1 |
#!/usr/bin/perl |
2 |
|
3 |
package Fuse::DBI; |
4 |
|
5 |
use 5.008; |
6 |
use strict; |
7 |
use warnings; |
8 |
|
9 |
use POSIX qw(ENOENT EISDIR EINVAL ENOSYS O_RDWR); |
10 |
use Fuse; |
11 |
use DBI; |
12 |
use Carp; |
13 |
use Data::Dumper; |
14 |
|
15 |
|
16 |
our $VERSION = '0.06'; |
17 |
|
18 |
=head1 NAME |
19 |
|
20 |
Fuse::DBI - mount your database as filesystem and use it |
21 |
|
22 |
=head1 SYNOPSIS |
23 |
|
24 |
use Fuse::DBI; |
25 |
Fuse::DBI->mount( ... ); |
26 |
|
27 |
See C<run> below for examples how to set parameters. |
28 |
|
29 |
=head1 DESCRIPTION |
30 |
|
31 |
This module will use C<Fuse> module, part of C<FUSE (Filesystem in USErspace)> |
32 |
available at L<http://fuse.sourceforge.net/> to mount |
33 |
your database as file system. |
34 |
|
35 |
That will give you possibility to use normal file-system tools (cat, grep, vi) |
36 |
to manipulate data in database. |
37 |
|
38 |
It's actually opposite of Oracle's intention to put everything into database. |
39 |
|
40 |
|
41 |
=head1 METHODS |
42 |
|
43 |
=cut |
44 |
|
45 |
=head2 mount |
46 |
|
47 |
Mount your database as filesystem. |
48 |
|
49 |
Let's suppose that your database have table C<files> with following structure: |
50 |
|
51 |
id: int |
52 |
filename: text |
53 |
size: int |
54 |
content: text |
55 |
writable: boolean |
56 |
|
57 |
Following is example how to mount table like that to C</mnt>: |
58 |
|
59 |
my $mnt = Fuse::DBI->mount({ |
60 |
'filenames' => 'select id,filename,size,writable from files', |
61 |
'read' => 'select content from files where id = ?', |
62 |
'update' => 'update files set content = ? where id = ?', |
63 |
'dsn' => 'DBI:Pg:dbname=test_db', |
64 |
'user' => 'database_user', |
65 |
'password' => 'database_password', |
66 |
'invalidate' => sub { ... }, |
67 |
}); |
68 |
|
69 |
Options: |
70 |
|
71 |
=over 5 |
72 |
|
73 |
=item filenames |
74 |
|
75 |
SQL query which returns C<id> (unique id for that row), C<filename>, |
76 |
C<size> and C<writable> boolean flag. |
77 |
|
78 |
=item read |
79 |
|
80 |
SQL query which returns only one column with content of file and has |
81 |
placeholder C<?> for C<id>. |
82 |
|
83 |
=item update |
84 |
|
85 |
SQL query with two pace-holders, one for new content and one for C<id>. |
86 |
|
87 |
=item dsn |
88 |
|
89 |
C<DBI> dsn to connect to (contains database driver and name of database). |
90 |
|
91 |
=item user |
92 |
|
93 |
User with which to connect to database |
94 |
|
95 |
=item password |
96 |
|
97 |
Password for connecting to database |
98 |
|
99 |
=item invalidate |
100 |
|
101 |
Optional anonymous code reference which will be executed when data is updated in |
102 |
database. It can be used as hook to delete cache (for example on-disk-cache) |
103 |
which is created from data edited through C<Fuse::DBI>. |
104 |
|
105 |
=item fork |
106 |
|
107 |
Optional flag which forks after mount so that executing script will continue |
108 |
running. Implementation is experimental. |
109 |
|
110 |
=back |
111 |
|
112 |
=cut |
113 |
|
114 |
my $dbh; |
115 |
my $sth; |
116 |
my $ctime_start; |
117 |
|
118 |
sub read_filenames; |
119 |
sub fuse_module_loaded; |
120 |
|
121 |
# evil, evil way to solve this. It makes this module non-reentrant. But, since |
122 |
# fuse calls another copy of this script for each mount anyway, this shouldn't |
123 |
# be a problem. |
124 |
my $fuse_self; |
125 |
|
126 |
sub mount { |
127 |
my $class = shift; |
128 |
my $self = {}; |
129 |
bless($self, $class); |
130 |
|
131 |
my $arg = shift; |
132 |
|
133 |
print Dumper($arg); |
134 |
|
135 |
carp "mount needs 'dsn' to connect to (e.g. dsn => 'DBI:Pg:dbname=test')" unless ($arg->{'dsn'}); |
136 |
carp "mount needs 'mount' as mountpoint" unless ($arg->{'mount'}); |
137 |
|
138 |
# save (some) arguments in self |
139 |
foreach (qw(mount invalidate)) { |
140 |
$self->{$_} = $arg->{$_}; |
141 |
} |
142 |
|
143 |
foreach (qw(filenames read update)) { |
144 |
carp "mount needs '$_' SQL" unless ($arg->{$_}); |
145 |
} |
146 |
|
147 |
$ctime_start = time(); |
148 |
|
149 |
my $pid; |
150 |
if ($arg->{'fork'}) { |
151 |
$pid = fork(); |
152 |
die "fork() failed: $!" unless defined $pid; |
153 |
# child will return to caller |
154 |
if ($pid) { |
155 |
return $self; |
156 |
} |
157 |
} |
158 |
|
159 |
$dbh = DBI->connect($arg->{'dsn'},$arg->{'user'},$arg->{'password'}, {AutoCommit => 0, RaiseError => 1}) || die $DBI::errstr; |
160 |
|
161 |
$sth->{'filenames'} = $dbh->prepare($arg->{'filenames'}) || die $dbh->errstr(); |
162 |
|
163 |
$sth->{'read'} = $dbh->prepare($arg->{'read'}) || die $dbh->errstr(); |
164 |
$sth->{'update'} = $dbh->prepare($arg->{'update'}) || die $dbh->errstr(); |
165 |
|
166 |
|
167 |
$self->{'sth'} = $sth; |
168 |
|
169 |
$self->{'read_filenames'} = sub { $self->read_filenames }; |
170 |
$self->read_filenames; |
171 |
|
172 |
$fuse_self = \$self; |
173 |
|
174 |
Fuse::main( |
175 |
mountpoint=>$arg->{'mount'}, |
176 |
getattr=>\&e_getattr, |
177 |
getdir=>\&e_getdir, |
178 |
open=>\&e_open, |
179 |
statfs=>\&e_statfs, |
180 |
read=>\&e_read, |
181 |
write=>\&e_write, |
182 |
utime=>\&e_utime, |
183 |
truncate=>\&e_truncate, |
184 |
unlink=>\&e_unlink, |
185 |
rmdir=>\&e_unlink, |
186 |
debug=>0, |
187 |
); |
188 |
|
189 |
exit(0) if ($arg->{'fork'}); |
190 |
|
191 |
return 1; |
192 |
|
193 |
}; |
194 |
|
195 |
=head2 umount |
196 |
|
197 |
Unmount your database as filesystem. |
198 |
|
199 |
$mnt->umount; |
200 |
|
201 |
This will also kill background process which is translating |
202 |
database to filesystem. |
203 |
|
204 |
=cut |
205 |
|
206 |
sub umount { |
207 |
my $self = shift; |
208 |
|
209 |
if ($self->{'mount'}) { |
210 |
if (open(MTAB, "/etc/mtab")) { |
211 |
my $mounted = 0; |
212 |
my $mount = $self->{'mount'}; |
213 |
while(<MTAB>) { |
214 |
$mounted = 1 if (/ $mount fuse /i); |
215 |
} |
216 |
close(MTAB); |
217 |
|
218 |
if ($mounted) { |
219 |
system "fusermount -u ".$self->{'mount'}." 2>&1 >/dev/null" || return 0; |
220 |
return 1; |
221 |
} |
222 |
|
223 |
} else { |
224 |
warn "can't open /etc/mtab: $!"; |
225 |
return 0; |
226 |
} |
227 |
} |
228 |
} |
229 |
|
230 |
$SIG{'INT'} = sub { |
231 |
if ($fuse_self && $$fuse_self->umount) { |
232 |
print STDERR "umount called by SIG INT\n"; |
233 |
} |
234 |
}; |
235 |
|
236 |
$SIG{'QUIT'} = sub { |
237 |
if ($fuse_self && $$fuse_self->umount) { |
238 |
print STDERR "umount called by SIG QUIT\n"; |
239 |
} |
240 |
}; |
241 |
|
242 |
sub DESTROY { |
243 |
my $self = shift; |
244 |
if ($self->umount) { |
245 |
print STDERR "umount called by DESTROY\n"; |
246 |
} |
247 |
} |
248 |
|
249 |
=head2 fuse_module_loaded |
250 |
|
251 |
Checks if C<fuse> module is loaded in kernel. |
252 |
|
253 |
die "no fuse module loaded in kernel" |
254 |
unless (Fuse::DBI::fuse_module_loaded); |
255 |
|
256 |
This function in called by C<mount>, but might be useful alone also. |
257 |
|
258 |
=cut |
259 |
|
260 |
sub fuse_module_loaded { |
261 |
my $lsmod = `lsmod`; |
262 |
die "can't start lsmod: $!" unless ($lsmod); |
263 |
if ($lsmod =~ m/fuse/s) { |
264 |
return 1; |
265 |
} else { |
266 |
return 0; |
267 |
} |
268 |
} |
269 |
|
270 |
my %files; |
271 |
my %dirs; |
272 |
|
273 |
sub read_filenames { |
274 |
my $self = shift; |
275 |
|
276 |
my $sth = $self->{'sth'} || die "no sth argument"; |
277 |
|
278 |
# create empty filesystem |
279 |
(%files) = ( |
280 |
'.' => { |
281 |
type => 0040, |
282 |
mode => 0755, |
283 |
}, |
284 |
'..' => { |
285 |
type => 0040, |
286 |
mode => 0755, |
287 |
}, |
288 |
# a => { |
289 |
# cont => "File 'a'.\n", |
290 |
# type => 0100, |
291 |
# ctime => time()-2000 |
292 |
# }, |
293 |
); |
294 |
|
295 |
# fetch new filename list from database |
296 |
$sth->{'filenames'}->execute() || die $sth->{'filenames'}->errstr(); |
297 |
|
298 |
# read them in with sesible defaults |
299 |
while (my $row = $sth->{'filenames'}->fetchrow_hashref() ) { |
300 |
$files{$row->{'filename'}} = { |
301 |
size => $row->{'size'}, |
302 |
mode => $row->{'writable'} ? 0644 : 0444, |
303 |
id => $row->{'id'} || 99, |
304 |
}; |
305 |
|
306 |
my $d; |
307 |
foreach (split(m!/!, $row->{'filename'})) { |
308 |
# first, entry is assumed to be file |
309 |
if ($d) { |
310 |
$files{$d} = { |
311 |
size => $dirs{$d}++, |
312 |
mode => 0755, |
313 |
type => 0040 |
314 |
}; |
315 |
$files{$d.'/.'} = { |
316 |
mode => 0755, |
317 |
type => 0040 |
318 |
}; |
319 |
$files{$d.'/..'} = { |
320 |
mode => 0755, |
321 |
type => 0040 |
322 |
}; |
323 |
} |
324 |
$d .= "/" if ($d); |
325 |
$d .= "$_"; |
326 |
} |
327 |
} |
328 |
|
329 |
print "found ",scalar(keys %files)-scalar(keys %dirs)," files, ",scalar(keys %dirs), " dirs\n"; |
330 |
} |
331 |
|
332 |
|
333 |
sub filename_fixup { |
334 |
my ($file) = shift; |
335 |
$file =~ s,^/,,; |
336 |
$file = '.' unless length($file); |
337 |
return $file; |
338 |
} |
339 |
|
340 |
sub e_getattr { |
341 |
my ($file) = filename_fixup(shift); |
342 |
$file =~ s,^/,,; |
343 |
$file = '.' unless length($file); |
344 |
return -ENOENT() unless exists($files{$file}); |
345 |
my ($size) = $files{$file}{size} || 1; |
346 |
my ($dev, $ino, $rdev, $blocks, $gid, $uid, $nlink, $blksize) = (0,0,0,1,0,0,1,1024); |
347 |
my ($atime, $ctime, $mtime); |
348 |
$atime = $ctime = $mtime = $files{$file}{ctime} || $ctime_start; |
349 |
|
350 |
my ($modes) = (($files{$file}{type} || 0100)<<9) + $files{$file}{mode}; |
351 |
|
352 |
# 2 possible types of return values: |
353 |
#return -ENOENT(); # or any other error you care to |
354 |
#print(join(",",($dev,$ino,$modes,$nlink,$uid,$gid,$rdev,$size,$atime,$mtime,$ctime,$blksize,$blocks)),"\n"); |
355 |
return ($dev,$ino,$modes,$nlink,$uid,$gid,$rdev,$size,$atime,$mtime,$ctime,$blksize,$blocks); |
356 |
} |
357 |
|
358 |
sub e_getdir { |
359 |
my ($dirname) = shift; |
360 |
$dirname =~ s!^/!!; |
361 |
# return as many text filenames as you like, followed by the retval. |
362 |
print((scalar keys %files)." files total\n"); |
363 |
my %out; |
364 |
foreach my $f (sort keys %files) { |
365 |
if ($dirname) { |
366 |
if ($f =~ s/^\Q$dirname\E\///) { |
367 |
$out{$f}++ if ($f =~ /^[^\/]+$/); |
368 |
} |
369 |
} else { |
370 |
$out{$f}++ if ($f =~ /^[^\/]+$/); |
371 |
} |
372 |
} |
373 |
if (! %out) { |
374 |
$out{'no files? bug?'}++; |
375 |
} |
376 |
print scalar keys %out," files in dir '$dirname'\n"; |
377 |
print "## ",join(" ",keys %out),"\n"; |
378 |
return (keys %out),0; |
379 |
} |
380 |
|
381 |
sub read_content { |
382 |
my ($file,$id) = @_; |
383 |
|
384 |
die "read_content needs file and id" unless ($file && $id); |
385 |
|
386 |
$sth->{'read'}->execute($id) || die $sth->{'read'}->errstr; |
387 |
$files{$file}{cont} = $sth->{'read'}->fetchrow_array; |
388 |
# I should modify ctime only if content in database changed |
389 |
#$files{$file}{ctime} = time() unless ($files{$file}{ctime}); |
390 |
print "file '$file' content [",length($files{$file}{cont})," bytes] read in cache\n"; |
391 |
} |
392 |
|
393 |
|
394 |
sub e_open { |
395 |
# VFS sanity check; it keeps all the necessary state, not much to do here. |
396 |
my $file = filename_fixup(shift); |
397 |
my $flags = shift; |
398 |
|
399 |
return -ENOENT() unless exists($files{$file}); |
400 |
return -EISDIR() unless exists($files{$file}{id}); |
401 |
|
402 |
read_content($file,$files{$file}{id}) unless exists($files{$file}{cont}); |
403 |
|
404 |
print "open '$file' ",length($files{$file}{cont})," bytes\n"; |
405 |
return 0; |
406 |
} |
407 |
|
408 |
sub e_read { |
409 |
# return an error numeric, or binary/text string. |
410 |
# (note: 0 means EOF, "0" will give a byte (ascii "0") |
411 |
# to the reading program) |
412 |
my ($file) = filename_fixup(shift); |
413 |
my ($buf_len,$off) = @_; |
414 |
|
415 |
return -ENOENT() unless exists($files{$file}); |
416 |
|
417 |
my $len = length($files{$file}{cont}); |
418 |
|
419 |
print "read '$file' [$len bytes] offset $off length $buf_len\n"; |
420 |
|
421 |
return -EINVAL() if ($off > $len); |
422 |
return 0 if ($off == $len); |
423 |
|
424 |
$buf_len = $len-$off if ($len - $off < $buf_len); |
425 |
|
426 |
return substr($files{$file}{cont},$off,$buf_len); |
427 |
} |
428 |
|
429 |
sub clear_cont { |
430 |
print "transaction rollback\n"; |
431 |
$dbh->rollback || die $dbh->errstr; |
432 |
print "invalidate all cached content\n"; |
433 |
foreach my $f (keys %files) { |
434 |
delete $files{$f}{cont}; |
435 |
delete $files{$f}{ctime}; |
436 |
} |
437 |
print "begin new transaction\n"; |
438 |
#$dbh->begin_work || die $dbh->errstr; |
439 |
} |
440 |
|
441 |
|
442 |
sub update_db { |
443 |
my $file = shift || die; |
444 |
|
445 |
$files{$file}{ctime} = time(); |
446 |
|
447 |
my ($cont,$id) = ( |
448 |
$files{$file}{cont}, |
449 |
$files{$file}{id} |
450 |
); |
451 |
|
452 |
if (!$sth->{'update'}->execute($cont,$id)) { |
453 |
print "update problem: ",$sth->{'update'}->errstr; |
454 |
clear_cont; |
455 |
return 0; |
456 |
} else { |
457 |
if (! $dbh->commit) { |
458 |
print "ERROR: commit problem: ",$sth->{'update'}->errstr; |
459 |
clear_cont; |
460 |
return 0; |
461 |
} |
462 |
print "updated '$file' [",$files{$file}{id},"]\n"; |
463 |
|
464 |
$$fuse_self->{'invalidate'}->() if (ref $$fuse_self->{'invalidate'}); |
465 |
} |
466 |
return 1; |
467 |
} |
468 |
|
469 |
sub e_write { |
470 |
my $file = filename_fixup(shift); |
471 |
my ($buffer,$off) = @_; |
472 |
|
473 |
return -ENOENT() unless exists($files{$file}); |
474 |
|
475 |
my $cont = $files{$file}{cont}; |
476 |
my $len = length($cont); |
477 |
|
478 |
print "write '$file' [$len bytes] offset $off length ",length($buffer),"\n"; |
479 |
|
480 |
$files{$file}{cont} = ""; |
481 |
|
482 |
$files{$file}{cont} .= substr($cont,0,$off) if ($off > 0); |
483 |
$files{$file}{cont} .= $buffer; |
484 |
$files{$file}{cont} .= substr($cont,$off+length($buffer),$len-$off-length($buffer)) if ($off+length($buffer) < $len); |
485 |
|
486 |
$files{$file}{size} = length($files{$file}{cont}); |
487 |
|
488 |
if (! update_db($file)) { |
489 |
return -ENOSYS(); |
490 |
} else { |
491 |
return length($buffer); |
492 |
} |
493 |
} |
494 |
|
495 |
sub e_truncate { |
496 |
my $file = filename_fixup(shift); |
497 |
my $size = shift; |
498 |
|
499 |
print "truncate to $size\n"; |
500 |
|
501 |
$files{$file}{cont} = substr($files{$file}{cont},0,$size); |
502 |
$files{$file}{size} = $size; |
503 |
return 0 |
504 |
}; |
505 |
|
506 |
|
507 |
sub e_utime { |
508 |
my ($atime,$mtime,$file) = @_; |
509 |
$file = filename_fixup($file); |
510 |
|
511 |
return -ENOENT() unless exists($files{$file}); |
512 |
|
513 |
print "utime '$file' $atime $mtime\n"; |
514 |
|
515 |
$files{$file}{time} = $mtime; |
516 |
return 0; |
517 |
} |
518 |
|
519 |
sub e_statfs { return 255, 1, 1, 1, 1, 2 } |
520 |
|
521 |
sub e_unlink { |
522 |
my $file = filename_fixup(shift); |
523 |
|
524 |
if (exists( $dirs{$file} )) { |
525 |
print "unlink '$file' will re-read template names\n"; |
526 |
print Dumper($fuse_self); |
527 |
$$fuse_self->{'read_filenames'}->(); |
528 |
return 0; |
529 |
} elsif (exists( $files{$file} )) { |
530 |
print "unlink '$file' will invalidate cache\n"; |
531 |
read_content($file,$files{$file}{id}); |
532 |
return 0; |
533 |
} |
534 |
|
535 |
return -ENOENT(); |
536 |
} |
537 |
1; |
538 |
__END__ |
539 |
|
540 |
=head1 EXPORT |
541 |
|
542 |
Nothing. |
543 |
|
544 |
=head1 BUGS |
545 |
|
546 |
Size information (C<ls -s>) is wrong. It's a problem in upstream Fuse module |
547 |
(for which I'm to blame lately), so when it gets fixes, C<Fuse::DBI> will |
548 |
automagically pick it up. |
549 |
|
550 |
=head1 SEE ALSO |
551 |
|
552 |
C<FUSE (Filesystem in USErspace)> website |
553 |
L<http://fuse.sourceforge.net/> |
554 |
|
555 |
Example for WebGUI which comes with this distribution in |
556 |
directory C<examples/webgui.pl>. It also contains a lot of documentation |
557 |
about design of this module, usage and limitations. |
558 |
|
559 |
=head1 AUTHOR |
560 |
|
561 |
Dobrica Pavlinusic, E<lt>dpavlin@rot13.orgE<gt> |
562 |
|
563 |
=head1 COPYRIGHT AND LICENSE |
564 |
|
565 |
Copyright (C) 2004 by Dobrica Pavlinusic |
566 |
|
567 |
This library is free software; you can redistribute it and/or modify |
568 |
it under the same terms as Perl itself, either Perl version 5.8.4 or, |
569 |
at your option, any later version of Perl 5 you may have available. |
570 |
|
571 |
|
572 |
=cut |
573 |
|