12 |
# |
# |
13 |
# |
# |
14 |
# usage: |
# usage: |
15 |
# $ backup.pl mountscript |
# $ psinib.pl mountscript |
16 |
|
|
17 |
use strict 'vars'; |
use strict 'vars'; |
18 |
use Data::Dumper; |
use Data::Dumper; |
22 |
use Filesys::SmbClient; |
use Filesys::SmbClient; |
23 |
#use Taint; |
#use Taint; |
24 |
use Fcntl qw(LOCK_EX LOCK_NB); |
use Fcntl qw(LOCK_EX LOCK_NB); |
25 |
|
use Digest::MD5; |
26 |
|
use File::Basename; |
27 |
|
|
28 |
# configuration |
# configuration |
29 |
my $LOG_TIME_FMT = '%Y-%m-%d %H:%M:%S'; # strftime format for logfile |
my $LOG_TIME_FMT = '%Y-%m-%d %H:%M:%S'; # strftime format for logfile |
30 |
my $DIR_TIME_FMT = '%Y%m%d'; # strftime format for backup dir |
my $DIR_TIME_FMT = '%Y%m%d'; # strftime format for backup dir |
31 |
|
|
32 |
my $LOG = '/var/log/backup.log'; # add path here... |
my $LOG = '/var/log/backup.log'; # add path here... |
33 |
$LOG = '/tmp/backup.log'; |
#$LOG = '/tmp/backup.log'; |
34 |
|
|
35 |
# store backups in which directory |
# store backups in which directory |
36 |
my $BACKUP_DEST = '/data/isis_backup'; |
my $BACKUP_DEST = '/backup/isis_backup'; |
37 |
|
|
38 |
# files to ignore in backup |
# files to ignore in backup |
39 |
my @ignore = ('.md5sum', '.backupignore', 'backupignore.txt'); |
my @ignore = ('.md5sum', '.backupignore', 'backupignore.txt'); |
40 |
|
|
41 |
# open log |
# open log |
42 |
open(L, "> $LOG") || die "can't open log $LOG: $!"; |
open(L, ">> $LOG") || die "can't open log $LOG: $!"; |
43 |
select((select(L), $|=1)[0]); # flush output |
select((select(L), $|=1)[0]); # flush output |
44 |
|
|
45 |
# make a lock on logfile |
# make a lock on logfile |
64 |
|
|
65 |
my @in_backup; # shares which are backeduped this run |
my @in_backup; # shares which are backeduped this run |
66 |
|
|
67 |
my $p = new Net::Ping->new(); |
my $p = new Net::Ping->new("tcp", 2); |
68 |
|
# ping will try tcp connect to netbios-ssn (139) |
69 |
|
$p->{port_num} = getservbyname("netbios-ssn", "tcp"); |
70 |
|
|
71 |
my $backup_ok = 0; |
my $backup_ok = 0; |
72 |
|
|
73 |
my $smb; |
my $smb; |
74 |
my %smb_atime; |
my %smb_atime; |
75 |
my %smb_mtime; |
my %smb_mtime; |
76 |
|
my %file_md5; |
77 |
|
|
78 |
open(M, $mounts) || die "can't open $mounts: $!"; |
open(M, $mounts) || die "can't open $mounts: $!"; |
79 |
while(<M>) { |
while(<M>) { |
102 |
|
|
103 |
push @in_backup,$share; |
push @in_backup,$share; |
104 |
|
|
105 |
|
|
106 |
|
my ($host,$dir,$date_dir) = share2host_dir($share); |
107 |
|
my $bl = "$BACKUP_DEST/$host/$dir/latest"; # latest backup |
108 |
|
my $bc = "$BACKUP_DEST/$host/$dir/$date_dir"; # current one |
109 |
|
my $real_bl; |
110 |
|
if (-l $bl) { |
111 |
|
$real_bl=readlink($bl) || die "can't read link $bl: $!"; |
112 |
|
$real_bl="$BACKUP_DEST/$host/$dir/$real_bl" if (substr($real_bl,0,1) ne "/"); |
113 |
|
if (-l $bc && $real_bl eq $bc) { |
114 |
|
print "$share allready backuped...\n"; |
115 |
|
$backup_ok++; |
116 |
|
next; |
117 |
|
} |
118 |
|
|
119 |
|
} |
120 |
|
|
121 |
|
|
122 |
print "working on $share\n"; |
print "working on $share\n"; |
123 |
|
|
124 |
|
|
125 |
my $ip = get_ip($share); |
my $ip = get_ip($share); |
126 |
|
|
127 |
if ($ip) { |
if ($ip) { |
140 |
|
|
141 |
#------------------------------------------------------------------------- |
#------------------------------------------------------------------------- |
142 |
|
|
143 |
|
|
144 |
# get IP number from share |
# get IP number from share |
145 |
sub get_ip { |
sub get_ip { |
146 |
my $share = shift; |
my $share = shift; |
153 |
} |
} |
154 |
} |
} |
155 |
|
|
156 |
|
|
157 |
|
# write entry to screen and log |
158 |
sub xlog { |
sub xlog { |
159 |
my $share = shift; |
my $share = shift; |
160 |
my $t = strftime $LOG_TIME_FMT, localtime; |
my $t = strftime $LOG_TIME_FMT, localtime; |
163 |
print L "$t $share\t$m\n"; |
print L "$t $share\t$m\n"; |
164 |
} |
} |
165 |
|
|
166 |
sub snap_share { |
# dump warn and dies into log |
167 |
|
BEGIN { $SIG{'__WARN__'} = sub { xlog('WARN',$_[0]) ; warn $_[0] } } |
168 |
my $share = shift; |
BEGIN { $SIG{'__DIE__'} = sub { xlog('DIE',$_[0]) ; die $_[0] } } |
169 |
|
|
|
my %param = ( debug => 0 ); |
|
|
|
|
|
$param{username} = shift; |
|
|
$param{password} = shift; |
|
|
$param{workgroup} = shift; |
|
170 |
|
|
171 |
|
# split share name to host, dir and currnet date dir |
172 |
|
sub share2host_dir { |
173 |
|
my $share = shift; |
174 |
my ($host,$dir); |
my ($host,$dir); |
175 |
if ($share =~ m#//([^/]+)/(.+)$#) { |
if ($share =~ m#//([^/]+)/(.+)$#) { |
176 |
($host,$dir) = ($1,$2); |
($host,$dir) = ($1,$2); |
181 |
print "Can't parse share $share into host and directory!\n"; |
print "Can't parse share $share into host and directory!\n"; |
182 |
return; |
return; |
183 |
} |
} |
184 |
|
return ($host,$dir,strftime $DIR_TIME_FMT, localtime); |
185 |
|
} |
186 |
|
|
187 |
|
|
188 |
|
# make a snapshot of a share |
189 |
|
sub snap_share { |
190 |
|
|
191 |
|
my $share = shift; |
192 |
|
|
193 |
|
my %param = ( debug => 0 ); |
194 |
|
|
195 |
|
$param{username} = shift || warn "can't find username for share $share"; |
196 |
|
$param{password} = shift || warn "can't find passwod for share $share"; |
197 |
|
$param{workgroup} = shift || warn "can't find workgroup for share $share"; |
198 |
|
|
199 |
my $date_dir = strftime $DIR_TIME_FMT, localtime; |
my ($host,$dir,$date_dir) = share2host_dir($share); |
200 |
|
|
201 |
# latest backup directory |
# latest backup directory |
202 |
my $bl = "$BACKUP_DEST/$host/$dir/latest"; |
my $bl = "$BACKUP_DEST/$host/$dir/latest"; |
204 |
my $bc = "$BACKUP_DEST/$host/$dir/$date_dir"; |
my $bc = "$BACKUP_DEST/$host/$dir/$date_dir"; |
205 |
|
|
206 |
my $real_bl; |
my $real_bl; |
207 |
if (-e $bl) { |
if (-l $bl) { |
208 |
$real_bl=readlink($bl) || die "can't read link $bl: $!"; |
$real_bl=readlink($bl) || die "can't read link $bl: $!"; |
209 |
$real_bl="$BACKUP_DEST/$host/$dir/$real_bl" if (substr($real_bl,0,1) ne "/"); |
$real_bl="$BACKUP_DEST/$host/$dir/$real_bl" if (substr($real_bl,0,1) ne "/"); |
210 |
} else { |
} else { |
211 |
print "no old backup, this is first run...\n"; |
print "no old backup, trying to find last backup, "; |
212 |
|
if (opendir(BL_DIR, "$BACKUP_DEST/$host/$dir")) { |
213 |
|
my @bl_dirs = sort grep { !/^\./ && -d "$BACKUP_DEST/$host/$dir/$_" } readdir(BL_DIR); |
214 |
|
closedir(BL_DIR); |
215 |
|
$real_bl=pop @bl_dirs; |
216 |
|
print "using $real_bl as latest...\n"; |
217 |
|
$real_bl="$BACKUP_DEST/$host/$dir/$real_bl" if (substr($real_bl,0,1) ne "/"); |
218 |
|
if ($real_bl eq $bc) { |
219 |
|
xlog($share,"latest from today (possible partial backup)"); |
220 |
|
rename $real_bl,$real_bl.".partial" || warn "can't reaname partial backup: $!"; |
221 |
|
$real_bl .= ".partial"; |
222 |
|
} |
223 |
|
} else { |
224 |
|
print "this is first run...\n"; |
225 |
|
} |
226 |
} |
} |
227 |
|
|
228 |
if (-e $bc && $real_bl && $real_bl eq $bc) { |
if (-l $bc && $real_bl && $real_bl eq $bc) { |
229 |
print "$share allready backuped...\n"; |
print "$share allready backuped...\n"; |
230 |
return; |
return; |
231 |
} |
} |
254 |
my %file_size; |
my %file_size; |
255 |
my %file_atime; |
my %file_atime; |
256 |
my %file_mtime; |
my %file_mtime; |
257 |
my %file_md5; |
#my %file_md5; |
258 |
|
|
259 |
my @smb_files; |
my @smb_files; |
260 |
my %smb_size; |
my %smb_size; |
261 |
#my %smb_atime; |
#my %smb_atime; |
262 |
#my %smb_mtime; |
#my %smb_mtime; |
|
my %smb_md5; |
|
|
|
|
263 |
|
|
264 |
sub norm_dir { |
sub norm_dir { |
265 |
my $foo = shift; |
my $foo = shift; |
275 |
my $di = 0; |
my $di = 0; |
276 |
while ($di <= $#dirs && $real_bl) { |
while ($di <= $#dirs && $real_bl) { |
277 |
my $d=$dirs[$di++]; |
my $d=$dirs[$di++]; |
278 |
opendir(DIR,"$bl/$d") || warn "opendir($bl/$d): $!\n"; |
opendir(DIR,"$real_bl/$d") || warn "opendir($real_bl/$d): $!\n"; |
279 |
|
|
280 |
# read .backupignore if exists |
# read .backupignore if exists |
281 |
if (-f "$bl/$d/.backupignore") { |
if (-f "$real_bl/$d/.backupignore") { |
282 |
open(I,"$bl/$d/.backupignore"); |
open(I,"$real_bl/$d/.backupignore"); |
283 |
while(<I>) { |
while(<I>) { |
284 |
chomp; |
chomp; |
285 |
push @ignore,norm_dir("$d/$_"); |
push @ignore,norm_dir("$d/$_"); |
286 |
} |
} |
287 |
close(I); |
close(I); |
288 |
print STDERR "ignore: ",join("|",@ignore),"\n"; |
#print STDERR "ignore: ",join("|",@ignore),"\n"; |
289 |
link "$bl/$d/.backupignore","$bc/$d/.backupignore" || |
link "$real_bl/$d/.backupignore","$bc/$d/.backupignore" || |
290 |
warn "can't copy $bl/$d/.backupignore to current backup dir: $!\n"; |
warn "can't copy $real_bl/$d/.backupignore to current backup dir: $!\n"; |
291 |
} |
} |
292 |
|
|
293 |
# read .md5sum if exists |
# read .md5sum if exists |
294 |
if (-f "$bl/$d/.md5sum") { |
if (-f "$real_bl/$d/.md5sum") { |
295 |
open(I,"$bl/$d/.md5sum"); |
open(I,"$real_bl/$d/.md5sum"); |
296 |
while(<I>) { |
while(<I>) { |
297 |
chomp; |
chomp; |
298 |
my ($md5,$f) = split(/\s+/,$_,2); |
my ($md5,$f) = split(/\s+/,$_,2); |
306 |
next if ($f eq '.'); |
next if ($f eq '.'); |
307 |
next if ($f eq '..'); |
next if ($f eq '..'); |
308 |
my $pr = norm_dir("$d/$f"); # path relative |
my $pr = norm_dir("$d/$f"); # path relative |
309 |
my $pf = norm_dir("$d/$f","$bl/"); # path full |
my $pf = norm_dir("$d/$f","$real_bl/"); # path full |
310 |
if (grep(/^\Q$pr\E$/,@ignore) == 0) { |
if (grep(/^\Q$pr\E$/,@ignore) == 0) { |
311 |
if (-f $pf) { |
if (-f $pf) { |
312 |
push @files,$pr; |
push @files,$pr; |
335 |
|
|
336 |
$di = 0; |
$di = 0; |
337 |
while ($di <= $#smb_dirs) { |
while ($di <= $#smb_dirs) { |
338 |
my $d=$smb_dirs[$di++]; |
my $d=$smb_dirs[$di]; |
339 |
my $pf = norm_dir($d,"smb:$share/"); # path full |
my $pf = norm_dir($d,"smb:$share/"); # path full |
340 |
my $D = $smb->opendir($pf) || warn "smb->opendir($pf): $!\n"; |
my $D = $smb->opendir($pf); |
341 |
|
if (! $D) { |
342 |
|
xlog($share,"FATAL: $share: $!"); |
343 |
|
# remove failing dir |
344 |
|
delete $smb_dirs[$di]; |
345 |
|
next; |
346 |
|
} |
347 |
|
$di++; |
348 |
|
|
349 |
my @clutter = $smb->readdir_struct($D); |
my @clutter = $smb->readdir_struct($D); |
350 |
foreach my $item (@clutter) { |
foreach my $item (@clutter) { |
407 |
next; |
next; |
408 |
} |
} |
409 |
|
|
410 |
|
my $md5 = Digest::MD5->new; |
411 |
|
|
412 |
my $fd = $smb->open("$from/$f"); |
my $fd = $smb->open("$from/$f"); |
413 |
if (! $fd) { |
if (! $fd) { |
414 |
print STDERR "can't open smb file $from/$f: $!\n"; |
print STDERR "can't open smb file $from/$f: $!\n"; |
418 |
while (defined(my $b=$smb->read($fd,4096))) { |
while (defined(my $b=$smb->read($fd,4096))) { |
419 |
print F $b; |
print F $b; |
420 |
$l += length($b); |
$l += length($b); |
421 |
|
$md5->add($b); |
422 |
} |
} |
423 |
|
|
424 |
$smb->close($fd); |
$smb->close($fd); |
425 |
close(F); |
close(F); |
426 |
|
|
427 |
|
$file_md5{$f} = $md5->hexdigest; |
428 |
|
|
429 |
# FIX: this fails with -T |
# FIX: this fails with -T |
430 |
my ($a,$m) = ($smb->stat("$from/$f"))[10,11]; |
my ($a,$m) = ($smb->stat("$from/$f"))[10,11]; |
431 |
utime $a, $m, "$to/$f" || |
utime $a, $m, "$to/$f" || |
479 |
xlog($share,"$transfer bytes transfered..."); |
xlog($share,"$transfer bytes transfered..."); |
480 |
|
|
481 |
foreach (@ln_files) { |
foreach (@ln_files) { |
482 |
link "$bl/$_","$bc/$_" || warn "link $bl/$_ -> $bc/$_: $!\n"; |
link "$real_bl/$_","$bc/$_" || warn "link $real_bl/$_ -> $bc/$_: $!\n"; |
483 |
} |
} |
484 |
|
|
485 |
# remove files |
# remove files |
492 |
rmdir "$bc/$_" || warn "rmdir $_: $!\n"; |
rmdir "$bc/$_" || warn "rmdir $_: $!\n"; |
493 |
} |
} |
494 |
|
|
495 |
|
# remove old .md5sum |
496 |
# FIX: create .md5sum |
foreach (sort @dirs) { |
497 |
|
unlink "$bc/$_/.md5sum" if (-e "$bc/$_/.md5sum"); |
498 |
|
} |
499 |
|
|
500 |
|
# create .md5sum |
501 |
|
my $last_dir = ''; |
502 |
|
my $md5; |
503 |
|
foreach my $f (sort { $file_md5{$a} cmp $file_md5{$b} } keys %file_md5) { |
504 |
|
my $dir = dirname($f); |
505 |
|
my $file = basename($f); |
506 |
|
print "$f -- $dir / $file<--\n"; |
507 |
|
if ($dir ne $last_dir) { |
508 |
|
close($md5) if ($md5); |
509 |
|
open($md5, ">> $bc/$dir/.md5sum") || warn "can't create $bc/$dir/.md5sum: $!"; |
510 |
|
$last_dir = $dir; |
511 |
|
#print STDERR "writing $last_dir/.md5sum\n"; |
512 |
|
} |
513 |
|
print $md5 $file_md5{$f}," $file\n"; |
514 |
|
} |
515 |
|
close($md5); |
516 |
|
|
517 |
# create leatest link |
# create leatest link |
518 |
|
#print "ln -s $bc $real_bl\n"; |
519 |
|
if (-l $bl) { |
520 |
|
unlink $bl || warn "can't remove old latest symlink $bl: $!\n"; |
521 |
|
} |
522 |
symlink $bc,$bl || warn "can't create latest symlink $bl -> $bc: $!\n"; |
symlink $bc,$bl || warn "can't create latest symlink $bl -> $bc: $!\n"; |
523 |
|
|
524 |
|
# FIX: sanity check -- remove for speedup |
525 |
|
xlog($share,"failed to create latest symlink $bl -> $bc...") if (readlink($bl) ne $bc || ! -l $bl); |
526 |
|
|
527 |
xlog($share,"backup completed..."); |
xlog($share,"backup completed..."); |
528 |
} |
} |
529 |
|
|
530 |
|
__END__ |
531 |
#------------------------------------------------------------------------- |
#------------------------------------------------------------------------- |
532 |
|
|
533 |
|
|
534 |
|
=head1 NAME |
535 |
|
|
536 |
|
psinib - Perl Snapshot Is Not Incremental Backup |
537 |
|
|
538 |
|
=head1 SYNOPSIS |
539 |
|
|
540 |
|
./psinib.pl |
541 |
|
|
542 |
|
=head1 DESCRIPTION |
543 |
|
|
544 |
|
This script in current version support just backup of Samba (or Micro$oft |
545 |
|
Winblowz) shares to central disk space. Central disk space is organized in |
546 |
|
multiple directories named after: |
547 |
|
|
548 |
|
=over 4 |
549 |
|
|
550 |
|
=item * |
551 |
|
server which is sharing files to be backed up |
552 |
|
|
553 |
|
=item * |
554 |
|
name of share on server |
555 |
|
|
556 |
|
=item * |
557 |
|
dated directory named like standard ISO date format (YYYYMMDD). |
558 |
|
|
559 |
|
=back |
560 |
|
|
561 |
|
In each dated directory you will find I<snapshot> of all files on |
562 |
|
exported share on that particular date. |
563 |
|
|
564 |
|
You can also use symlink I<latest> which will lead you to |
565 |
|
last completed backup. After that you can use some other backup |
566 |
|
software to transfer I<snapshot> to tape, CD-ROM or some other media. |
567 |
|
|
568 |
|
=head2 Design considerations |
569 |
|
|
570 |
|
Since taking of share snapshot every day requires a lot of disk space and |
571 |
|
network bandwidth, B<psinib> uses several techniques to keep disk usage and |
572 |
|
network traffic at acceptable level: |
573 |
|
|
574 |
|
=over 3 |
575 |
|
|
576 |
|
=item - usage of hard-links to provide same files in each snapshot (as opposed |
577 |
|
to have multiple copies of same file) |
578 |
|
|
579 |
|
=item - usage of file size, atime and mtime to find changes of files without |
580 |
|
transferring whole file over network (just share browsing is transfered |
581 |
|
over network) |
582 |
|
|
583 |
|
=item - usage of C<.md5sum> files (compatible with command-line utility |
584 |
|
C<md5sum>) to keep file between snapshots hard-linked |
585 |
|
|
586 |
|
=back |
587 |
|
|
588 |
|
=head1 CONFIGURATION |
589 |
|
|
590 |
|
This section is not yet written. |
591 |
|
|
592 |
|
=head1 HACKS, TRICKS, BUGS and LIMITATIONS |
593 |
|
|
594 |
|
This chapter will have all content that doesn't fit anywhere else. |
595 |
|
|
596 |
|
=head2 Can snapshots be more frequent than daily? |
597 |
|
|
598 |
|
There is not real reason why you can't take snapshot more often than |
599 |
|
once a day. Actually, if you are using B<psinib> to backup Windows |
600 |
|
workstations you already know that they tend to come-and-go during the day |
601 |
|
(reboots probably ;-), so running B<psinib> several times a day increases |
602 |
|
your chance of having up-to-date backup (B<psinib> will not make multiple |
603 |
|
snapshots for same day, nor will it update snapshot for current day if |
604 |
|
it already exists). |
605 |
|
|
606 |
|
However, changing B<psinib> to produce snapshots which are, for example, hourly |
607 |
|
is a simple change of C<$DIR_TIME_FMT> which is currently set to |
608 |
|
C<'%Y%m%d'> (see I<strftime> documentation for explanation of that |
609 |
|
format). If you change that to C<'%Y%m%d-%H> you can have hourly snapshots |
610 |
|
(if your network is fast enough, that is...). Also, some of messages in |
611 |
|
program will sound strange, but other than that it should work. |
612 |
|
I<You have been warned>. |
613 |
|
|
614 |
|
=head2 Do I really need to share every directory which I want to snapshot? |
615 |
|
|
616 |
|
Actually, no. Due to usage of C<Filesys::SmbClient> module, you can also |
617 |
|
specify sub-directory inside your share that you want to backup. This feature |
618 |
|
is most useful if you want to use administrative shares (but, have in mind |
619 |
|
that you have to enter your Win administrator password in unencrypted file on |
620 |
|
disk to do that) like this: |
621 |
|
|
622 |
|
smbmount //server/c$/WinNT/fonts /mnt -o username=administrator%win |
623 |
|
|
624 |
|
After that you will get directories with snapshots like: |
625 |
|
|
626 |
|
server/c_WinNT_fonts/yyyymmdd/.... |
627 |
|
|
628 |
|
=head2 Won't I run out of disk space? |
629 |
|
|
630 |
|
Of course you will... Snapshots and logfiles will eventually fill-up your disk. |
631 |
|
However, you can do two things to stop that: |
632 |
|
|
633 |
|
=head3 Clean snapshort older than x days |
634 |
|
|
635 |
|
You can add following command to your C<root> crontab: |
636 |
|
|
637 |
|
find /backup/isis_backup -type d -mindepth 3 -maxdepth 3 -mtime +11 -exec rm -Rf {} \; |
638 |
|
|
639 |
|
I assume that C</backup/isis_backup> is directory in which are your snapshots |
640 |
|
and that you don't want to keep snapshots older than 11 days (that's |
641 |
|
C<-mtime +11> part of command). |
642 |
|
|
643 |
|
=head3 Rotate your logs |
644 |
|
|
645 |
|
I will leave that to you. I relay on GNU/Debian's C<logrotate> to do it for me. |
646 |
|
|
647 |
|
=head2 What are I<YYYYMMDD.partial> directories? |
648 |
|
|
649 |
|
If there isn't I<latest> symlink in snapshot directory, it's preatty safe to |
650 |
|
assume that previous backup from that day failed. So, that directory will |
651 |
|
be renamed to I<YYYYMMDD.partial> and snapshot will be performed again, |
652 |
|
linking same files (other alternative would be to erase that dir and find |
653 |
|
second-oldest directory, but this seemed like more correct approach). |
654 |
|
|
655 |
|
=head1 AUTHOR |
656 |
|
|
657 |
|
Dobrica Pavlinusic <dpavlin@rot13.org> |
658 |
|
|
659 |
|
L<http://www.rot13.org/~dpavlin/> |
660 |
|
|
661 |
|
=head1 LICENSE |
662 |
|
|
663 |
|
This product is licensed under GNU Public License (GPL) v2 or later. |
664 |
|
|
665 |
|
=cut |