2 # Copyright (C) 2015-2020 all contributors <meta@public-inbox.org>
3 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
4 # Basic tool to create a Xapian search index for a public-inbox.
5 # Usage with libeatmydata <https://www.flamingspork.com/projects/libeatmydata/>
6 # highly recommended: eatmydata public-inbox-index INBOX_DIR
10 use Getopt::Long qw(:config gnu_getopt no_ignore_case auto_abbrev);
11 my $usage = 'public-inbox-index [options] INBOX_DIR';
12 my $help = <<EOF; # the following should fit w/o scrolling in 80x24 term:
15 Create and update search indices
19 --no-fsync speed up indexing, risk corruption on power outage
20 --indexlevel=LEVEL `basic', 'medium', or `full' (default: full)
21 --compact | -c run public-inbox-compact(1) after indexing
22 --sequential-shard index Xapian shards sequentially for slow storage
23 --jobs=NUM set or disable parallelization (NUM=0)
24 --batch-size=BYTES flush changes to OS after a given number of bytes
25 --max-size=BYTES do not index messages larger than the given size
26 --reindex index previously indexed data (if upgrading)
27 --rethread regenerate thread IDs (if upgrading, use sparingly)
28 --prune prune git storage on discontiguous history
29 --verbose | -v increase verbosity (may be repeated)
30 --help | -? show this help
32 BYTES may use `k', `m', and `g' suffixes (e.g. `10m' for 10 megabytes)
33 See public-inbox-index(1) man page for full documentation.
36 my $opt = { quiet => -1, compact => 0, maxsize => undef, fsync => 1 };
37 GetOptions($opt, qw(verbose|v+ reindex rethread compact|c+ jobs|j=i prune
38 fsync|sync! xapianonly|xapian-only
39 indexlevel|L=s maxsize|max-size=s batchsize|batch-size=s
40 sequentialshard|seq-shard|sequential-shard
42 or die "bad command-line args\n$usage";
43 if ($opt->{help}) { print $help; exit 0 };
44 die "--jobs must be >= 0\n" if defined $opt->{jobs} && $opt->{jobs} < 0;
46 # require lazily to speed up --help
47 require PublicInbox::Admin;
48 PublicInbox::Admin::require_or_die('-index');
50 if ($opt->{compact}) {
51 require PublicInbox::Xapcmd;
52 PublicInbox::Xapcmd::check_compact();
53 $compact_opt = { -coarse_lock => 1, compact => 1 };
54 if (defined(my $jobs = $opt->{jobs})) {
55 $compact_opt->{jobs} = $jobs;
59 my $cfg = PublicInbox::Config->new; # Config is loaded by Admin
60 my @ibxs = PublicInbox::Admin::resolve_inboxes(\@ARGV, undef, $cfg);
61 PublicInbox::Admin::require_or_die('-index');
62 unless (@ibxs) { print STDERR "Usage: $usage\n"; exit 1 }
64 my $max_size = $opt->{maxsize} // $cfg->{lc('publicInbox.indexMaxSize')};
65 if (defined $max_size) {
66 PublicInbox::Admin::parse_unsigned(\$max_size) or
67 die "`publicInbox.indexMaxSize=$max_size' not parsed\n";
70 my $bs = $opt->{batchsize} // $cfg->{lc('publicInbox.indexBatchSize')};
72 PublicInbox::Admin::parse_unsigned(\$bs) or
73 die "`publicInbox.indexBatchSize=$bs' not parsed\n";
76 local $PublicInbox::SearchIdx::BATCH_BYTES = $bs if defined($bs);
79 # out-of-the-box builds of Xapian 1.4.x are still limited to 32-bit
80 # https://getting-started-with-xapian.readthedocs.io/en/latest/concepts/indexing/limitations.html
81 local $ENV{XAPIAN_FLUSH_THRESHOLD} ||= '4294967295' if defined($bs);
83 my $s = $opt->{sequentialshard} //
84 $cfg->{lc('publicInbox.indexSequentialShard')};
86 my $v = $cfg->git_bool($s);
88 die "`publicInbox.indexSequentialShard=$s' not boolean\n";
89 $opt->{sequentialshard} = $v;
93 foreach my $ibx (@ibxs) {
94 # XXX: users can shoot themselves in the foot, with opt->{indexlevel}
95 $ibx->{indexlevel} //= $opt->{indexlevel} // ($opt->{xapianonly} ?
96 'full' : PublicInbox::Admin::detect_indexlevel($ibx));
97 $ibx->{index_max_size} = $max_size;
98 PublicInbox::Admin::scan_ibx_modules($mods, $ibx);
101 PublicInbox::Admin::require_or_die(keys %$mods);
102 require PublicInbox::InboxWritable;
103 PublicInbox::Admin::progress_prepare($opt);
104 for my $ibx (@ibxs) {
105 $ibx = PublicInbox::InboxWritable->new($ibx);
106 if ($opt->{compact} >= 2) {
107 PublicInbox::Xapcmd::run($ibx, 'compact', $compact_opt);
109 $ibx->{-no_fsync} = 1 if !$opt->{fsync};
112 if (defined(my $s = $ibx->{indexsequentialshard})) {
113 defined(my $v = $cfg->git_bool($s)) or die <<EOL;
114 publicInbox.$ibx->{name}.indexSequentialShard not boolean
116 $ibx_opt = { %$opt, sequentialshard => $v };
118 PublicInbox::Admin::index_inbox($ibx, undef, $ibx_opt);
120 local $compact_opt->{jobs} = 0 if $ibx_opt->{sequentialshard};
121 PublicInbox::Xapcmd::run($ibx, 'compact', $compact_opt);