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 ($$$$) {
20 my ($ibx, $im, $tmp, $opt) = @_;
21 my $reshard = $opt->{reshard};
22 my $reindex = $opt->{reindex};
24 $SIG{INT} or die 'BUG: $SIG{INT} not handled';
27 while (my ($old, $newdir) = each %$tmp) {
28 next if $old eq ''; # no invalid paths
30 if (!@st && !defined($opt->{reshard})) {
31 die "failed to stat($old): $!";
34 my $new = $newdir->dirname if defined($newdir);
35 my $over = "$old/over.sqlite3";
36 if (-f $over) { # only for v1, v2 over is untouched
37 defined $new or die "BUG: $over exists when culling v2";
38 $over = PublicInbox::Over->new($over);
39 my $tmp_over = "$new/over.sqlite3";
40 $over->connect->sqlite_backup_to_file($tmp_over);
44 if (!defined($new)) { # culled shard
45 push @old_shard, $old;
50 chmod($st[2] & 07777, $new) or die "chmod $old: $!\n";
51 rename($old, "$new/old") or
52 die "rename $old => $new/old: $!\n";
54 # Xtmpdir->DESTROY won't remove $new after this:
55 rename($new, $old) or die "rename $new => $old: $!\n";
57 my $prev = "$old/old";
59 die "failed to remove $prev: $!\n";
62 remove_tree(@old_shard);
64 if (!$opt->{-coarse_lock}) {
65 $opt->{-skip_lock} = 1;
67 if ($im->can('count_shards')) {
68 my $pr = $opt->{-progress};
69 my $n = $im->count_shards;
70 if (defined $reshard && $n != $reshard) {
72 "BUG: counted $n shards after resharding to $reshard";
74 my $prev = $im->{shards};
75 if ($pr && $prev != $n) {
76 $pr->("shard count changed: $prev => $n\n");
81 PublicInbox::Admin::index_inbox($ibx, $im, $opt);
86 my ($cb, $args, $opt) = @_; # $cb = cpdb() or compact()
87 defined(my $pid = fork) or die "fork: $!";
88 return $pid if $pid > 0;
93 sub runnable_or_die ($) {
95 which($exe) or die "$exe not found in PATH\n";
98 sub prepare_reindex ($$$) {
99 my ($ibx, $im, $reindex) = @_;
100 if ($ibx->version == 1) {
101 my $dir = $ibx->search->xdir(1);
102 my $xdb = $PublicInbox::Search::X{Database}->new($dir);
103 if (my $lc = $xdb->get_metadata('last_commit')) {
104 $reindex->{from} = $lc;
108 $im->git_dir_latest(\$max) or return;
109 my $from = $reindex->{from};
111 my $v = PublicInbox::Search::SCHEMA_VERSION();
112 foreach my $i (0..$max) {
113 $from->[$i] = $mm->last_commit_xap($v, $i);
118 sub same_fs_or_die ($$) {
120 return if ((stat($x))[0] == (stat($y))[0]); # 0 - st_dev
121 die "$x and $y reside on different filesystems\n";
125 my ($queue, $cb, $opt) = @_;
126 my $max = $opt->{jobs} || scalar(@$queue);
128 while (defined(my $args = shift @$queue)) {
137 while (scalar(keys(%pids)) < $max && scalar(@$queue)) {
138 my $args = shift @$queue;
139 $pids{cb_spawn($cb, $args, $opt)} = $args;
142 while (scalar keys %pids) {
143 my $pid = waitpid(-1, 0);
144 my $args = delete $pids{$pid};
146 die join(' ', @$args)." failed: $?\n" if $?;
148 warn "unknown PID($pid) reaped: $?\n";
154 sub setup_signals () {
155 # http://www.tldp.org/LDP/abs/html/exitcodes.html
156 $SIG{INT} = sub { exit(130) };
157 $SIG{HUP} = $SIG{PIPE} = $SIG{TERM} = sub { exit(1) };
161 my ($ibx, $opt) = @_;
162 my $tmp = {}; # old shard dir => File::Temp->newdir object or undef
163 my @queue; # ([old//src,newdir]) - list of args for cpdb() or compact()
165 my $old = $ibx->search->xdir(1);
166 -d $old or die "$old does not exist\n";
167 my $reshard = $opt->{reshard};
168 if (defined $reshard && $reshard <= 0) {
169 die "--reshard must be a positive number\n";
172 # we want temporary directories to be as deep as possible,
173 # so v2 shards can keep "xap$SCHEMA_VERSION" on a separate FS.
174 if ($ibx->version == 1) {
175 if (defined $reshard) {
177 "--reshard=$reshard ignored for v1 $ibx->{inboxdir}\n";
179 my $dir = dirname($old);
180 same_fs_or_die($dir, $old);
181 my $v = PublicInbox::Search::SCHEMA_VERSION();
182 my $wip = File::Temp->newdir("xapian$v-XXXXXXXX", DIR => $dir);
184 push @queue, [ $old, $wip ];
186 opendir my $dh, $old or die "Failed to opendir $old: $!\n";
188 while (defined(my $dn = readdir($dh))) {
189 if ($dn =~ /\A[0-9]+\z/) {
190 push @old_shards, $dn;
191 } elsif ($dn eq '.' || $dn eq '..') {
192 } elsif ($dn =~ /\Aover\.sqlite3/) {
194 warn "W: skipping unknown dir: $old/$dn\n"
197 die "No Xapian shards found in $old\n" unless @old_shards;
199 my ($src, $max_shard);
200 if (!defined($reshard) || $reshard == scalar(@old_shards)) {
202 $max_shard = scalar(@old_shards) - 1;
205 $max_shard = $reshard - 1;
206 $src = [ map { "$old/$_" } @old_shards ];
208 foreach my $dn (0..$max_shard) {
209 my $tmpl = "$dn-XXXXXXXX";
210 my $wip = File::Temp->newdir($tmpl, DIR => $old);
211 same_fs_or_die($old, $wip->dirname);
212 my $cur = "$old/$dn";
213 push @queue, [ $src // $cur , $wip ];
216 # mark old shards to be unlinked
218 $tmp->{$_} ||= undef for @$src;
224 sub check_compact () { runnable_or_die($XAPIAN_COMPACT) }
227 my ($ibx, $task, $opt) = @_; # task = 'cpdb' or 'compact'
228 my $cb = \&${\"PublicInbox::Xapcmd::$task"};
229 PublicInbox::Admin::progress_prepare($opt ||= {});
230 defined(my $dir = $ibx->{inboxdir}) or die "no inboxdir defined\n";
231 -d $dir or die "inboxdir=$dir does not exist\n";
232 check_compact() if $opt->{compact};
233 my $reindex; # v1:{ from => $x40 }, v2:{ from => [ $x40, $x40, .. ] } }
235 if (!$opt->{-coarse_lock}) {
236 $reindex = $opt->{reindex} = {};
237 $reindex->{from} = []; # per-epoch ranges
238 require PublicInbox::SearchIdx;
239 PublicInbox::SearchIdx::load_xapian_writable();
245 $ibx->with_umask(sub {
246 my $im = $ibx->importer(0);
248 my ($tmp, $queue) = prepare_run($ibx, $opt);
250 # fine-grained locking if we prepare for reindex
251 if (!$opt->{-coarse_lock}) {
252 prepare_reindex($ibx, $im, $reindex);
257 process_queue($queue, $cb, $opt);
258 $im->lock_acquire if !$opt->{-coarse_lock};
259 commit_changes($ibx, $im, $tmp, $opt);
263 sub cpdb_retryable ($$) {
264 my ($src, $pfx) = @_;
265 if (ref($@) =~ /\bDatabaseModifiedError\b/) {
266 warn "$pfx Xapian DB modified, reopening and retrying\n";
271 warn "$pfx E: ", ref($@), "\n";
277 sub progress_pfx ($) {
278 my ($wip) = @_; # tempdir v2: ([0-9])+-XXXXXXXX
279 my @p = split('/', $wip);
281 # return "xap15/0" for v2, or "xapian15" for v1:
282 ($p[-1] =~ /\A([0-9]+)/) ? "$p[-2]/$1" : $p[-1];
285 # xapian-compact wrapper
287 my ($args, $opt) = @_;
288 my ($src, $newdir) = @$args;
289 my $dst = ref($newdir) ? $newdir->dirname : $newdir;
290 my $pfx = $opt->{-progress_pfx} ||= progress_pfx($src);
291 my $pr = $opt->{-progress};
294 foreach my $fd (0..2) {
295 defined(my $dfd = $opt->{$fd}) or next;
299 # we rely on --no-renumber to keep docids synched to NNTP
300 my $cmd = [ $XAPIAN_COMPACT, '--no-renumber' ];
301 for my $sw (qw(no-full fuller)) {
302 push @$cmd, "--$sw" if $opt->{$sw};
304 for my $sw (qw(blocksize)) {
305 defined(my $v = $opt->{$sw}) or next;
306 push @$cmd, "--$sw", $v;
308 $pr->("$pfx `".join(' ', @$cmd)."'\n") if $pr;
309 push @$cmd, $src, $dst;
310 my $rd = popen_rd($cmd, undef, $rdr);
317 close $rd or die join(' ', @$cmd)." failed: $?n";
320 sub cpdb_loop ($$$;$$) {
321 my ($src, $dst, $pr_data, $cur_shard, $reshard) = @_;
322 my ($pr, $fmt, $nr, $pfx);
324 $pr = $pr_data->{pr};
325 $fmt = $pr_data->{fmt};
326 $nr = \($pr_data->{nr});
327 $pfx = $pr_data->{pfx};
333 $it = $src->postlist_begin('');
334 $end = $src->postlist_end('');
336 } while (cpdb_retryable($src, $pfx));
340 for (; $it != $end; $it++) {
341 my $docid = $it->get_docid;
342 if (defined $reshard) {
343 my $dst_shard = $docid % $reshard;
344 next if $dst_shard != $cur_shard;
346 my $doc = $src->get_document($docid);
347 $dst->replace_document($docid, $doc);
348 if ($pr_data && !(++$$nr & 1023)) {
349 $pr->(sprintf($fmt, $$nr));
353 # unlike copydatabase(1), we don't copy spelling
354 # and synonym data (or other user metadata) since
355 # the Perl APIs don't expose iterators for them
356 # (and public-inbox does not use those features)
358 } while (cpdb_retryable($src, $pfx));
361 # Like copydatabase(1), this is horribly slow; and it doesn't seem due
362 # to the overhead of Perl.
364 my ($args, $opt) = @_;
365 my ($old, $newdir) = @$args;
366 my $new = $newdir->dirname;
367 my ($src, $cur_shard);
369 PublicInbox::SearchIdx::load_xapian_writable() or die;
370 my $XapianDatabase = $PublicInbox::Search::X{Database};
371 if (ref($old) eq 'ARRAY') {
372 ($cur_shard) = ($new =~ m!xap[0-9]+/([0-9]+)\b!);
373 defined $cur_shard or
374 die "BUG: could not extract shard # from $new";
375 $reshard = $opt->{reshard};
376 defined $reshard or die 'BUG: got array src w/o --reshard';
378 # resharding, M:N copy means have full read access
381 my $sub = $XapianDatabase->new($_);
382 $src->add_database($sub);
384 $src = $XapianDatabase->new($_);
388 $src = $XapianDatabase->new($old);
393 if ($opt->{compact}) {
394 my $dir = dirname($new);
395 same_fs_or_die($dir, $new);
396 $ft = File::Temp->newdir("$new.compact-XXXXXX", DIR => $dir);
403 # like copydatabase(1), be sure we don't overwrite anything in case
405 my $creat = eval($PublicInbox::Search::Xap.'::DB_CREATE()');
407 my $XapianWritableDatabase = $PublicInbox::Search::X{WritableDatabase};
408 my $dst = $XapianWritableDatabase->new($tmp, $creat);
409 my $pr = $opt->{-progress};
410 my $pfx = $opt->{-progress_pfx} = progress_pfx($new);
411 my $pr_data = { pr => $pr, pfx => $pfx, nr => 0 } if $pr;
415 # update the only metadata key for v1:
416 my $lc = $src->get_metadata('last_commit');
417 $dst->set_metadata('last_commit', $lc) if $lc;
419 # only the first xapian shard (0) gets 'indexlevel'
420 if ($new =~ m!(?:xapian[0-9]+|xap[0-9]+/0)\b!) {
421 my $l = $src->get_metadata('indexlevel');
422 if ($l eq 'medium') {
423 $dst->set_metadata('indexlevel', $l);
427 my $tot = $src->get_doccount;
429 # we can only estimate when resharding,
430 # because removed spam causes slight imbalance
432 if (defined $cur_shard && $reshard > 1) {
433 $tot = int($tot/$reshard);
436 my $fmt = "$pfx % ".length($tot)."u/$tot\n";
437 $pr->("$pfx copying $est$tot documents\n");
438 $pr_data->{fmt} = $fmt;
439 $pr_data->{total} = $tot;
442 } while (cpdb_retryable($src, $pfx));
444 if (defined $reshard) {
445 # we rely on document IDs matching NNTP article number,
446 # so we can't have the Xapian sharding DB support rewriting
447 # document IDs. Thus we iterate through each shard
451 my $old = $XapianDatabase->new($_);
452 cpdb_loop($old, $dst, $pr_data, $cur_shard, $reshard);
455 cpdb_loop($src, $dst, $pr_data);
458 $pr->(sprintf($pr_data->{fmt}, $pr_data->{nr})) if $pr;
459 return unless $opt->{compact};
461 $src = $dst = undef; # flushes and closes
463 # this is probably the best place to do xapian-compact
464 # since $dst isn't readable by HTTP or NNTP clients, yet:
465 compact([ $tmp, $new ], $opt);
466 remove_tree($tmp) or die "failed to remove $tmp: $!\n";