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;
5 use PublicInbox::Spawn qw(which popen_rd nodatacow_dir);
6 use PublicInbox::Admin qw(setup_signals);
8 use PublicInbox::SearchIdx;
9 use File::Temp 0.19 (); # ->newdir
10 use File::Path qw(remove_tree);
11 use File::Basename qw(dirname);
12 use POSIX qw(WNOHANG);
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 ($$$$) {
20 my ($ibx, $im, $tmp, $opt) = @_;
21 my $reshard = $opt->{reshard};
23 $SIG{INT} or die 'BUG: $SIG{INT} not handled';
27 while (my ($old, $newdir) = each %$tmp) {
28 next if $old eq ''; # no invalid paths
30 if (!@st && !defined($opt->{reshard})) {
31 die "failed to stat($old): $!";
34 my $new = $newdir->dirname if defined($newdir);
35 my $over = "$old/over.sqlite3";
36 if (-f $over) { # only for v1, v2 over is untouched
37 defined $new or die "BUG: $over exists when culling v2";
38 $over = PublicInbox::Over->new($over);
39 my $tmp_over = "$new/over.sqlite3";
40 $over->dbh->sqlite_backup_to_file($tmp_over);
45 if (!defined($new)) { # culled shard
46 push @old_shard, $old;
51 chmod($st[2] & 07777, $new) or die "chmod $old: $!\n";
52 rename($old, "$new/old") or
53 die "rename $old => $new/old: $!\n";
55 rename($new, $old) or die "rename $new => $old: $!\n";
57 my $prev = "$old/old";
59 die "failed to remove $prev: $!\n";
63 # trigger ->check_inodes in read-only daemons
64 syswrite($im->{lockfh}, '.') if $over_chg;
66 remove_tree(@old_shard);
68 if (!$opt->{-coarse_lock}) {
69 $opt->{-skip_lock} = 1;
71 if ($im->can('count_shards')) {
72 my $pr = $opt->{-progress};
73 my $n = $im->count_shards;
74 if (defined $reshard && $n != $reshard) {
76 "BUG: counted $n shards after resharding to $reshard";
78 my $prev = $im->{shards};
79 if ($pr && $prev != $n) {
80 $pr->("shard count changed: $prev => $n\n");
84 my $env = $opt->{-idx_env};
85 local %ENV = (%ENV, %$env) if $env;
86 PublicInbox::Admin::index_inbox($ibx, $im, $opt);
91 my ($cb, $args, $opt) = @_; # $cb = cpdb() or compact()
92 defined(my $pid = fork) or die "fork: $!";
93 return $pid if $pid > 0;
98 sub runnable_or_die ($) {
100 which($exe) or die "$exe not found in PATH\n";
103 sub prepare_reindex ($$$) {
104 my ($ibx, $im, $opt) = @_;
105 if ($ibx->version == 1) {
106 my $dir = $ibx->search->xdir(1);
107 my $xdb = $PublicInbox::Search::X{Database}->new($dir);
108 if (my $lc = $xdb->get_metadata('last_commit')) {
109 $opt->{reindex}->{from} = $lc;
112 my $max = $ibx->max_git_epoch // return;
113 my $from = $opt->{reindex}->{from};
115 my $v = PublicInbox::Search::SCHEMA_VERSION();
116 foreach my $i (0..$max) {
117 $from->[$i] = $mm->last_commit_xap($v, $i);
122 sub same_fs_or_die ($$) {
124 return if ((stat($x))[0] == (stat($y))[0]); # 0 - st_dev
125 die "$x and $y reside on different filesystems\n";
129 my ($sig, $pids) = @_;
130 kill($sig, keys %$pids); # pids may be empty
134 my ($queue, $cb, $opt) = @_;
135 my $max = $opt->{jobs} // scalar(@$queue);
137 while (defined(my $args = shift @$queue)) {
146 setup_signals(\&kill_pids, \%pids);
148 while (scalar(keys(%pids)) < $max && scalar(@$queue)) {
149 my $args = shift @$queue;
150 $pids{cb_spawn($cb, $args, $opt)} = $args;
154 while (scalar keys %pids) {
155 my $pid = waitpid(-1, $flags) or last;
157 my $args = delete $pids{$pid};
159 die join(' ', @$args)." failed: $?\n" if $?;
161 warn "unknown PID($pid) reaped: $?\n";
163 $flags = WNOHANG if scalar(@$queue);
169 my ($ibx, $opt) = @_;
170 my $tmp = {}; # old shard dir => File::Temp->newdir object or undef
171 my @queue; # ([old//src,newdir]) - list of args for cpdb() or compact()
173 if (my $srch = $ibx->search) {
174 $old = $srch->xdir(1);
175 -d $old or die "$old does not exist\n";
177 my $reshard = $opt->{reshard};
178 if (defined $reshard && $reshard <= 0) {
179 die "--reshard must be a positive number\n";
182 # we want temporary directories to be as deep as possible,
183 # so v2 shards can keep "xap$SCHEMA_VERSION" on a separate FS.
184 if ($old && $ibx->version == 1) {
185 if (defined $reshard) {
187 "--reshard=$reshard ignored for v1 $ibx->{inboxdir}\n";
189 my $dir = dirname($old);
190 same_fs_or_die($dir, $old);
191 my $v = PublicInbox::Search::SCHEMA_VERSION();
192 my $wip = File::Temp->newdir("xapian$v-XXXXXXXX", DIR => $dir);
194 nodatacow_dir($wip->dirname);
195 push @queue, [ $old, $wip ];
197 opendir my $dh, $old or die "Failed to opendir $old: $!\n";
199 while (defined(my $dn = readdir($dh))) {
200 if ($dn =~ /\A[0-9]+\z/) {
201 push @old_shards, $dn;
202 } elsif ($dn eq '.' || $dn eq '..') {
203 } elsif ($dn =~ /\Aover\.sqlite3/) {
205 warn "W: skipping unknown dir: $old/$dn\n"
208 die "No Xapian shards found in $old\n" unless @old_shards;
210 my ($src, $max_shard);
211 if (!defined($reshard) || $reshard == scalar(@old_shards)) {
213 $max_shard = scalar(@old_shards) - 1;
216 $max_shard = $reshard - 1;
217 $src = [ map { "$old/$_" } @old_shards ];
219 foreach my $dn (0..$max_shard) {
220 my $tmpl = "$dn-XXXXXXXX";
221 my $wip = File::Temp->newdir($tmpl, DIR => $old);
222 same_fs_or_die($old, $wip->dirname);
223 my $cur = "$old/$dn";
224 push @queue, [ $src // $cur , $wip ];
225 nodatacow_dir($wip->dirname);
228 # mark old shards to be unlinked
230 $tmp->{$_} ||= undef for @$src;
236 sub check_compact () { runnable_or_die($XAPIAN_COMPACT) }
239 my ($ibx, $cb, $opt) = @_;
240 my $im = $ibx->importer(0);
242 my ($tmp, $queue) = prepare_run($ibx, $opt);
244 # fine-grained locking if we prepare for reindex
245 if (!$opt->{-coarse_lock}) {
246 prepare_reindex($ibx, $im, $opt);
251 process_queue($queue, $cb, $opt);
252 $im->lock_acquire if !$opt->{-coarse_lock};
253 commit_changes($ibx, $im, $tmp, $opt);
257 my ($ibx, $task, $opt) = @_; # task = 'cpdb' or 'compact'
259 PublicInbox::Admin::progress_prepare($opt ||= {});
260 defined(my $dir = $ibx->{inboxdir}) or die "no inboxdir defined\n";
261 -d $dir or die "inboxdir=$dir does not exist\n";
262 check_compact() if $opt->{compact} && $ibx->search;
264 if (!$opt->{-coarse_lock}) {
265 # per-epoch ranges for v2
266 # v1:{ from => $OID }, v2:{ from => [ $OID, $OID, $OID ] } }
267 $opt->{reindex} = { from => $ibx->version == 1 ? '' : [] };
268 PublicInbox::SearchIdx::load_xapian_writable();
274 $ibx->with_umask(\&_run, $ibx, $cb, $opt);
277 sub cpdb_retryable ($$) {
278 my ($src, $pfx) = @_;
279 if (ref($@) =~ /\bDatabaseModifiedError\b/) {
280 warn "$pfx Xapian DB modified, reopening and retrying\n";
285 warn "$pfx E: ", ref($@), "\n";
291 sub progress_pfx ($) {
292 my ($wip) = @_; # tempdir v2: ([0-9])+-XXXXXXXX
293 my @p = split('/', $wip);
295 # return "xap15/0" for v2, or "xapian15" for v1:
296 ($p[-1] =~ /\A([0-9]+)/) ? "$p[-2]/$1" : $p[-1];
299 sub kill_compact { # setup_signals callback
300 my ($sig, $pidref) = @_;
301 kill($sig, $$pidref) if defined($$pidref);
304 # xapian-compact wrapper
306 my ($args, $opt) = @_;
307 my ($src, $newdir) = @$args;
308 my $dst = ref($newdir) ? $newdir->dirname : $newdir;
309 my $pfx = $opt->{-progress_pfx} ||= progress_pfx($src);
310 my $pr = $opt->{-progress};
313 foreach my $fd (0..2) {
314 defined(my $dfd = $opt->{$fd}) or next;
318 # we rely on --no-renumber to keep docids synched to NNTP
319 my $cmd = [ $XAPIAN_COMPACT, '--no-renumber' ];
320 for my $sw (qw(no-full fuller)) {
321 push @$cmd, "--$sw" if $opt->{$sw};
323 for my $sw (qw(blocksize)) {
324 defined(my $v = $opt->{$sw}) or next;
325 push @$cmd, "--$sw", $v;
327 $pr->("$pfx `".join(' ', @$cmd)."'\n") if $pr;
328 push @$cmd, $src, $dst;
331 setup_signals(\&kill_compact, \$pid);
332 ($rd, $pid) = popen_rd($cmd, undef, $rdr);
340 die "@$cmd failed: \$?=$?\n" if $?;
343 sub cpdb_loop ($$$;$$) {
344 my ($src, $dst, $pr_data, $cur_shard, $reshard) = @_;
345 my ($pr, $fmt, $nr, $pfx);
347 $pr = $pr_data->{pr};
348 $fmt = $pr_data->{fmt};
349 $nr = \($pr_data->{nr});
350 $pfx = $pr_data->{pfx};
356 $it = $src->postlist_begin('');
357 $end = $src->postlist_end('');
359 } while (cpdb_retryable($src, $pfx));
363 for (; $it != $end; $it++) {
364 my $docid = $it->get_docid;
365 if (defined $reshard) {
366 my $dst_shard = $docid % $reshard;
367 next if $dst_shard != $cur_shard;
369 my $doc = $src->get_document($docid);
370 $dst->replace_document($docid, $doc);
371 if ($pr_data && !(++$$nr & 1023)) {
372 $pr->(sprintf($fmt, $$nr));
376 # unlike copydatabase(1), we don't copy spelling
377 # and synonym data (or other user metadata) since
378 # the Perl APIs don't expose iterators for them
379 # (and public-inbox does not use those features)
381 } while (cpdb_retryable($src, $pfx));
384 # Like copydatabase(1), this is horribly slow; and it doesn't seem due
385 # to the overhead of Perl.
387 my ($args, $opt) = @_;
388 my ($old, $newdir) = @$args;
389 my $new = $newdir->dirname;
390 my ($src, $cur_shard);
392 PublicInbox::SearchIdx::load_xapian_writable() or die;
393 my $XapianDatabase = $PublicInbox::Search::X{Database};
394 if (ref($old) eq 'ARRAY') {
395 ($cur_shard) = ($new =~ m!xap[0-9]+/([0-9]+)\b!);
396 defined $cur_shard or
397 die "BUG: could not extract shard # from $new";
398 $reshard = $opt->{reshard};
399 defined $reshard or die 'BUG: got array src w/o --reshard';
401 # resharding, M:N copy means have full read access
404 my $sub = $XapianDatabase->new($_);
405 $src->add_database($sub);
407 $src = $XapianDatabase->new($_);
411 $src = $XapianDatabase->new($old);
416 if ($opt->{compact}) {
417 my $dir = dirname($new);
418 same_fs_or_die($dir, $new);
419 $ft = File::Temp->newdir("$new.compact-XXXXXX", DIR => $dir);
427 # like copydatabase(1), be sure we don't overwrite anything in case
429 my $flag = eval($PublicInbox::Search::Xap.'::DB_CREATE()');
431 my $XapianWritableDatabase = $PublicInbox::Search::X{WritableDatabase};
432 $flag |= $PublicInbox::SearchIdx::DB_NO_SYNC if !$opt->{fsync};
433 my $dst = $XapianWritableDatabase->new($tmp, $flag);
434 my $pr = $opt->{-progress};
435 my $pfx = $opt->{-progress_pfx} = progress_pfx($new);
436 my $pr_data = { pr => $pr, pfx => $pfx, nr => 0 } if $pr;
440 # update the only metadata key for v1:
441 my $lc = $src->get_metadata('last_commit');
442 $dst->set_metadata('last_commit', $lc) if $lc;
444 # only the first xapian shard (0) gets 'indexlevel'
445 if ($new =~ m!(?:xapian[0-9]+|xap[0-9]+/0)\b!) {
446 my $l = $src->get_metadata('indexlevel');
447 if ($l eq 'medium') {
448 $dst->set_metadata('indexlevel', $l);
452 my $tot = $src->get_doccount;
454 # we can only estimate when resharding,
455 # because removed spam causes slight imbalance
457 if (defined $cur_shard && $reshard > 1) {
458 $tot = int($tot/$reshard);
461 my $fmt = "$pfx % ".length($tot)."u/$tot\n";
462 $pr->("$pfx copying $est$tot documents\n");
463 $pr_data->{fmt} = $fmt;
464 $pr_data->{total} = $tot;
467 } while (cpdb_retryable($src, $pfx));
469 if (defined $reshard) {
470 # we rely on document IDs matching NNTP article number,
471 # so we can't have the Xapian sharding DB support rewriting
472 # document IDs. Thus we iterate through each shard
476 my $old = $XapianDatabase->new($_);
477 cpdb_loop($old, $dst, $pr_data, $cur_shard, $reshard);
480 cpdb_loop($src, $dst, $pr_data);
483 $pr->(sprintf($pr_data->{fmt}, $pr_data->{nr})) if $pr;
484 return unless $opt->{compact};
486 $src = $dst = undef; # flushes and closes
488 # this is probably the best place to do xapian-compact
489 # since $dst isn't readable by HTTP or NNTP clients, yet:
490 compact([ $tmp, $new ], $opt);
491 remove_tree($tmp) or die "failed to remove $tmp: $!\n";