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 qw(nodatacow_dir);
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 ($$$$) {
20 my ($ibx, $im, $tmp, $opt) = @_;
21 my $reshard = $opt->{reshard};
22 my $reindex = $opt->{reindex};
24 $SIG{INT} or die 'BUG: $SIG{INT} not handled';
28 while (my ($old, $newdir) = each %$tmp) {
29 next if $old eq ''; # no invalid paths
31 if (!@st && !defined($opt->{reshard})) {
32 die "failed to stat($old): $!";
35 my $new = $newdir->dirname if defined($newdir);
36 my $over = "$old/over.sqlite3";
37 if (-f $over) { # only for v1, v2 over is untouched
38 defined $new or die "BUG: $over exists when culling v2";
39 $over = PublicInbox::Over->new($over);
40 my $tmp_over = "$new/over.sqlite3";
41 $over->connect->sqlite_backup_to_file($tmp_over);
46 if (!defined($new)) { # culled shard
47 push @old_shard, $old;
52 chmod($st[2] & 07777, $new) or die "chmod $old: $!\n";
53 rename($old, "$new/old") or
54 die "rename $old => $new/old: $!\n";
56 # Xtmpdir->DESTROY won't remove $new after this:
57 rename($new, $old) or die "rename $new => $old: $!\n";
59 my $prev = "$old/old";
61 die "failed to remove $prev: $!\n";
65 # trigger ->check_inodes in read-only daemons
66 syswrite($im->{lockfh}, '.') if $over_chg;
68 remove_tree(@old_shard);
70 if (!$opt->{-coarse_lock}) {
71 $opt->{-skip_lock} = 1;
73 if ($im->can('count_shards')) {
74 my $pr = $opt->{-progress};
75 my $n = $im->count_shards;
76 if (defined $reshard && $n != $reshard) {
78 "BUG: counted $n shards after resharding to $reshard";
80 my $prev = $im->{shards};
81 if ($pr && $prev != $n) {
82 $pr->("shard count changed: $prev => $n\n");
87 PublicInbox::Admin::index_inbox($ibx, $im, $opt);
92 my ($cb, $args, $opt) = @_; # $cb = cpdb() or compact()
93 defined(my $pid = fork) or die "fork: $!";
94 return $pid if $pid > 0;
99 sub runnable_or_die ($) {
101 which($exe) or die "$exe not found in PATH\n";
104 sub prepare_reindex ($$$) {
105 my ($ibx, $im, $reindex) = @_;
106 if ($ibx->version == 1) {
107 my $dir = $ibx->search->xdir(1);
108 my $xdb = $PublicInbox::Search::X{Database}->new($dir);
109 if (my $lc = $xdb->get_metadata('last_commit')) {
110 $reindex->{from} = $lc;
114 $im->git_dir_latest(\$max) or return;
115 my $from = $reindex->{from};
117 my $v = PublicInbox::Search::SCHEMA_VERSION();
118 foreach my $i (0..$max) {
119 $from->[$i] = $mm->last_commit_xap($v, $i);
124 sub same_fs_or_die ($$) {
126 return if ((stat($x))[0] == (stat($y))[0]); # 0 - st_dev
127 die "$x and $y reside on different filesystems\n";
131 my ($queue, $cb, $opt) = @_;
132 my $max = $opt->{jobs} // scalar(@$queue);
134 while (defined(my $args = shift @$queue)) {
143 while (scalar(keys(%pids)) < $max && scalar(@$queue)) {
144 my $args = shift @$queue;
145 $pids{cb_spawn($cb, $args, $opt)} = $args;
148 while (scalar keys %pids) {
149 my $pid = waitpid(-1, 0);
150 my $args = delete $pids{$pid};
152 die join(' ', @$args)." failed: $?\n" if $?;
154 warn "unknown PID($pid) reaped: $?\n";
160 sub setup_signals () {
161 # http://www.tldp.org/LDP/abs/html/exitcodes.html
162 $SIG{INT} = sub { exit(130) };
163 $SIG{HUP} = $SIG{PIPE} = $SIG{TERM} = sub { exit(1) };
167 my ($ibx, $opt) = @_;
168 my $tmp = {}; # old shard dir => File::Temp->newdir object or undef
169 my @queue; # ([old//src,newdir]) - list of args for cpdb() or compact()
171 my $old = $ibx->search->xdir(1);
172 -d $old or die "$old does not exist\n";
173 my $reshard = $opt->{reshard};
174 if (defined $reshard && $reshard <= 0) {
175 die "--reshard must be a positive number\n";
178 # we want temporary directories to be as deep as possible,
179 # so v2 shards can keep "xap$SCHEMA_VERSION" on a separate FS.
180 if ($ibx->version == 1) {
181 if (defined $reshard) {
183 "--reshard=$reshard ignored for v1 $ibx->{inboxdir}\n";
185 my $dir = dirname($old);
186 same_fs_or_die($dir, $old);
187 my $v = PublicInbox::Search::SCHEMA_VERSION();
188 my $wip = File::Temp->newdir("xapian$v-XXXXXXXX", DIR => $dir);
190 nodatacow_dir($wip->dirname);
191 push @queue, [ $old, $wip ];
193 opendir my $dh, $old or die "Failed to opendir $old: $!\n";
195 while (defined(my $dn = readdir($dh))) {
196 if ($dn =~ /\A[0-9]+\z/) {
197 push @old_shards, $dn;
198 } elsif ($dn eq '.' || $dn eq '..') {
199 } elsif ($dn =~ /\Aover\.sqlite3/) {
201 warn "W: skipping unknown dir: $old/$dn\n"
204 die "No Xapian shards found in $old\n" unless @old_shards;
206 my ($src, $max_shard);
207 if (!defined($reshard) || $reshard == scalar(@old_shards)) {
209 $max_shard = scalar(@old_shards) - 1;
212 $max_shard = $reshard - 1;
213 $src = [ map { "$old/$_" } @old_shards ];
215 foreach my $dn (0..$max_shard) {
216 my $tmpl = "$dn-XXXXXXXX";
217 my $wip = File::Temp->newdir($tmpl, DIR => $old);
218 same_fs_or_die($old, $wip->dirname);
219 my $cur = "$old/$dn";
220 push @queue, [ $src // $cur , $wip ];
221 nodatacow_dir($wip->dirname);
224 # mark old shards to be unlinked
226 $tmp->{$_} ||= undef for @$src;
232 sub check_compact () { runnable_or_die($XAPIAN_COMPACT) }
235 my ($ibx, $cb, $opt, $reindex) = @_;
236 my $im = $ibx->importer(0);
238 my ($tmp, $queue) = prepare_run($ibx, $opt);
240 # fine-grained locking if we prepare for reindex
241 if (!$opt->{-coarse_lock}) {
242 prepare_reindex($ibx, $im, $reindex);
247 process_queue($queue, $cb, $opt);
248 $im->lock_acquire if !$opt->{-coarse_lock};
249 commit_changes($ibx, $im, $tmp, $opt);
253 my ($ibx, $task, $opt) = @_; # task = 'cpdb' or 'compact'
254 my $cb = \&${\"PublicInbox::Xapcmd::$task"};
255 PublicInbox::Admin::progress_prepare($opt ||= {});
256 defined(my $dir = $ibx->{inboxdir}) or die "no inboxdir defined\n";
257 -d $dir or die "inboxdir=$dir does not exist\n";
258 check_compact() if $opt->{compact};
259 my $reindex; # v1:{ from => $x40 }, v2:{ from => [ $x40, $x40, .. ] } }
261 if (!$opt->{-coarse_lock}) {
262 $reindex = $opt->{reindex} = { # per-epoch ranges for v2
263 from => $ibx->version == 1 ? '' : [],
265 require PublicInbox::SearchIdx;
266 PublicInbox::SearchIdx::load_xapian_writable();
272 $ibx->with_umask(\&_run, $ibx, $cb, $opt, $reindex);
275 sub cpdb_retryable ($$) {
276 my ($src, $pfx) = @_;
277 if (ref($@) =~ /\bDatabaseModifiedError\b/) {
278 warn "$pfx Xapian DB modified, reopening and retrying\n";
283 warn "$pfx E: ", ref($@), "\n";
289 sub progress_pfx ($) {
290 my ($wip) = @_; # tempdir v2: ([0-9])+-XXXXXXXX
291 my @p = split('/', $wip);
293 # return "xap15/0" for v2, or "xapian15" for v1:
294 ($p[-1] =~ /\A([0-9]+)/) ? "$p[-2]/$1" : $p[-1];
297 # xapian-compact wrapper
299 my ($args, $opt) = @_;
300 my ($src, $newdir) = @$args;
301 my $dst = ref($newdir) ? $newdir->dirname : $newdir;
302 my $pfx = $opt->{-progress_pfx} ||= progress_pfx($src);
303 my $pr = $opt->{-progress};
306 foreach my $fd (0..2) {
307 defined(my $dfd = $opt->{$fd}) or next;
311 # we rely on --no-renumber to keep docids synched to NNTP
312 my $cmd = [ $XAPIAN_COMPACT, '--no-renumber' ];
313 for my $sw (qw(no-full fuller)) {
314 push @$cmd, "--$sw" if $opt->{$sw};
316 for my $sw (qw(blocksize)) {
317 defined(my $v = $opt->{$sw}) or next;
318 push @$cmd, "--$sw", $v;
320 $pr->("$pfx `".join(' ', @$cmd)."'\n") if $pr;
321 push @$cmd, $src, $dst;
322 my $rd = popen_rd($cmd, undef, $rdr);
329 close $rd or die join(' ', @$cmd)." failed: $?n";
332 sub cpdb_loop ($$$;$$) {
333 my ($src, $dst, $pr_data, $cur_shard, $reshard) = @_;
334 my ($pr, $fmt, $nr, $pfx);
336 $pr = $pr_data->{pr};
337 $fmt = $pr_data->{fmt};
338 $nr = \($pr_data->{nr});
339 $pfx = $pr_data->{pfx};
345 $it = $src->postlist_begin('');
346 $end = $src->postlist_end('');
348 } while (cpdb_retryable($src, $pfx));
352 for (; $it != $end; $it++) {
353 my $docid = $it->get_docid;
354 if (defined $reshard) {
355 my $dst_shard = $docid % $reshard;
356 next if $dst_shard != $cur_shard;
358 my $doc = $src->get_document($docid);
359 $dst->replace_document($docid, $doc);
360 if ($pr_data && !(++$$nr & 1023)) {
361 $pr->(sprintf($fmt, $$nr));
365 # unlike copydatabase(1), we don't copy spelling
366 # and synonym data (or other user metadata) since
367 # the Perl APIs don't expose iterators for them
368 # (and public-inbox does not use those features)
370 } while (cpdb_retryable($src, $pfx));
373 # Like copydatabase(1), this is horribly slow; and it doesn't seem due
374 # to the overhead of Perl.
376 my ($args, $opt) = @_;
377 my ($old, $newdir) = @$args;
378 my $new = $newdir->dirname;
379 my ($src, $cur_shard);
381 PublicInbox::SearchIdx::load_xapian_writable() or die;
382 my $XapianDatabase = $PublicInbox::Search::X{Database};
383 if (ref($old) eq 'ARRAY') {
384 ($cur_shard) = ($new =~ m!xap[0-9]+/([0-9]+)\b!);
385 defined $cur_shard or
386 die "BUG: could not extract shard # from $new";
387 $reshard = $opt->{reshard};
388 defined $reshard or die 'BUG: got array src w/o --reshard';
390 # resharding, M:N copy means have full read access
393 my $sub = $XapianDatabase->new($_);
394 $src->add_database($sub);
396 $src = $XapianDatabase->new($_);
400 $src = $XapianDatabase->new($old);
405 if ($opt->{compact}) {
406 my $dir = dirname($new);
407 same_fs_or_die($dir, $new);
408 $ft = File::Temp->newdir("$new.compact-XXXXXX", DIR => $dir);
416 # like copydatabase(1), be sure we don't overwrite anything in case
418 my $flag = eval($PublicInbox::Search::Xap.'::DB_CREATE()');
420 my $XapianWritableDatabase = $PublicInbox::Search::X{WritableDatabase};
421 $flag |= $PublicInbox::SearchIdx::DB_NO_SYNC if !$opt->{sync};
422 my $dst = $XapianWritableDatabase->new($tmp, $flag);
423 my $pr = $opt->{-progress};
424 my $pfx = $opt->{-progress_pfx} = progress_pfx($new);
425 my $pr_data = { pr => $pr, pfx => $pfx, nr => 0 } if $pr;
429 # update the only metadata key for v1:
430 my $lc = $src->get_metadata('last_commit');
431 $dst->set_metadata('last_commit', $lc) if $lc;
433 # only the first xapian shard (0) gets 'indexlevel'
434 if ($new =~ m!(?:xapian[0-9]+|xap[0-9]+/0)\b!) {
435 my $l = $src->get_metadata('indexlevel');
436 if ($l eq 'medium') {
437 $dst->set_metadata('indexlevel', $l);
441 my $tot = $src->get_doccount;
443 # we can only estimate when resharding,
444 # because removed spam causes slight imbalance
446 if (defined $cur_shard && $reshard > 1) {
447 $tot = int($tot/$reshard);
450 my $fmt = "$pfx % ".length($tot)."u/$tot\n";
451 $pr->("$pfx copying $est$tot documents\n");
452 $pr_data->{fmt} = $fmt;
453 $pr_data->{total} = $tot;
456 } while (cpdb_retryable($src, $pfx));
458 if (defined $reshard) {
459 # we rely on document IDs matching NNTP article number,
460 # so we can't have the Xapian sharding DB support rewriting
461 # document IDs. Thus we iterate through each shard
465 my $old = $XapianDatabase->new($_);
466 cpdb_loop($old, $dst, $pr_data, $cur_shard, $reshard);
469 cpdb_loop($src, $dst, $pr_data);
472 $pr->(sprintf($pr_data->{fmt}, $pr_data->{nr})) if $pr;
473 return unless $opt->{compact};
475 $src = $dst = undef; # flushes and closes
477 # this is probably the best place to do xapian-compact
478 # since $dst isn't readable by HTTP or NNTP clients, yet:
479 compact([ $tmp, $new ], $opt);
480 remove_tree($tmp) or die "failed to remove $tmp: $!\n";