = 1 ) or die Can t create new ftp with host: $ftp- {host} ; It s part of my FTP module Thanks,
                      Message 10 of 16 , Aug 1, 2005
                      • 0 Attachment
                        $ftp->{handle} = Net::FTP->new( $ftp->{host}, Passive => 1 ) or die
                        "Can't create new ftp with host: $ftp->{host}";
                        It's part of my FTP module

                        Thanks,
                        Boysenberry

                        This message contains information that is confidential
                        and proprietary to Humaniteque and / or its affiliates.
                        It is intended only for the recipient named and for
                        the express purpose(s) described therein.
                        Any other use is prohibited.

                        http://www.habitatlife.com
                        The World's Best Site Builder
                        On Aug 1, 2005, at 6:31 PM, Philip M. Gollucci wrote:

                        > Boysenberry Payne wrote:
                        >> my $h = $ftp->{handle};
                        >> foreach my $directory ( @directories ) {
                        >> $h->cwd( $directory ) or die "can't change to directory:
                        >> $directory $!";
                        >> my $dir_ls = $h->ls;
                        >> foreach my $file_name ( @$dir_ls ) {
                        >> unless ( substr( $file_name, 0, 1 ) eq "." ) {
                        >> my $dir_nfo = $h->dir( $directory . $file_name );
                        >> $_ = $dir_nfo->[ 0 ];
                        >> s/(\s)+/ /g;
                        >> my @file_nfo = split / /, $_;
                        >> my $file_size = $file_nfo[ 4 ];
                        >> if( $file_size != 0 ) {
                        >> add to database
                        >> }
                        >> }
                        >> }
                        >> }
                        >> $h->quit;
                        > Whats this ftp->{handle} stuff ?
                        > Shouldn't it just be $ftp->xxx
                        > Thats not in perldoc Net::FTP.
                        >
                        > Are you using a relatively new version of it ?
                        > We've got Net::FTP 2.75,
                        > Linux wickedwitch 2.6.12.3 #2 SMP Mon Jul 18 17:14:55 EDT 2005 i686
                        > i686 i386 GNU/Linux
                        >
                        > What is the file size check for .. why do you have files of size 0 ?
                        >
                        > I think it might be faster to do
                        >
                        > next unless $file_name =~ /^\./o;
                        >
                        > instead of
                        > > unless ( substr( $file_name, 0, 1 ) eq "." ) {
                        >
                        > It might be your database connection... Do you prepare the handle
                        > outside of the loop? Is the database connect/disconnect outside of
                        > the loop? What are you inserting into the database ? If you're
                        > inserting a BLOB of the image/file data it could be the bandwidth
                        > transfer now thats its not a local socket anymore.
                        >
                        > Luck
                        >
                        > --
                        > END
                        > -----------------------------------------------------------------------
                        > ------
                        > Philip M. Gollucci
                        > Senior Developer - Liquidity Services Inc.
                        > Phone: 202.558.6268 (Direct)
                        > Cell: 301.254.5198
                        > E-Mail: pgollucci@...
                        > Web: http://www.liquidityservicesinc.com
                        > http://www.liquidation.com
                        > http://www.uksurplus.com
                        > http://www.govliquidation.com
                        > http://www.gowholesale.com
                        >
                        >
                        >
                      • Goddard Lee
                        ... I trust you are putting thousands into perl, mod_perl and other good things, then ;)
                        Message 11 of 16 , Aug 2, 2005
                        • 0 Attachment
                          > From: Philip M. Gollucci [mailto:pgollucci@...]

                          > I didn't write this, but ...[it]... makes us millions.

                          I trust "you" are putting thousands into perl, mod_perl and other good
                          things, then ;)
                        • Torsten Foertsch
                          ... mod_dav may be an option. Torsten
                          Message 12 of 16 , Aug 2, 2005
                          • 0 Attachment
                            On Monday 01 August 2005 23:12, Boysenberry Payne wrote:
                            > Hello All,
                            >
                            > I've got a two server platform one a static server for files and runs
                            > the mysql server
                            > and the other runs mod_perl. I'm trying to figure out the fastest way
                            > to get info on directories
                            > and files from the static server to the mod_perl server. Right now I'm
                            > using Net::FTP which
                            > is really slow, especially when they're a lot of files. Unfortunately,
                            > I need to check the file info
                            > quite frequently. I was wondering if anyone knew what was the fast way
                            > to get this info, LDAP,
                            > SSH, etc?

                            mod_dav may be an option.

                            Torsten
                          • Boysenberry Payne
                            Thank You Everyone, I think now that I know I can use $ftp- ls( -lR ), which I couldn t find anywhere in the Net::FTP docs or other O Reilly books I have, I
                            Message 13 of 16 , Aug 2, 2005
                            • 0 Attachment
                              Thank You Everyone,

                              I think now that I know I can use $ftp->ls( "-lR" ), which I couldn't
                              find
                              anywhere in the Net::FTP docs or other O'Reilly books I have, I can
                              stick to Net::FTP without is being slow. What was causing my script
                              to take so long was the multiple $ftp->cwd( $directory ), $ftp->ls() and
                              $ftp->dir( $directory . $file ) calls for each directory in my
                              directory loop.

                              Now I use one cwd and ls("-lR") from my public html area then process
                              the return array, which is a lot faster. It would be nice to be able
                              to specify
                              the directory as well as the "-lR" without using cwd( $directory ); does
                              anyone know how to do it?

                              Thanks for the tips on making my code more efficient too.

                              Boysenberry

                              This message contains information that is confidential
                              and proprietary to Humaniteque and / or its affiliates.
                              It is intended only for the recipient named and for
                              the express purpose(s) described therein.
                              Any other use is prohibited.

                              http://www.habitatlife.com
                              The World's Best Site Builder
                              On Aug 1, 2005, at 6:28 PM, Randy Kobes wrote:

                              > On Mon, 1 Aug 2005, Boysenberry Payne wrote:
                              >
                              >> I'm not sure if HEAD would work.
                              >> Basically, I'm trying to read a directory's files.
                              >> After I confirm a file exists and doesn't have zero
                              >> size I check that it has the appropriate extension
                              >> for the directory then I add the directory address,
                              >> file name and extension to a table in our database.
                              >
                              > Can you get someone on the remote server to do a
                              > cd top_level_directory
                              > ls -lR > ls-lR # or find -fls find-ls
                              > gzip ls-lR # or gzip find-ls
                              > periodically, and then you can grab and parse ls-lR.gz or find-ls.gz?
                              >
                              > --
                              > best regards,
                              > randy kobes
                              >
                              >
                            Your message has been successfully submitted and would be delivered to recipients shortly.