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>
3 package PublicInbox::Xapcmd;
6 use PublicInbox::Spawn qw(which popen_rd);
8 use PublicInbox::SearchIdx;
9 use File::Temp 0.19 (); # ->newdir
10 use File::Path qw(remove_tree);
11 use File::Basename qw(dirname);
14 # support testing with dev versions of Xapian which installs
15 # commands with a version number suffix (e.g. "xapian-compact-1.5")
16 our $XAPIAN_COMPACT = $ENV{XAPIAN_COMPACT} || 'xapian-compact';
17 our @COMPACT_OPT = qw(jobs|j=i quiet|q blocksize|b=s no-full|n fuller|F);
19 sub commit_changes ($$$$) {
20 my ($ibx, $im, $tmp, $opt) = @_;
21 my $reshard = $opt->{reshard};
22 my $reindex = $opt->{reindex};
24 $SIG{INT} or die 'BUG: $SIG{INT} not handled';
27 while (my ($old, $newdir) = each %$tmp) {
28 next if $old eq ''; # no invalid paths
30 if (!@st && !defined($opt->{reshard})) {
31 die "failed to stat($old): $!";
34 my $new = $newdir->dirname if defined($newdir);
35 my $over = "$old/over.sqlite3";
36 if (-f $over) { # only for v1, v2 over is untouched
37 defined $new or die "BUG: $over exists when culling v2";
38 $over = PublicInbox::Over->new($over);
39 my $tmp_over = "$new/over.sqlite3";
40 $over->connect->sqlite_backup_to_file($tmp_over);
42 syswrite($im->{lockfh}, '.'); # trigger ->check_inodes
45 if (!defined($new)) { # culled shard
46 push @old_shard, $old;
51 chmod($st[2] & 07777, $new) or die "chmod $old: $!\n";
52 rename($old, "$new/old") or
53 die "rename $old => $new/old: $!\n";
55 # Xtmpdir->DESTROY won't remove $new after this:
56 rename($new, $old) or die "rename $new => $old: $!\n";
58 my $prev = "$old/old";
60 die "failed to remove $prev: $!\n";
63 remove_tree(@old_shard);
65 if (!$opt->{-coarse_lock}) {
66 $opt->{-skip_lock} = 1;
68 if ($im->can('count_shards')) {
69 my $pr = $opt->{-progress};
70 my $n = $im->count_shards;
71 if (defined $reshard && $n != $reshard) {
73 "BUG: counted $n shards after resharding to $reshard";
75 my $prev = $im->{shards};
76 if ($pr && $prev != $n) {
77 $pr->("shard count changed: $prev => $n\n");
82 PublicInbox::Admin::index_inbox($ibx, $im, $opt);
87 my ($cb, $args, $opt) = @_; # $cb = cpdb() or compact()
88 defined(my $pid = fork) or die "fork: $!";
89 return $pid if $pid > 0;
94 sub runnable_or_die ($) {
96 which($exe) or die "$exe not found in PATH\n";
99 sub prepare_reindex ($$$) {
100 my ($ibx, $im, $reindex) = @_;
101 if ($ibx->version == 1) {
102 my $dir = $ibx->search->xdir(1);
103 my $xdb = $PublicInbox::Search::X{Database}->new($dir);
104 if (my $lc = $xdb->get_metadata('last_commit')) {
105 $reindex->{from} = $lc;
109 $im->git_dir_latest(\$max) or return;
110 my $from = $reindex->{from};
112 my $v = PublicInbox::Search::SCHEMA_VERSION();
113 foreach my $i (0..$max) {
114 $from->[$i] = $mm->last_commit_xap($v, $i);
119 sub same_fs_or_die ($$) {
121 return if ((stat($x))[0] == (stat($y))[0]); # 0 - st_dev
122 die "$x and $y reside on different filesystems\n";
126 my ($queue, $cb, $opt) = @_;
127 my $max = $opt->{jobs} || scalar(@$queue);
129 while (defined(my $args = shift @$queue)) {
138 while (scalar(keys(%pids)) < $max && scalar(@$queue)) {
139 my $args = shift @$queue;
140 $pids{cb_spawn($cb, $args, $opt)} = $args;
143 while (scalar keys %pids) {
144 my $pid = waitpid(-1, 0);
145 my $args = delete $pids{$pid};
147 die join(' ', @$args)." failed: $?\n" if $?;
149 warn "unknown PID($pid) reaped: $?\n";
155 sub setup_signals () {
156 # http://www.tldp.org/LDP/abs/html/exitcodes.html
157 $SIG{INT} = sub { exit(130) };
158 $SIG{HUP} = $SIG{PIPE} = $SIG{TERM} = sub { exit(1) };
162 my ($ibx, $opt) = @_;
163 my $tmp = {}; # old shard dir => File::Temp->newdir object or undef
164 my @queue; # ([old//src,newdir]) - list of args for cpdb() or compact()
166 my $old = $ibx->search->xdir(1);
167 -d $old or die "$old does not exist\n";
168 my $reshard = $opt->{reshard};
169 if (defined $reshard && $reshard <= 0) {
170 die "--reshard must be a positive number\n";
173 # we want temporary directories to be as deep as possible,
174 # so v2 shards can keep "xap$SCHEMA_VERSION" on a separate FS.
175 if ($ibx->version == 1) {
176 if (defined $reshard) {
178 "--reshard=$reshard ignored for v1 $ibx->{inboxdir}\n";
180 my $dir = dirname($old);
181 same_fs_or_die($dir, $old);
182 my $v = PublicInbox::Search::SCHEMA_VERSION();
183 my $wip = File::Temp->newdir("xapian$v-XXXXXXXX", DIR => $dir);
185 push @queue, [ $old, $wip ];
187 opendir my $dh, $old or die "Failed to opendir $old: $!\n";
189 while (defined(my $dn = readdir($dh))) {
190 if ($dn =~ /\A[0-9]+\z/) {
191 push @old_shards, $dn;
192 } elsif ($dn eq '.' || $dn eq '..') {
193 } elsif ($dn =~ /\Aover\.sqlite3/) {
195 warn "W: skipping unknown dir: $old/$dn\n"
198 die "No Xapian shards found in $old\n" unless @old_shards;
200 my ($src, $max_shard);
201 if (!defined($reshard) || $reshard == scalar(@old_shards)) {
203 $max_shard = scalar(@old_shards) - 1;
206 $max_shard = $reshard - 1;
207 $src = [ map { "$old/$_" } @old_shards ];
209 foreach my $dn (0..$max_shard) {
210 my $tmpl = "$dn-XXXXXXXX";
211 my $wip = File::Temp->newdir($tmpl, DIR => $old);
212 same_fs_or_die($old, $wip->dirname);
213 my $cur = "$old/$dn";
214 push @queue, [ $src // $cur , $wip ];
217 # mark old shards to be unlinked
219 $tmp->{$_} ||= undef for @$src;
225 sub check_compact () { runnable_or_die($XAPIAN_COMPACT) }
228 my ($ibx, $task, $opt) = @_; # task = 'cpdb' or 'compact'
229 my $cb = \&${\"PublicInbox::Xapcmd::$task"};
230 PublicInbox::Admin::progress_prepare($opt ||= {});
231 defined(my $dir = $ibx->{inboxdir}) or die "no inboxdir defined\n";
232 -d $dir or die "inboxdir=$dir does not exist\n";
233 check_compact() if $opt->{compact};
234 my $reindex; # v1:{ from => $x40 }, v2:{ from => [ $x40, $x40, .. ] } }
236 if (!$opt->{-coarse_lock}) {
237 $reindex = $opt->{reindex} = {};
238 $reindex->{from} = []; # per-epoch ranges
239 require PublicInbox::SearchIdx;
240 PublicInbox::SearchIdx::load_xapian_writable();
246 $ibx->with_umask(sub {
247 my $im = $ibx->importer(0);
249 my ($tmp, $queue) = prepare_run($ibx, $opt);
251 # fine-grained locking if we prepare for reindex
252 if (!$opt->{-coarse_lock}) {
253 prepare_reindex($ibx, $im, $reindex);
258 process_queue($queue, $cb, $opt);
259 $im->lock_acquire if !$opt->{-coarse_lock};
260 commit_changes($ibx, $im, $tmp, $opt);
264 sub cpdb_retryable ($$) {
265 my ($src, $pfx) = @_;
266 if (ref($@) =~ /\bDatabaseModifiedError\b/) {
267 warn "$pfx Xapian DB modified, reopening and retrying\n";
272 warn "$pfx E: ", ref($@), "\n";
278 sub progress_pfx ($) {
279 my ($wip) = @_; # tempdir v2: ([0-9])+-XXXXXXXX
280 my @p = split('/', $wip);
282 # return "xap15/0" for v2, or "xapian15" for v1:
283 ($p[-1] =~ /\A([0-9]+)/) ? "$p[-2]/$1" : $p[-1];
286 # xapian-compact wrapper
288 my ($args, $opt) = @_;
289 my ($src, $newdir) = @$args;
290 my $dst = ref($newdir) ? $newdir->dirname : $newdir;
291 my $pfx = $opt->{-progress_pfx} ||= progress_pfx($src);
292 my $pr = $opt->{-progress};
295 foreach my $fd (0..2) {
296 defined(my $dfd = $opt->{$fd}) or next;
300 # we rely on --no-renumber to keep docids synched to NNTP
301 my $cmd = [ $XAPIAN_COMPACT, '--no-renumber' ];
302 for my $sw (qw(no-full fuller)) {
303 push @$cmd, "--$sw" if $opt->{$sw};
305 for my $sw (qw(blocksize)) {
306 defined(my $v = $opt->{$sw}) or next;
307 push @$cmd, "--$sw", $v;
309 $pr->("$pfx `".join(' ', @$cmd)."'\n") if $pr;
310 push @$cmd, $src, $dst;
311 my $rd = popen_rd($cmd, undef, $rdr);
318 close $rd or die join(' ', @$cmd)." failed: $?n";
321 sub cpdb_loop ($$$;$$) {
322 my ($src, $dst, $pr_data, $cur_shard, $reshard) = @_;
323 my ($pr, $fmt, $nr, $pfx);
325 $pr = $pr_data->{pr};
326 $fmt = $pr_data->{fmt};
327 $nr = \($pr_data->{nr});
328 $pfx = $pr_data->{pfx};
334 $it = $src->postlist_begin('');
335 $end = $src->postlist_end('');
337 } while (cpdb_retryable($src, $pfx));
341 for (; $it != $end; $it++) {
342 my $docid = $it->get_docid;
343 if (defined $reshard) {
344 my $dst_shard = $docid % $reshard;
345 next if $dst_shard != $cur_shard;
347 my $doc = $src->get_document($docid);
348 $dst->replace_document($docid, $doc);
349 if ($pr_data && !(++$$nr & 1023)) {
350 $pr->(sprintf($fmt, $$nr));
354 # unlike copydatabase(1), we don't copy spelling
355 # and synonym data (or other user metadata) since
356 # the Perl APIs don't expose iterators for them
357 # (and public-inbox does not use those features)
359 } while (cpdb_retryable($src, $pfx));
362 # Like copydatabase(1), this is horribly slow; and it doesn't seem due
363 # to the overhead of Perl.
365 my ($args, $opt) = @_;
366 my ($old, $newdir) = @$args;
367 my $new = $newdir->dirname;
368 my ($src, $cur_shard);
370 PublicInbox::SearchIdx::load_xapian_writable() or die;
371 my $XapianDatabase = $PublicInbox::Search::X{Database};
372 if (ref($old) eq 'ARRAY') {
373 ($cur_shard) = ($new =~ m!xap[0-9]+/([0-9]+)\b!);
374 defined $cur_shard or
375 die "BUG: could not extract shard # from $new";
376 $reshard = $opt->{reshard};
377 defined $reshard or die 'BUG: got array src w/o --reshard';
379 # resharding, M:N copy means have full read access
382 my $sub = $XapianDatabase->new($_);
383 $src->add_database($sub);
385 $src = $XapianDatabase->new($_);
389 $src = $XapianDatabase->new($old);
394 if ($opt->{compact}) {
395 my $dir = dirname($new);
396 same_fs_or_die($dir, $new);
397 $ft = File::Temp->newdir("$new.compact-XXXXXX", DIR => $dir);
404 # like copydatabase(1), be sure we don't overwrite anything in case
406 my $creat = eval($PublicInbox::Search::Xap.'::DB_CREATE()');
408 my $XapianWritableDatabase = $PublicInbox::Search::X{WritableDatabase};
409 my $dst = $XapianWritableDatabase->new($tmp, $creat);
410 my $pr = $opt->{-progress};
411 my $pfx = $opt->{-progress_pfx} = progress_pfx($new);
412 my $pr_data = { pr => $pr, pfx => $pfx, nr => 0 } if $pr;
416 # update the only metadata key for v1:
417 my $lc = $src->get_metadata('last_commit');
418 $dst->set_metadata('last_commit', $lc) if $lc;
420 # only the first xapian shard (0) gets 'indexlevel'
421 if ($new =~ m!(?:xapian[0-9]+|xap[0-9]+/0)\b!) {
422 my $l = $src->get_metadata('indexlevel');
423 if ($l eq 'medium') {
424 $dst->set_metadata('indexlevel', $l);
428 my $tot = $src->get_doccount;
430 # we can only estimate when resharding,
431 # because removed spam causes slight imbalance
433 if (defined $cur_shard && $reshard > 1) {
434 $tot = int($tot/$reshard);
437 my $fmt = "$pfx % ".length($tot)."u/$tot\n";
438 $pr->("$pfx copying $est$tot documents\n");
439 $pr_data->{fmt} = $fmt;
440 $pr_data->{total} = $tot;
443 } while (cpdb_retryable($src, $pfx));
445 if (defined $reshard) {
446 # we rely on document IDs matching NNTP article number,
447 # so we can't have the Xapian sharding DB support rewriting
448 # document IDs. Thus we iterate through each shard
452 my $old = $XapianDatabase->new($_);
453 cpdb_loop($old, $dst, $pr_data, $cur_shard, $reshard);
456 cpdb_loop($src, $dst, $pr_data);
459 $pr->(sprintf($pr_data->{fmt}, $pr_data->{nr})) if $pr;
460 return unless $opt->{compact};
462 $src = $dst = undef; # flushes and closes
464 # this is probably the best place to do xapian-compact
465 # since $dst isn't readable by HTTP or NNTP clients, yet:
466 compact([ $tmp, $new ], $opt);
467 remove_tree($tmp) or die "failed to remove $tmp: $!\n";