1 # Copyright (C) 2018-2020 all contributors <meta@public-inbox.org>
2 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
4 # used to interface with a single Xapian shard in V2 repos.
5 # See L<public-inbox-v2-format(5)> for more info on how we shard Xapian
6 package PublicInbox::SearchIdxShard;
9 use parent qw(PublicInbox::SearchIdx);
10 use IO::Handle (); # autoflush
14 my ($class, $v2w, $shard) = @_;
15 my $ibx = $v2w->{ibx};
16 my $self = $class->SUPER::new($ibx, 1, $shard);
17 # create the DB before forking:
19 $self->set_indexlevel;
21 $self->spawn_worker($v2w, $shard) if $v2w->{parallel};
26 my ($self, $v2w, $shard) = @_;
28 pipe($r, $w) or die "pipe failed: $!\n";
31 defined $pid or die "fork failed: $!\n";
33 my $bnote = $v2w->atfork_child;
34 close $w or die "failed to close: $!";
36 # F_SETPIPE_SZ = 1031 on Linux; increasing the pipe size here
37 # speeds V2Writable batch imports across 8 cores by nearly 20%
38 fcntl($r, 1031, 1048576) if $^O eq 'linux';
40 eval { shard_worker_loop($self, $v2w, $r, $shard, $bnote) };
41 die "worker $shard died: $@\n" if $@;
42 die "unexpected MM $self->{mm}" if $self->{mm};
47 close $r or die "failed to close: $!";
50 sub shard_worker_loop ($$$$$) {
51 my ($self, $v2w, $r, $shard, $bnote) = @_;
52 $0 = "pi-v2-shard[$shard]";
53 $self->begin_txn_lazy;
54 while (my $line = readline($r)) {
55 $v2w->{current_info} = "[$shard] $line";
56 if ($line eq "commit\n") {
57 $self->commit_txn_lazy;
58 } elsif ($line eq "close\n") {
60 } elsif ($line eq "barrier\n") {
61 $self->commit_txn_lazy;
62 # no need to lock < 512 bytes is atomic under POSIX
63 print $bnote "barrier $shard\n" or
64 die "write failed for barrier $!\n";
65 } elsif ($line =~ /\AD ([a-f0-9]{40,}) ([0-9]+)\n\z/s) {
66 $self->remove_by_oid($1, $2 + 0);
69 # n.b. $mid may contain spaces(!)
70 my ($to_read, $bytes, $num, $blob, $ds, $ts, $mid) =
72 $self->begin_txn_lazy;
73 my $n = read($r, my $msg, $to_read) or die "read: $!\n";
74 $n == $to_read or die "short read: $n != $to_read\n";
75 my $mime = PublicInbox::Eml->new(\$msg);
83 }, 'PublicInbox::Smsg';
84 $self->add_message($mime, $smsg);
90 # called by V2Writable
92 my ($self, $msgref, $mime, $smsg) = @_;
93 if (my $w = $self->{w}) {
94 # mid must be last, it can contain spaces (but not LF)
95 print $w join(' ', @$smsg{qw(raw_bytes bytes
96 num blob ds ts mid)}),
97 "\n", $$msgref or die "failed to write shard $!\n";
100 $self->begin_txn_lazy;
101 $self->add_message($mime, $smsg);
106 close $_[0]->{w} or die "failed to close write pipe: $!\n";
109 # called by V2Writable:
112 if (my $w = $self->{w}) {
113 print $w "barrier\n" or die "failed to print: $!";
115 $self->commit_txn_lazy;