1 # Copyright (C) 2018-2019 all contributors <meta@public-inbox.org>
2 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
3 package PublicInbox::Xapcmd;
6 use PublicInbox::Spawn qw(which spawn);
8 use PublicInbox::Search;
9 use File::Temp qw(tempdir);
10 use File::Path qw(remove_tree);
11 use File::Basename qw(dirname);
13 # support testing with dev versions of Xapian which installs
14 # commands with a version number suffix (e.g. "xapian-compact-1.5")
15 our $XAPIAN_COMPACT = $ENV{XAPIAN_COMPACT} || 'xapian-compact';
17 sub commit_changes ($$$) {
18 my ($ibx, $tmp, $opt) = @_;
20 my $reindex = $opt->{reindex};
21 my $im = $ibx->importer(0);
22 $im->lock_acquire if !$opt->{-coarse_lock};
24 while (my ($old, $new) = each %$tmp) {
25 my @st = stat($old) or die "failed to stat($old): $!\n";
27 my $over = "$old/over.sqlite3";
28 if (-f $over) { # only for v1, v2 over is untouched
29 $over = PublicInbox::Over->new($over);
30 my $tmp_over = "$new/over.sqlite3";
31 $over->connect->sqlite_backup_to_file($tmp_over);
34 chmod($st[2] & 07777, $new) or die "chmod $old: $!\n";
36 # Xtmpdir->DESTROY won't remove $new after this:
37 rename($old, "$new/old") or die "rename $old => $new/old: $!\n";
38 rename($new, $old) or die "rename $new => $old: $!\n";
39 my $prev = "$old/old";
40 remove_tree($prev) or die "failed to remove $prev: $!\n";
43 if (!$opt->{-coarse_lock}) {
44 $opt->{-skip_lock} = 1;
45 PublicInbox::Admin::index_inbox($ibx, $opt);
46 # implicit lock_release
53 my ($cb, $args, $opt) = @_; # $cb = cpdb() or compact()
54 defined(my $pid = fork) or die "fork: $!";
55 return $pid if $pid > 0;
56 eval { $cb->($args, $opt) };
61 sub runnable_or_die ($) {
63 which($exe) or die "$exe not found in PATH\n";
66 sub prepare_reindex ($$) {
67 my ($ibx, $reindex) = @_;
68 if ($ibx->{version} == 1) {
69 my $dir = $ibx->search->xdir(1);
70 my $xdb = Search::Xapian::Database->new($dir);
71 if (my $lc = $xdb->get_metadata('last_commit')) {
72 $reindex->{from} = $lc;
75 my $v2w = $ibx->importer(0);
77 $v2w->git_dir_latest(\$max) or return;
78 my $from = $reindex->{from};
80 my $v = PublicInbox::Search::SCHEMA_VERSION();
81 foreach my $i (0..$max) {
82 $from->[$i] = $mm->last_commit_xap($v, $i);
87 sub progress_prepare ($) {
90 open my $null, '>', '/dev/null' or
91 die "failed to open /dev/null: $!\n";
92 $opt->{1} = fileno($null);
93 $opt->{-dev_null} = $null;
95 $opt->{-progress} = sub { print STDERR @_ };
99 sub same_fs_or_die ($$) {
101 return if ((stat($x))[0] == (stat($y))[0]); # 0 - st_dev
102 die "$x and $y reside on different filesystems\n";
106 my ($ibx, $task, $opt) = @_; # task = 'cpdb' or 'compact'
107 my $cb = \&${\"PublicInbox::Xapcmd::$task"};
108 progress_prepare($opt ||= {});
109 my $dir = $ibx->{mainrepo} or die "no mainrepo in inbox\n";
110 runnable_or_die($XAPIAN_COMPACT) if $opt->{compact};
111 my $reindex; # v1:{ from => $x40 }, v2:{ from => [ $x40, $x40, .. ] } }
112 my $from; # per-epoch ranges
114 if (!$opt->{-coarse_lock}) {
115 $reindex = $opt->{reindex} = {};
116 $from = $reindex->{from} = [];
117 require Search::Xapian::WritableDatabase;
121 my $old = $ibx->search->xdir(1);
122 -d $old or die "$old does not exist\n";
124 my $tmp = PublicInbox::Xtmpdirs->new;
125 my $v = $ibx->{version} ||= 1;
128 # we want temporary directories to be as deep as possible,
129 # so v2 partitions can keep "xap$SCHEMA_VERSION" on a separate FS.
131 my $old_parent = dirname($old);
132 same_fs_or_die($old_parent, $old);
133 $tmp->{$old} = tempdir('xapcmd-XXXXXXXX', DIR => $old_parent);
134 push @q, [ $old, $tmp->{$old} ];
136 opendir my $dh, $old or die "Failed to opendir $old: $!\n";
137 while (defined(my $dn = readdir($dh))) {
138 if ($dn =~ /\A\d+\z/) {
139 my $tmpl = "$dn-XXXXXXXX";
140 my $dst = tempdir($tmpl, DIR => $old);
141 same_fs_or_die($old, $dst);
142 my $cur = "$old/$dn";
143 push @q, [ $cur, $dst ];
145 } elsif ($dn eq '.' || $dn eq '..') {
146 } elsif ($dn =~ /\Aover\.sqlite3/) {
148 warn "W: skipping unknown dir: $old/$dn\n"
151 die "No Xapian parts found in $old\n" unless @q;
153 my $im = $ibx->importer(0);
154 my $max = $opt->{jobs} || scalar(@q);
155 $ibx->with_umask(sub {
158 # fine-grained locking if we prepare for reindex
159 if (!$opt->{-coarse_lock}) {
160 prepare_reindex($ibx, $reindex);
164 delete($ibx->{$_}) for (qw(mm over search)); # cleanup
167 while (scalar(keys(%pids)) < $max && scalar(@q)) {
169 $pids{cb_spawn($cb, $args, $opt)} = $args;
172 while (scalar keys %pids) {
173 my $pid = waitpid(-1, 0);
174 my $args = delete $pids{$pid};
175 die join(' ', @$args)." failed: $?\n" if $?;
178 commit_changes($ibx, $tmp, $opt);
182 sub cpdb_retryable ($$) {
183 my ($src, $pfx) = @_;
184 if (ref($@) eq 'Search::Xapian::DatabaseModifiedError') {
185 warn "$pfx Xapian DB modified, reopening and retrying\n";
190 warn "$pfx E: ", ref($@), "\n";
196 sub progress_pfx ($) {
197 my @p = split('/', $_[0]);
199 # return "xap15/0" for v2, or "xapian15" for v1:
200 ($p[-1] =~ /\A\d+\z/) ? "$p[-2]/$p[-1]" : $p[-1];
203 # xapian-compact wrapper
205 my ($args, $opt) = @_;
206 my ($src, $dst) = @$args;
208 my $pfx = $opt->{-progress_pfx} ||= progress_pfx($src);
209 my $pr = $opt->{-progress};
212 foreach my $fd (0..2) {
213 defined(my $dfd = $opt->{$fd}) or next;
217 $pr->("$pfx compacting...\n");
218 $rdr->{1} = fileno($w) if pipe($r, $w);
221 # we rely on --no-renumber to keep docids synched to NNTP
222 my $cmd = [ $XAPIAN_COMPACT, '--no-renumber', $src, $dst ];
223 my $pid = spawn($cmd, undef, $rdr);
225 close $w or die "close: \$w: $!";
231 my $rp = waitpid($pid, 0);
232 if ($? || $rp != $pid) {
233 die join(' ', @$cmd)." failed: $? (pid=$pid, reaped=$rp)\n";
237 # Like copydatabase(1), this is horribly slow; and it doesn't seem due
238 # to the overhead of Perl.
240 my ($args, $opt) = @_;
241 my ($old, $new) = @$args;
242 my $src = Search::Xapian::Database->new($old);
243 my $tmp = $opt->{compact} ? "$new.compact" : $new;
245 # like copydatabase(1), be sure we don't overwrite anything in case
247 my $creat = Search::Xapian::DB_CREATE();
248 my $dst = Search::Xapian::WritableDatabase->new($tmp, $creat);
250 my ($nr, $tot, $fmt); # progress output
251 my $pr = $opt->{-progress};
252 my $pfx = $opt->{-progress_pfx} = progress_pfx($old);
256 # update the only metadata key for v1:
257 my $lc = $src->get_metadata('last_commit');
258 $dst->set_metadata('last_commit', $lc) if $lc;
260 $it = $src->postlist_begin('');
261 $end = $src->postlist_end('');
264 $tot = $src->get_doccount;
265 $fmt = "$pfx % ".length($tot)."u/$tot\n";
266 $pr->("$pfx copying $tot documents\n");
269 } while (cpdb_retryable($src, $pfx));
273 while ($it != $end) {
274 my $docid = $it->get_docid;
275 my $doc = $src->get_document($docid);
276 $dst->replace_document($docid, $doc);
278 if ($pr && !(++$nr & 1023)) {
279 $pr->(sprintf($fmt, $nr));
283 # unlike copydatabase(1), we don't copy spelling
284 # and synonym data (or other user metadata) since
285 # the Perl APIs don't expose iterators for them
286 # (and public-inbox does not use those features)
288 } while (cpdb_retryable($src, $pfx));
290 $pr->(sprintf($fmt, $nr)) if $pr;
291 return unless $opt->{compact};
293 $src = $dst = undef; # flushes and closes
295 # this is probably the best place to do xapian-compact
296 # since $dst isn't readable by HTTP or NNTP clients, yet:
297 compact([ $tmp, $new ], $opt);
298 remove_tree($tmp) or die "failed to remove $tmp: $!\n";
301 # slightly easier-to-manage manage than END{} blocks
302 package PublicInbox::Xtmpdirs;
305 use File::Path qw(remove_tree);
309 # http://www.tldp.org/LDP/abs/html/exitcodes.html
310 $SIG{INT} = sub { exit(130) };
311 $SIG{HUP} = $SIG{PIPE} = $SIG{TERM} = sub { exit(1) };
312 my $self = bless {}, $_[0]; # old partition => new (tmp) partition
313 $owner{"$self"} = $$;
319 delete $owner{"$self"};
320 $SIG{INT} = $SIG{HUP} = $SIG{PIPE} = $SIG{TERM} = 'DEFAULT';
326 my $owner_pid = delete $owner{"$self"} or return;
327 return if $owner_pid != $$;
328 foreach my $new (values %$self) {
329 remove_tree($new) unless -d "$new/old";
331 $SIG{INT} = $SIG{HUP} = $SIG{PIPE} = $SIG{TERM} = 'DEFAULT';