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';
16 our @COMPACT_OPT = qw(jobs|j=i quiet|q blocksize|b=s no-full|n fuller|F);
18 sub commit_changes ($$$) {
19 my ($ibx, $tmp, $opt) = @_;
21 my $reindex = $opt->{reindex};
22 my $im = $ibx->importer(0);
23 $im->lock_acquire if !$opt->{-coarse_lock};
25 $SIG{INT} or die 'BUG: $SIG{INT} not handled';
27 while (my ($old, $new) = each %$tmp) {
28 my @st = stat($old) or die "failed to stat($old): $!\n";
30 my $over = "$old/over.sqlite3";
31 if (-f $over) { # only for v1, v2 over is untouched
32 $over = PublicInbox::Over->new($over);
33 my $tmp_over = "$new/over.sqlite3";
34 $over->connect->sqlite_backup_to_file($tmp_over);
37 chmod($st[2] & 07777, $new) or die "chmod $old: $!\n";
39 # Xtmpdir->DESTROY won't remove $new after this:
40 rename($old, "$new/old") or die "rename $old => $new/old: $!\n";
41 rename($new, $old) or die "rename $new => $old: $!\n";
42 my $prev = "$old/old";
43 remove_tree($prev) or die "failed to remove $prev: $!\n";
46 if (!$opt->{-coarse_lock}) {
47 $opt->{-skip_lock} = 1;
48 PublicInbox::Admin::index_inbox($ibx, $opt);
49 # implicit lock_release
56 my ($cb, $args, $opt) = @_; # $cb = cpdb() or compact()
57 defined(my $pid = fork) or die "fork: $!";
58 return $pid if $pid > 0;
63 sub runnable_or_die ($) {
65 which($exe) or die "$exe not found in PATH\n";
68 sub prepare_reindex ($$) {
69 my ($ibx, $reindex) = @_;
70 if ($ibx->{version} == 1) {
71 my $dir = $ibx->search->xdir(1);
72 my $xdb = Search::Xapian::Database->new($dir);
73 if (my $lc = $xdb->get_metadata('last_commit')) {
74 $reindex->{from} = $lc;
77 my $v2w = $ibx->importer(0);
79 $v2w->git_dir_latest(\$max) or return;
80 my $from = $reindex->{from};
82 my $v = PublicInbox::Search::SCHEMA_VERSION();
83 foreach my $i (0..$max) {
84 $from->[$i] = $mm->last_commit_xap($v, $i);
89 sub same_fs_or_die ($$) {
91 return if ((stat($x))[0] == (stat($y))[0]); # 0 - st_dev
92 die "$x and $y reside on different filesystems\n";
96 my ($queue, $cb, $max, $opt) = @_;
98 while (defined(my $args = shift @$queue)) {
107 while (scalar(keys(%pids)) < $max && scalar(@$queue)) {
108 my $args = shift @$queue;
109 $pids{cb_spawn($cb, $args, $opt)} = $args;
112 while (scalar keys %pids) {
113 my $pid = waitpid(-1, 0);
114 my $args = delete $pids{$pid};
115 die join(' ', @$args)." failed: $?\n" if $?;
121 my ($ibx, $task, $opt) = @_; # task = 'cpdb' or 'compact'
122 my $cb = \&${\"PublicInbox::Xapcmd::$task"};
123 PublicInbox::Admin::progress_prepare($opt ||= {});
124 my $dir = $ibx->{mainrepo} or die "no mainrepo in inbox\n";
125 runnable_or_die($XAPIAN_COMPACT) if $opt->{compact};
126 my $reindex; # v1:{ from => $x40 }, v2:{ from => [ $x40, $x40, .. ] } }
127 my $from; # per-epoch ranges
129 if (!$opt->{-coarse_lock}) {
130 $reindex = $opt->{reindex} = {};
131 $from = $reindex->{from} = [];
132 require Search::Xapian::WritableDatabase;
136 my $old = $ibx->search->xdir(1);
137 -d $old or die "$old does not exist\n";
139 my $tmp = PublicInbox::Xtmpdirs->new;
140 my $v = $ibx->{version} ||= 1;
143 # we want temporary directories to be as deep as possible,
144 # so v2 partitions can keep "xap$SCHEMA_VERSION" on a separate FS.
146 my $old_parent = dirname($old);
147 same_fs_or_die($old_parent, $old);
148 $tmp->{$old} = tempdir('xapcmd-XXXXXXXX', DIR => $old_parent);
149 push @q, [ $old, $tmp->{$old} ];
151 opendir my $dh, $old or die "Failed to opendir $old: $!\n";
152 while (defined(my $dn = readdir($dh))) {
153 if ($dn =~ /\A\d+\z/) {
154 my $tmpl = "$dn-XXXXXXXX";
155 my $dst = tempdir($tmpl, DIR => $old);
156 same_fs_or_die($old, $dst);
157 my $cur = "$old/$dn";
158 push @q, [ $cur, $dst ];
160 } elsif ($dn eq '.' || $dn eq '..') {
161 } elsif ($dn =~ /\Aover\.sqlite3/) {
163 warn "W: skipping unknown dir: $old/$dn\n"
166 die "No Xapian parts found in $old\n" unless @q;
168 my $im = $ibx->importer(0);
169 my $max = $opt->{jobs} || scalar(@q);
170 $ibx->with_umask(sub {
173 # fine-grained locking if we prepare for reindex
174 if (!$opt->{-coarse_lock}) {
175 prepare_reindex($ibx, $reindex);
179 delete($ibx->{$_}) for (qw(mm over search)); # cleanup
180 process_queue(\@q, $cb, $max, $opt);
181 commit_changes($ibx, $tmp, $opt);
185 sub cpdb_retryable ($$) {
186 my ($src, $pfx) = @_;
187 if (ref($@) eq 'Search::Xapian::DatabaseModifiedError') {
188 warn "$pfx Xapian DB modified, reopening and retrying\n";
193 warn "$pfx E: ", ref($@), "\n";
199 sub progress_pfx ($) {
200 my @p = split('/', $_[0]);
202 # return "xap15/0" for v2, or "xapian15" for v1:
203 ($p[-1] =~ /\A\d+\z/) ? "$p[-2]/$p[-1]" : $p[-1];
206 # xapian-compact wrapper
208 my ($args, $opt) = @_;
209 my ($src, $dst) = @$args;
211 my $pfx = $opt->{-progress_pfx} ||= progress_pfx($src);
212 my $pr = $opt->{-progress};
215 foreach my $fd (0..2) {
216 defined(my $dfd = $opt->{$fd}) or next;
219 $rdr->{1} = fileno($w) if $pr && pipe($r, $w);
221 # we rely on --no-renumber to keep docids synched to NNTP
222 my $cmd = [ $XAPIAN_COMPACT, '--no-renumber' ];
223 for my $sw (qw(no-full fuller)) {
224 push @$cmd, "--$sw" if $opt->{$sw};
226 for my $sw (qw(blocksize)) {
227 defined(my $v = $opt->{$sw}) or next;
228 push @$cmd, "--$sw", $v;
230 $pr->("$pfx `".join(' ', @$cmd)."'\n") if $pr;
231 push @$cmd, $src, $dst;
232 my $pid = spawn($cmd, undef, $rdr);
234 close $w or die "close: \$w: $!";
240 my $rp = waitpid($pid, 0);
241 if ($? || $rp != $pid) {
242 die join(' ', @$cmd)." failed: $? (pid=$pid, reaped=$rp)\n";
246 # Like copydatabase(1), this is horribly slow; and it doesn't seem due
247 # to the overhead of Perl.
249 my ($args, $opt) = @_;
250 my ($old, $new) = @$args;
251 my $src = Search::Xapian::Database->new($old);
253 if ($opt->{compact}) {
254 my $newdir = dirname($new);
255 same_fs_or_die($newdir, $new);
256 $tmp = tempdir("$new.compact-XXXXXX", DIR => $newdir);
257 $xtmp = PublicInbox::Xtmpdirs->new;
258 $xtmp->{$new} = $tmp;
263 # like copydatabase(1), be sure we don't overwrite anything in case
265 my $creat = Search::Xapian::DB_CREATE();
266 my $dst = Search::Xapian::WritableDatabase->new($tmp, $creat);
268 my ($nr, $tot, $fmt); # progress output
269 my $pr = $opt->{-progress};
270 my $pfx = $opt->{-progress_pfx} = progress_pfx($old);
274 # update the only metadata key for v1:
275 my $lc = $src->get_metadata('last_commit');
276 $dst->set_metadata('last_commit', $lc) if $lc;
278 # only the first xapian partition (0) gets 'indexlevel'
279 if ($old =~ m!(?:xapian\d+|xap\d+/0)\z!) {
280 my $l = $src->get_metadata('indexlevel');
281 if ($l eq 'medium') {
282 $dst->set_metadata('indexlevel', $l);
286 $it = $src->postlist_begin('');
287 $end = $src->postlist_end('');
290 $tot = $src->get_doccount;
291 $fmt = "$pfx % ".length($tot)."u/$tot\n";
292 $pr->("$pfx copying $tot documents\n");
295 } while (cpdb_retryable($src, $pfx));
299 while ($it != $end) {
300 my $docid = $it->get_docid;
301 my $doc = $src->get_document($docid);
302 $dst->replace_document($docid, $doc);
304 if ($pr && !(++$nr & 1023)) {
305 $pr->(sprintf($fmt, $nr));
309 # unlike copydatabase(1), we don't copy spelling
310 # and synonym data (or other user metadata) since
311 # the Perl APIs don't expose iterators for them
312 # (and public-inbox does not use those features)
314 } while (cpdb_retryable($src, $pfx));
316 $pr->(sprintf($fmt, $nr)) if $pr;
319 $src = $dst = undef; # flushes and closes
321 # this is probably the best place to do xapian-compact
322 # since $dst isn't readable by HTTP or NNTP clients, yet:
323 compact([ $tmp, $new ], $opt);
324 remove_tree($tmp) or die "failed to remove $tmp: $!\n";
328 # slightly easier-to-manage manage than END{} blocks
329 package PublicInbox::Xtmpdirs;
332 use File::Path qw(remove_tree);
336 # http://www.tldp.org/LDP/abs/html/exitcodes.html
337 $SIG{INT} = sub { exit(130) };
338 $SIG{HUP} = $SIG{PIPE} = $SIG{TERM} = sub { exit(1) };
339 my $self = bless {}, $_[0]; # old partition => new (tmp) partition
340 $owner{"$self"} = $$;
346 delete $owner{"$self"};
349 $known_pids{$_}++ foreach values %owner;
350 if (!$known_pids{$$}) {
351 $SIG{INT} = $SIG{HUP} = $SIG{PIPE} = $SIG{TERM} = 'DEFAULT';
358 my $owner_pid = delete $owner{"$self"} or return;
359 return if $owner_pid != $$;
360 foreach my $new (values %$self) {
361 remove_tree($new) unless -d "$new/old";