1 =head1 HOW TO MIRROR CPAN
3 As of this writing the single purpose for this distribution is to
4 provide a backbone for the CPAN.
6 The idea is to get a handful backbone nodes that mirror directly from
7 PAUSE and a second layer that mirrors from them. Targetting at a
8 mirroring interval of 20 seconds.
10 The rsync daemon on PAUSE runs on port 8732 and you need a password to
11 access it. There is a second rsync daemon running on the standard port
12 873 but it has very limited access policies to leave the bandwidth for
15 If you have username and password you can mirror directly from PAUSE.
16 If you haven't and maintain a public CPAN mirror, ask me for one.
17 Otherwise pick your best CPAN mirror with an rsync server and try the
18 same script below for getting the quickest update cycle possible.
20 You find a list of potential rsync servers at
21 http://cpan.perl.org/SITES.html#RSYNC
23 The first thing you should prepare is the CPAN tree itself on your
24 disk. The source where you take it from does not matter that much.
25 Take it from where you always took it. The setup I suggest is to
26 mirror authors/ and modules/ with this program. And once per day or
27 per week or so do a traditional rsync from that old school mirroring
28 host, maybe excluding authors/ and modules/.
30 The loop is something like this:
32 $ENV{USER}="sandy"; # fill in your name
33 $ENV{RSYNC_PASSWORD} = "secret"; # fill in you passwd
35 use File::Rsync::Mirror::Recent;
37 File::Rsync::Mirror::Recent->new
39 ignore_link_stat_errors => 1,
40 localroot => "/home/ftp/pub/PAUSE/$_", # your local path
41 remote => "pause.perl.org::PAUSE/$_/RECENT.recent", # your upstream
42 max_files_per_connection => 863,
45 port => 8732, # only for PAUSE
51 tempdir => "/home/ftp/tmp", # optional tempdir to hide temporary files
54 )} "authors", "modules";
55 die "directory $_ doesn't exist, giving up" for grep { ! -d $_->localroot } @rrr;
57 my $ttgo = time + 1200; # pick less if you have a password and/or consent from the upstream
59 $rrr->rmirror ( "skip-deletes" => 1 );
61 my $sleep = $ttgo - time;
63 print STDERR "sleeping $sleep ... ";
70 Don't forget to fill in your user name and password.
72 You see the 'skip-deletes' and guess, this is mirroring without doing
73 deletes. You can do it with deletes or you can leave the deletion to
74 the other, the traditional rsync loop. Worry about that later.
76 You see the option 'max_files_per_connection' which I filled with 863
77 and you need to find your favorite number. The larger the number the
78 faster the whole download goes. As you already have a mirror, you
79 probably want to take 10 or 20 times times that much. CPAN is at the
80 moment (2009-03) consisting of 142000 files. This option chunks the
81 downloads into piecemeal units and between these units the process
82 takes a peek at the most recent files to download them with higher
85 The localroot parameter contains your target directory. Because only
86 the authors/ and the modules/ directory are hosted by the PAUSE, you
87 need the loop over two directories. The other directories of CPAN are
88 currently not available for downloading with
89 File::Rsync::Mirror::Recent.
91 The 'port' is only needed for users who have a password for PAUSE,
92 other rsync servers listen on the rsync server port and the option can
95 The timeslice in the while loop above needs to be large enough to let
96 the rsync server survive. If you choose a random rsync server and are
97 not an rsync server yourself please be modest and choose 1200. Choose
98 less if you're offering rsync yourself and have a fat pipe, and
99 especially if you know your upstream can manage it. If you have a
100 PAUSE password choose 20 seconds. We will watch how well this works
101 and will adjust this recommendation according to our findings.
103 Set the key 'verbose' to 0 if you have no debugging demands. In this
104 case you want to omit the 'sleeping $sleep ...' noise further down the
107 You can leave everything else as it stands. Start experimenting and
108 let me know how it works out.