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, $i);
89 sub progress_prepare ($) {
92 open my $null, '>', '/dev/null' or
93 die "failed to open /dev/null: $!\n";
94 $opt->{1} = fileno($null);
95 $opt->{-dev_null} = $null;
97 $opt->{-progress} = sub { print STDERR @_ };
101 sub same_fs_or_die ($$) {
103 return if ((stat($x))[0] == (stat($y))[0]); # 0 - st_dev
104 die "$x and $y reside on different filesystems\n";
108 my ($queue, $cb, $max, $opt) = @_;
110 while (defined(my $args = shift @$queue)) {
119 while (scalar(keys(%pids)) < $max && scalar(@$queue)) {
120 my $args = shift @$queue;
121 $pids{cb_spawn($cb, $args, $opt)} = $args;
124 while (scalar keys %pids) {
125 my $pid = waitpid(-1, 0);
126 my $args = delete $pids{$pid};
127 die join(' ', @$args)." failed: $?\n" if $?;
133 my ($ibx, $task, $opt) = @_; # task = 'cpdb' or 'compact'
134 my $cb = \&${\"PublicInbox::Xapcmd::$task"};
135 progress_prepare($opt ||= {});
136 my $dir = $ibx->{mainrepo} or die "no mainrepo in inbox\n";
137 runnable_or_die($XAPIAN_COMPACT) if $opt->{compact};
138 my $reindex; # v1:{ from => $x40 }, v2:{ from => [ $x40, $x40, .. ] } }
139 my $from; # per-epoch ranges
141 if (!$opt->{-coarse_lock}) {
142 $reindex = $opt->{reindex} = {};
143 $from = $reindex->{from} = [];
144 require Search::Xapian::WritableDatabase;
148 my $old = $ibx->search->xdir(1);
149 -d $old or die "$old does not exist\n";
151 my $tmp = PublicInbox::Xtmpdirs->new;
152 my $v = $ibx->{version} ||= 1;
155 # we want temporary directories to be as deep as possible,
156 # so v2 partitions can keep "xap$SCHEMA_VERSION" on a separate FS.
158 my $old_parent = dirname($old);
159 same_fs_or_die($old_parent, $old);
160 $tmp->{$old} = tempdir('xapcmd-XXXXXXXX', DIR => $old_parent);
161 push @q, [ $old, $tmp->{$old} ];
163 opendir my $dh, $old or die "Failed to opendir $old: $!\n";
164 while (defined(my $dn = readdir($dh))) {
165 if ($dn =~ /\A\d+\z/) {
166 my $tmpl = "$dn-XXXXXXXX";
167 my $dst = tempdir($tmpl, DIR => $old);
168 same_fs_or_die($old, $dst);
169 my $cur = "$old/$dn";
170 push @q, [ $cur, $dst ];
172 } elsif ($dn eq '.' || $dn eq '..') {
173 } elsif ($dn =~ /\Aover\.sqlite3/) {
175 warn "W: skipping unknown dir: $old/$dn\n"
178 die "No Xapian parts found in $old\n" unless @q;
180 my $im = $ibx->importer(0);
181 my $max = $opt->{jobs} || scalar(@q);
182 $ibx->with_umask(sub {
185 # fine-grained locking if we prepare for reindex
186 if (!$opt->{-coarse_lock}) {
187 prepare_reindex($ibx, $reindex);
191 delete($ibx->{$_}) for (qw(mm over search)); # cleanup
192 process_queue(\@q, $cb, $max, $opt);
193 commit_changes($ibx, $tmp, $opt);
197 sub cpdb_retryable ($$) {
198 my ($src, $pfx) = @_;
199 if (ref($@) eq 'Search::Xapian::DatabaseModifiedError') {
200 warn "$pfx Xapian DB modified, reopening and retrying\n";
205 warn "$pfx E: ", ref($@), "\n";
211 sub progress_pfx ($) {
212 my @p = split('/', $_[0]);
214 # return "xap15/0" for v2, or "xapian15" for v1:
215 ($p[-1] =~ /\A\d+\z/) ? "$p[-2]/$p[-1]" : $p[-1];
218 # xapian-compact wrapper
220 my ($args, $opt) = @_;
221 my ($src, $dst) = @$args;
223 my $pfx = $opt->{-progress_pfx} ||= progress_pfx($src);
224 my $pr = $opt->{-progress};
227 foreach my $fd (0..2) {
228 defined(my $dfd = $opt->{$fd}) or next;
231 $rdr->{1} = fileno($w) if $pr && pipe($r, $w);
233 # we rely on --no-renumber to keep docids synched to NNTP
234 my $cmd = [ $XAPIAN_COMPACT, '--no-renumber' ];
235 for my $sw (qw(no-full fuller)) {
236 push @$cmd, "--$sw" if $opt->{$sw};
238 for my $sw (qw(blocksize)) {
239 defined(my $v = $opt->{$sw}) or next;
240 push @$cmd, "--$sw", $v;
242 $pr->("$pfx `".join(' ', @$cmd)."'\n") if $pr;
243 push @$cmd, $src, $dst;
244 my $pid = spawn($cmd, undef, $rdr);
246 close $w or die "close: \$w: $!";
252 my $rp = waitpid($pid, 0);
253 if ($? || $rp != $pid) {
254 die join(' ', @$cmd)." failed: $? (pid=$pid, reaped=$rp)\n";
258 # Like copydatabase(1), this is horribly slow; and it doesn't seem due
259 # to the overhead of Perl.
261 my ($args, $opt) = @_;
262 my ($old, $new) = @$args;
263 my $src = Search::Xapian::Database->new($old);
265 if ($opt->{compact}) {
266 my $newdir = dirname($new);
267 same_fs_or_die($newdir, $new);
268 $tmp = tempdir("$new.compact-XXXXXX", DIR => $newdir);
269 $xtmp = PublicInbox::Xtmpdirs->new;
270 $xtmp->{$new} = $tmp;
275 # like copydatabase(1), be sure we don't overwrite anything in case
277 my $creat = Search::Xapian::DB_CREATE();
278 my $dst = Search::Xapian::WritableDatabase->new($tmp, $creat);
280 my ($nr, $tot, $fmt); # progress output
281 my $pr = $opt->{-progress};
282 my $pfx = $opt->{-progress_pfx} = progress_pfx($old);
286 # update the only metadata key for v1:
287 my $lc = $src->get_metadata('last_commit');
288 $dst->set_metadata('last_commit', $lc) if $lc;
290 $it = $src->postlist_begin('');
291 $end = $src->postlist_end('');
294 $tot = $src->get_doccount;
295 $fmt = "$pfx % ".length($tot)."u/$tot\n";
296 $pr->("$pfx copying $tot documents\n");
299 } while (cpdb_retryable($src, $pfx));
303 while ($it != $end) {
304 my $docid = $it->get_docid;
305 my $doc = $src->get_document($docid);
306 $dst->replace_document($docid, $doc);
308 if ($pr && !(++$nr & 1023)) {
309 $pr->(sprintf($fmt, $nr));
313 # unlike copydatabase(1), we don't copy spelling
314 # and synonym data (or other user metadata) since
315 # the Perl APIs don't expose iterators for them
316 # (and public-inbox does not use those features)
318 } while (cpdb_retryable($src, $pfx));
320 $pr->(sprintf($fmt, $nr)) if $pr;
323 $src = $dst = undef; # flushes and closes
325 # this is probably the best place to do xapian-compact
326 # since $dst isn't readable by HTTP or NNTP clients, yet:
327 compact([ $tmp, $new ], $opt);
328 remove_tree($tmp) or die "failed to remove $tmp: $!\n";
332 # slightly easier-to-manage manage than END{} blocks
333 package PublicInbox::Xtmpdirs;
336 use File::Path qw(remove_tree);
340 # http://www.tldp.org/LDP/abs/html/exitcodes.html
341 $SIG{INT} = sub { exit(130) };
342 $SIG{HUP} = $SIG{PIPE} = $SIG{TERM} = sub { exit(1) };
343 my $self = bless {}, $_[0]; # old partition => new (tmp) partition
344 $owner{"$self"} = $$;
350 delete $owner{"$self"};
353 $known_pids{$_}++ foreach values %owner;
354 if (!$known_pids{$$}) {
355 $SIG{INT} = $SIG{HUP} = $SIG{PIPE} = $SIG{TERM} = 'DEFAULT';
362 my $owner_pid = delete $owner{"$self"} or return;
363 return if $owner_pid != $$;
364 foreach my $new (values %$self) {
365 remove_tree($new) unless -d "$new/old";