#============================================================= -*-perl-*- # # BackupPC::Lib package # # DESCRIPTION # # This library defines a BackupPC::Lib class and a variety of utility # functions used by BackupPC. # # AUTHOR # Craig Barratt # # COPYRIGHT # Copyright (C) 2001-2009 Craig Barratt # # This program is free software; you can redistribute it and/or modify # it under the terms of the GNU General Public License as published by # the Free Software Foundation; either version 2 of the License, or # (at your option) any later version. # # This program is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # GNU General Public License for more details. # # You should have received a copy of the GNU General Public License # along with this program; if not, write to the Free Software # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA # #======================================================================== # # Version 3.2.0, released 31 Jul 2010. # # See http://backuppc.sourceforge.net. # #======================================================================== package BackupPC::Lib; use strict; use vars qw(%Conf %Lang); use BackupPC::Storage; use Fcntl ':mode'; use Carp; use File::Path; use File::Compare; use Socket; use Cwd; use Digest::MD5; use Config; use Encode qw/from_to encode_utf8/; use vars qw( $IODirentOk $IODirentLoaded ); use vars qw(@ISA @EXPORT @EXPORT_OK %EXPORT_TAGS); require Exporter; require DynaLoader; @ISA = qw(Exporter DynaLoader); @EXPORT_OK = qw( BPC_DT_UNKNOWN BPC_DT_FIFO BPC_DT_CHR BPC_DT_DIR BPC_DT_BLK BPC_DT_REG BPC_DT_LNK BPC_DT_SOCK ); @EXPORT = qw( ); %EXPORT_TAGS = ('BPC_DT_ALL' => [@EXPORT, @EXPORT_OK]); BEGIN { eval "use IO::Dirent qw( readdirent DT_DIR );"; $IODirentLoaded = 1 if ( !$@ ); }; # # The need to match the constants in IO::Dirent # use constant BPC_DT_UNKNOWN => 0; use constant BPC_DT_FIFO => 1; ## named pipe (fifo) use constant BPC_DT_CHR => 2; ## character special use constant BPC_DT_DIR => 4; ## directory use constant BPC_DT_BLK => 6; ## block special use constant BPC_DT_REG => 8; ## regular use constant BPC_DT_LNK => 10; ## symbolic link use constant BPC_DT_SOCK => 12; ## socket sub new { my $class = shift; my($topDir, $installDir, $confDir, $noUserCheck) = @_; # # Whether to use filesystem hierarchy standard for file layout. # If set, text config files are below /etc/BackupPC. # my $useFHS = 1; my $paths; # # Set defaults for $topDir and $installDir. # $topDir = '/data/BackupPC' if ( $topDir eq "" ); $installDir = '/usr/local/BackupPC' if ( $installDir eq "" ); $confDir = '/etc/BackupPC'; # FIXME remove this! XXX # # Pick some initial defaults. For FHS the only critical # path is the ConfDir, since we get everything else out # of the main config file. # if ( $useFHS ) { $paths = { useFHS => $useFHS, TopDir => $topDir, InstallDir => $installDir, ConfDir => $confDir eq "" ? '/data/BackupPC/conf' : $confDir, LogDir => '/var/log/BackupPC', }; } else { $paths = { useFHS => $useFHS, TopDir => $topDir, InstallDir => $installDir, ConfDir => $confDir eq "" ? "$topDir/conf" : $confDir, LogDir => "$topDir/log", }; } my $bpc = bless { %$paths, Version => '3.2.0', }, $class; $bpc->{storage} = BackupPC::Storage->new($paths); # # Clean up %ENV and setup other variables. # delete @ENV{qw(IFS CDPATH ENV BASH_ENV)}; if ( defined(my $error = $bpc->ConfigRead()) ) { print(STDERR $error, "\n"); return; } # # Update the paths based on the config file # foreach my $dir ( qw(TopDir ConfDir InstallDir LogDir) ) { next if ( $bpc->{Conf}{$dir} eq "" ); $paths->{$dir} = $bpc->{$dir} = $bpc->{Conf}{$dir}; } $bpc->{storage}->setPaths($paths); $bpc->{PoolDir} = "$bpc->{TopDir}/pool"; $bpc->{CPoolDir} = "$bpc->{TopDir}/cpool"; # # Verify we are running as the correct user # if ( !$noUserCheck && $bpc->{Conf}{BackupPCUserVerify} && $> != (my $uid = (getpwnam($bpc->{Conf}{BackupPCUser}))[2]) ) { print(STDERR "$0: Wrong user: my userid is $>, instead of $uid" . " ($bpc->{Conf}{BackupPCUser})\n"); print(STDERR "Please su $bpc->{Conf}{BackupPCUser} first\n"); return; } return $bpc; } sub TopDir { my($bpc) = @_; return $bpc->{TopDir}; } sub BinDir { my($bpc) = @_; return "$bpc->{InstallDir}/bin"; } sub LogDir { my($bpc) = @_; return $bpc->{LogDir}; } sub ConfDir { my($bpc) = @_; return $bpc->{ConfDir}; } sub LibDir { my($bpc) = @_; return "$bpc->{InstallDir}/lib"; } sub InstallDir { my($bpc) = @_; return $bpc->{InstallDir}; } sub useFHS { my($bpc) = @_; return $bpc->{useFHS}; } sub Version { my($bpc) = @_; return $bpc->{Version}; } sub Conf { my($bpc) = @_; return %{$bpc->{Conf}}; } sub Lang { my($bpc) = @_; return $bpc->{Lang}; } sub adminJob { my($bpc, $num) = @_; return " admin " if ( !$num ); return " admin$num "; } sub isAdminJob { my($bpc, $str) = @_; return $str =~ /^ admin/; } sub trashJob { return " trashClean "; } sub ConfValue { my($bpc, $param) = @_; return $bpc->{Conf}{$param}; } sub verbose { my($bpc, $param) = @_; $bpc->{verbose} = $param if ( defined($param) ); return $bpc->{verbose}; } sub sigName2num { my($bpc, $sig) = @_; if ( !defined($bpc->{SigName2Num}) ) { my $i = 0; foreach my $name ( split(' ', $Config{sig_name}) ) { $bpc->{SigName2Num}{$name} = $i; $i++; } } return $bpc->{SigName2Num}{$sig}; } # # Generate an ISO 8601 format timeStamp (but without the "T"). # See http://www.w3.org/TR/NOTE-datetime and # http://www.cl.cam.ac.uk/~mgk25/iso-time.html # sub timeStamp { my($bpc, $t, $noPad) = @_; my($sec,$min,$hour,$mday,$mon,$year,$wday,$yday,$isdst) = localtime($t || time); return sprintf("%04d-%02d-%02d %02d:%02d:%02d", $year + 1900, $mon + 1, $mday, $hour, $min, $sec) . ($noPad ? "" : " "); } sub BackupInfoRead { my($bpc, $host) = @_; return $bpc->{storage}->BackupInfoRead($host); } sub BackupInfoWrite { my($bpc, $host, @Backups) = @_; return $bpc->{storage}->BackupInfoWrite($host, @Backups); } sub RestoreInfoRead { my($bpc, $host) = @_; return $bpc->{storage}->RestoreInfoRead($host); } sub RestoreInfoWrite { my($bpc, $host, @Restores) = @_; return $bpc->{storage}->RestoreInfoWrite($host, @Restores); } sub ArchiveInfoRead { my($bpc, $host) = @_; return $bpc->{storage}->ArchiveInfoRead($host); } sub ArchiveInfoWrite { my($bpc, $host, @Archives) = @_; return $bpc->{storage}->ArchiveInfoWrite($host, @Archives); } sub ConfigDataRead { my($bpc, $host) = @_; return $bpc->{storage}->ConfigDataRead($host); } sub ConfigDataWrite { my($bpc, $host, $conf) = @_; return $bpc->{storage}->ConfigDataWrite($host, $conf); } sub ConfigRead { my($bpc, $host) = @_; my($ret); # # Read main config file # my($mesg, $config) = $bpc->{storage}->ConfigDataRead(); return $mesg if ( defined($mesg) ); $bpc->{Conf} = $config; # # Read host config file # if ( $host ne "" ) { ($mesg, $config) = $bpc->{storage}->ConfigDataRead($host, $config); return $mesg if ( defined($mesg) ); $bpc->{Conf} = $config; } # # Load optional perl modules # if ( defined($bpc->{Conf}{PerlModuleLoad}) ) { # # Load any user-specified perl modules. This is for # optional user-defined extensions. # $bpc->{Conf}{PerlModuleLoad} = [$bpc->{Conf}{PerlModuleLoad}] if ( ref($bpc->{Conf}{PerlModuleLoad}) ne "ARRAY" ); foreach my $module ( @{$bpc->{Conf}{PerlModuleLoad}} ) { eval("use $module;"); } } # # Load language file # return "No language setting" if ( !defined($bpc->{Conf}{Language}) ); my $langFile = "$bpc->{InstallDir}/lib/BackupPC/Lang/$bpc->{Conf}{Language}.pm"; if ( !defined($ret = do $langFile) && ($! || $@) ) { $mesg = "Couldn't open language file $langFile: $!" if ( $! ); $mesg = "Couldn't execute language file $langFile: $@" if ( $@ ); $mesg =~ s/[\n\r]+//; return $mesg; } $bpc->{Lang} = \%Lang; # # Make sure IncrLevels is defined # $bpc->{Conf}{IncrLevels} = [1] if ( !defined($bpc->{Conf}{IncrLevels}) ); return; } # # Return the mtime of the config file # sub ConfigMTime { my($bpc) = @_; return $bpc->{storage}->ConfigMTime(); } # # Returns information from the host file in $bpc->{TopDir}/conf/hosts. # With no argument a ref to a hash of hosts is returned. Each # hash contains fields as specified in the hosts file. With an # argument a ref to a single hash is returned with information # for just that host. # sub HostInfoRead { my($bpc, $host) = @_; return $bpc->{storage}->HostInfoRead($host); } sub HostInfoWrite { my($bpc, $host) = @_; return $bpc->{storage}->HostInfoWrite($host); } # # Return the mtime of the hosts file # sub HostsMTime { my($bpc) = @_; return $bpc->{storage}->HostsMTime(); } # # Read a directory and return the entries in sorted inode order. # This relies on the IO::Dirent module being installed. If not, # the inode data is empty and the default directory order is # returned. # # The returned data is a list of hashes with entries {name, type, inode, nlink}. # The returned data includes "." and "..". # # $need is a hash of file attributes we need: type, inode, or nlink. # If set, these parameters are added to the returned hash. # # To support browsing pre-3.0.0 backups where the charset encoding # is typically iso-8859-1, the charsetLegacy option can be set in # $need to convert the path from utf8 and convert the names to utf8. # # If IO::Dirent is successful if will get type and inode for free. # Otherwise, a stat is done on each file, which is more expensive. # sub dirRead { my($bpc, $path, $need) = @_; my(@entries, $addInode); from_to($path, "utf8", $need->{charsetLegacy}) if ( $need->{charsetLegacy} ne "" ); return if ( !opendir(my $fh, $path) ); if ( $IODirentLoaded && !$IODirentOk ) { # # Make sure the IO::Dirent really works - some installs # on certain file systems (eg: XFS) don't return a valid type. # if ( opendir(my $fh, $bpc->{TopDir}) ) { my $dt_dir = eval("DT_DIR"); foreach my $e ( readdirent($fh) ) { if ( $e->{name} eq "." && $e->{type} == $dt_dir ) { $IODirentOk = 1; last; } } closedir($fh); } # # if it isn't ok then don't check again. # $IODirentLoaded = 0 if ( !$IODirentOk ); } if ( $IODirentOk ) { @entries = sort({ $a->{inode} <=> $b->{inode} } readdirent($fh)); map { $_->{type} = 0 + $_->{type} } @entries; # make type numeric } else { @entries = map { { name => $_} } readdir($fh); } closedir($fh); if ( defined($need) ) { for ( my $i = 0 ; $i < @entries ; $i++ ) { next if ( (!$need->{inode} || defined($entries[$i]{inode})) && (!$need->{type} || defined($entries[$i]{type})) && (!$need->{nlink} || defined($entries[$i]{nlink})) ); my @s = stat("$path/$entries[$i]{name}"); $entries[$i]{nlink} = $s[3] if ( $need->{nlink} ); if ( $need->{inode} && !defined($entries[$i]{inode}) ) { $addInode = 1; $entries[$i]{inode} = $s[1]; } if ( $need->{type} && !defined($entries[$i]{type}) ) { my $mode = S_IFMT($s[2]); $entries[$i]{type} = BPC_DT_FIFO if ( S_ISFIFO($mode) ); $entries[$i]{type} = BPC_DT_CHR if ( S_ISCHR($mode) ); $entries[$i]{type} = BPC_DT_DIR if ( S_ISDIR($mode) ); $entries[$i]{type} = BPC_DT_BLK if ( S_ISBLK($mode) ); $entries[$i]{type} = BPC_DT_REG if ( S_ISREG($mode) ); $entries[$i]{type} = BPC_DT_LNK if ( S_ISLNK($mode) ); $entries[$i]{type} = BPC_DT_SOCK if ( S_ISSOCK($mode) ); } } } # # Sort the entries if inodes were added (the IO::Dirent case already # sorted above) # @entries = sort({ $a->{inode} <=> $b->{inode} } @entries) if ( $addInode ); # # for browing pre-3.0.0 backups, map iso-8859-1 to utf8 if requested # if ( $need->{charsetLegacy} ne "" ) { for ( my $i = 0 ; $i < @entries ; $i++ ) { from_to($entries[$i]{name}, $need->{charsetLegacy}, "utf8"); } } return \@entries; } # # Same as dirRead, but only returns the names (which will be sorted in # inode order if IO::Dirent is installed) # sub dirReadNames { my($bpc, $path, $need) = @_; my $entries = $bpc->dirRead($path, $need); return if ( !defined($entries) ); my @names = map { $_->{name} } @$entries; return \@names; } sub find { my($bpc, $param, $dir, $dontDoCwd) = @_; return if ( !chdir($dir) ); my $entries = $bpc->dirRead(".", {inode => 1, type => 1}); #print Dumper($entries); foreach my $f ( @$entries ) { next if ( $f->{name} eq ".." || $f->{name} eq "." && $dontDoCwd ); $param->{wanted}($f->{name}, "$dir/$f->{name}"); next if ( $f->{type} != BPC_DT_DIR || $f->{name} eq "." ); chdir($f->{name}); $bpc->find($param, "$dir/$f->{name}", 1); return if ( !chdir("..") ); } } # # Stripped down from File::Path. In particular we don't print # many warnings and we try three times to delete each directory # and file -- for some reason the original File::Path rmtree # didn't always completely remove a directory tree on a NetApp. # # Warning: this routine changes the cwd. # sub RmTreeQuiet { my($bpc, $pwd, $roots) = @_; my(@files, $root); if ( defined($roots) && length($roots) ) { $roots = [$roots] unless ref $roots; } else { print(STDERR "RmTreeQuiet: No root path(s) specified\n"); } chdir($pwd); foreach $root (@{$roots}) { $root = $1 if ( $root =~ m{(.*?)/*$} ); # # Try first to simply unlink the file: this avoids an # extra stat for every file. If it fails (which it # will for directories), check if it is a directory and # then recurse. # if ( !unlink($root) ) { if ( -d $root ) { my $d = $bpc->dirReadNames($root); if ( !defined($d) ) { print(STDERR "Can't read $pwd/$root: $!\n"); } else { @files = grep $_ !~ /^\.{1,2}$/, @$d; $bpc->RmTreeQuiet("$pwd/$root", \@files); chdir($pwd); rmdir($root) || rmdir($root); } } else { unlink($root) || unlink($root); } } } } # # Move a directory or file away for later deletion # sub RmTreeDefer { my($bpc, $trashDir, $file) = @_; my($i, $f); return if ( !-e $file ); if ( !-d $trashDir ) { eval { mkpath($trashDir, 0, 0777) }; if ( $@ ) { # # There's no good place to send this error - use stderr # print(STDERR "RmTreeDefer: can't create directory $trashDir"); } } for ( $i = 0 ; $i < 1000 ; $i++ ) { $f = sprintf("%s/%d_%d_%d", $trashDir, time, $$, $i); next if ( -e $f ); return if ( rename($file, $f) ); } # shouldn't get here, but might if you tried to call this # across file systems.... just remove the tree right now. if ( $file =~ /(.*)\/([^\/]*)/ ) { my($d) = $1; my($f) = $2; my($cwd) = Cwd::fastcwd(); $cwd = $1 if ( $cwd =~ /(.*)/ ); $bpc->RmTreeQuiet($d, $f); chdir($cwd) if ( $cwd ); } } # # Empty the trash directory. Returns 0 if it did nothing, 1 if it # did something, -1 if it failed to remove all the files. # sub RmTreeTrashEmpty { my($bpc, $trashDir) = @_; my(@files); my($cwd) = Cwd::fastcwd(); $cwd = $1 if ( $cwd =~ /(.*)/ ); return if ( !-d $trashDir ); my $d = $bpc->dirReadNames($trashDir) or carp "Can't read $trashDir: $!"; @files = grep $_ !~ /^\.{1,2}$/, @$d; return 0 if ( !@files ); $bpc->RmTreeQuiet($trashDir, \@files); foreach my $f ( @files ) { return -1 if ( -e $f ); } chdir($cwd) if ( $cwd ); return 1; } # # Open a connection to the server. Returns an error string on failure. # Returns undef on success. # sub ServerConnect { my($bpc, $host, $port, $justConnect) = @_; local(*FH); return if ( defined($bpc->{ServerFD}) ); # # First try the unix-domain socket # my $sockFile = "$bpc->{LogDir}/BackupPC.sock"; socket(*FH, PF_UNIX, SOCK_STREAM, 0) || return "unix socket: $!"; if ( !connect(*FH, sockaddr_un($sockFile)) ) { my $err = "unix connect: $!"; close(*FH); if ( $port > 0 ) { my $proto = getprotobyname('tcp'); my $iaddr = inet_aton($host) || return "unknown host $host"; my $paddr = sockaddr_in($port, $iaddr); socket(*FH, PF_INET, SOCK_STREAM, $proto) || return "inet socket: $!"; connect(*FH, $paddr) || return "inet connect: $!"; } else { return $err; } } my($oldFH) = select(*FH); $| = 1; select($oldFH); $bpc->{ServerFD} = *FH; return if ( $justConnect ); # # Read the seed that we need for our MD5 message digest. See # ServerMesg below. # sysread($bpc->{ServerFD}, $bpc->{ServerSeed}, 1024); $bpc->{ServerMesgCnt} = 0; return; } # # Check that the server connection is still ok # sub ServerOK { my($bpc) = @_; return 0 if ( !defined($bpc->{ServerFD}) ); vec(my $FDread, fileno($bpc->{ServerFD}), 1) = 1; my $ein = $FDread; return 0 if ( select(my $rout = $FDread, undef, $ein, 0.0) < 0 ); return 1 if ( !vec($rout, fileno($bpc->{ServerFD}), 1) ); } # # Disconnect from the server # sub ServerDisconnect { my($bpc) = @_; return if ( !defined($bpc->{ServerFD}) ); close($bpc->{ServerFD}); delete($bpc->{ServerFD}); } # # Sends a message to the server and returns with the reply. # # To avoid possible attacks via the TCP socket interface, every client # message is protected by an MD5 digest. The MD5 digest includes four # items: # - a seed that is sent to us when we first connect # - a sequence number that increments for each message # - a shared secret that is stored in $Conf{ServerMesgSecret} # - the message itself. # The message is sent in plain text preceded by the MD5 digest. A # snooper can see the plain-text seed sent by BackupPC and plain-text # message, but cannot construct a valid MD5 digest since the secret in # $Conf{ServerMesgSecret} is unknown. A replay attack is not possible # since the seed changes on a per-connection and per-message basis. # sub ServerMesg { my($bpc, $mesg) = @_; return if ( !defined(my $fh = $bpc->{ServerFD}) ); $mesg =~ s/\n/\\n/g; $mesg =~ s/\r/\\r/g; my $md5 = Digest::MD5->new; $mesg = encode_utf8($mesg); $md5->add($bpc->{ServerSeed} . $bpc->{ServerMesgCnt} . $bpc->{Conf}{ServerMesgSecret} . $mesg); print($fh $md5->b64digest . " $mesg\n"); $bpc->{ServerMesgCnt}++; return <$fh>; } # # Do initialization for child processes # sub ChildInit { my($bpc) = @_; close(STDERR); open(STDERR, ">&STDOUT"); select(STDERR); $| = 1; select(STDOUT); $| = 1; $ENV{PATH} = $bpc->{Conf}{MyPath}; } # # Compute the MD5 digest of a file. For efficiency we don't # use the whole file for big files: # - for files <= 256K we use the file size and the whole file. # - for files <= 1M we use the file size, the first 128K and # the last 128K. # - for files > 1M, we use the file size, the first 128K and # the 8th 128K (ie: the 128K up to 1MB). # See the documentation for a discussion of the tradeoffs in # how much data we use and how many collisions we get. # # Returns the MD5 digest (a hex string) and the file size. # sub File2MD5 { my($bpc, $md5, $name) = @_; my($data, $fileSize); local(*N); $fileSize = (stat($name))[7]; return ("", -1) if ( !-f _ ); $name = $1 if ( $name =~ /(.*)/ ); return ("", 0) if ( $fileSize == 0 ); return ("", -1) if ( !open(N, $name) ); binmode(N); $md5->reset(); $md5->add($fileSize); if ( $fileSize > 262144 ) { # # read the first and last 131072 bytes of the file, # up to 1MB. # my $seekPosn = ($fileSize > 1048576 ? 1048576 : $fileSize) - 131072; $md5->add($data) if ( sysread(N, $data, 131072) ); $md5->add($data) if ( sysseek(N, $seekPosn, 0) && sysread(N, $data, 131072) ); } else { # # read the whole file # $md5->add($data) if ( sysread(N, $data, $fileSize) ); } close(N); return ($md5->hexdigest, $fileSize); } # # Compute the MD5 digest of a buffer (string). For efficiency we don't # use the whole string for big strings: # - for files <= 256K we use the file size and the whole file. # - for files <= 1M we use the file size, the first 128K and # the last 128K. # - for files > 1M, we use the file size, the first 128K and # the 8th 128K (ie: the 128K up to 1MB). # See the documentation for a discussion of the tradeoffs in # how much data we use and how many collisions we get. # # Returns the MD5 digest (a hex string). # sub Buffer2MD5 { my($bpc, $md5, $fileSize, $dataRef) = @_; $md5->reset(); $md5->add($fileSize); if ( $fileSize > 262144 ) { # # add the first and last 131072 bytes of the string, # up to 1MB. # my $seekPosn = ($fileSize > 1048576 ? 1048576 : $fileSize) - 131072; $md5->add(substr($$dataRef, 0, 131072)); $md5->add(substr($$dataRef, $seekPosn, 131072)); } else { # # add the whole string # $md5->add($$dataRef); } return $md5->hexdigest; } # # Given an MD5 digest $d and a compress flag, return the full # path in the pool. # sub MD52Path { my($bpc, $d, $compress, $poolDir) = @_; return if ( $d !~ m{(.)(.)(.)(.*)} ); $poolDir = ($compress ? $bpc->{CPoolDir} : $bpc->{PoolDir}) if ( !defined($poolDir) ); return "$poolDir/$1/$2/$3/$1$2$3$4"; } # # For each file, check if the file exists in $bpc->{TopDir}/pool. # If so, remove the file and make a hardlink to the file in # the pool. Otherwise, if the newFile flag is set, make a # hardlink in the pool to the new file. # # Returns 0 if a link should be made to a new file (ie: when the file # is a new file but the newFile flag is 0). # Returns 1 if a link to an existing file is made, # Returns 2 if a link to a new file is made (only if $newFile is set) # Returns negative on error. # sub MakeFileLink { my($bpc, $name, $d, $newFile, $compress) = @_; my($i, $rawFile); return -1 if ( !-f $name ); for ( $i = -1 ; ; $i++ ) { return -2 if ( !defined($rawFile = $bpc->MD52Path($d, $compress)) ); $rawFile .= "_$i" if ( $i >= 0 ); if ( -f $rawFile ) { if ( (stat(_))[3] < $bpc->{Conf}{HardLinkMax} && !compare($name, $rawFile) ) { unlink($name); return -3 if ( !link($rawFile, $name) ); return 1; } } elsif ( $newFile && -f $name && (stat($name))[3] == 1 ) { my($newDir); ($newDir = $rawFile) =~ s{(.*)/.*}{$1}; if ( !-d $newDir ) { eval { mkpath($newDir, 0, 0777) }; return -5 if ( $@ ); } return -4 if ( !link($name, $rawFile) ); return 2; } else { return 0; } } } # # Tests if we can create a hardlink from a file in directory # $newDir to a file in directory $targetDir. A temporary # file in $targetDir is created and an attempt to create a # hardlink of the same name in $newDir is made. The temporary # files are removed. # # Like link(), returns true on success and false on failure. # sub HardlinkTest { my($bpc, $targetDir, $newDir) = @_; my($targetFile, $newFile, $fd); for ( my $i = 0 ; ; $i++ ) { $targetFile = "$targetDir/.TestFileLink.$$.$i"; $newFile = "$newDir/.TestFileLink.$$.$i"; last if ( !-e $targetFile && !-e $newFile ); } return 0 if ( !open($fd, ">", $targetFile) ); close($fd); my $ret = link($targetFile, $newFile); unlink($targetFile); unlink($newFile); return $ret; } sub CheckHostAlive { my($bpc, $host) = @_; my($s, $pingCmd, $ret); # # Return success if the ping cmd is undefined or empty. # if ( $bpc->{Conf}{PingCmd} eq "" ) { print(STDERR "CheckHostAlive: return ok because \$Conf{PingCmd}" . " is empty\n") if ( $bpc->{verbose} ); return 0; } my $args = { pingPath => $bpc->{Conf}{PingPath}, host => $host, }; $pingCmd = $bpc->cmdVarSubstitute($bpc->{Conf}{PingCmd}, $args); # # Do a first ping in case the PC needs to wakeup # $s = $bpc->cmdSystemOrEval($pingCmd, undef, $args); if ( $? ) { print(STDERR "CheckHostAlive: first ping failed ($?, $!)\n") if ( $bpc->{verbose} ); return -1; } # # Do a second ping and get the round-trip time in msec # $s = $bpc->cmdSystemOrEval($pingCmd, undef, $args); if ( $? ) { print(STDERR "CheckHostAlive: second ping failed ($?, $!)\n") if ( $bpc->{verbose} ); return -1; } if ( $s =~ /rtt\s*min\/avg\/max\/mdev\s*=\s*[\d.]+\/([\d.]+)\/[\d.]+\/[\d.]+\s*(ms|usec)/i ) { $ret = $1; $ret /= 1000 if ( lc($2) eq "usec" ); } elsif ( $s =~ /time=([\d.]+)\s*(ms|usec)/i ) { $ret = $1; $ret /= 1000 if ( lc($2) eq "usec" ); } else { print(STDERR "CheckHostAlive: can't extract round-trip time" . " (not fatal)\n") if ( $bpc->{verbose} ); $ret = 0; } print(STDERR "CheckHostAlive: returning $ret\n") if ( $bpc->{verbose} ); return $ret; } sub CheckFileSystemUsage { my($bpc) = @_; my($topDir) = $bpc->{TopDir}; my($s, $dfCmd); return 0 if ( $bpc->{Conf}{DfCmd} eq "" ); my $args = { dfPath => $bpc->{Conf}{DfPath}, topDir => $bpc->{TopDir}, }; $dfCmd = $bpc->cmdVarSubstitute($bpc->{Conf}{DfCmd}, $args); $s = $bpc->cmdSystemOrEval($dfCmd, undef, $args); return 0 if ( $? || $s !~ /(\d+)%/s ); return $1; } # # Given an IP address, return the host name and user name via # NetBios. # sub NetBiosInfoGet { my($bpc, $host) = @_; my($netBiosHostName, $netBiosUserName); my($s, $nmbCmd); # # Skip NetBios check if NmbLookupCmd is emtpy # if ( $bpc->{Conf}{NmbLookupCmd} eq "" ) { print(STDERR "NetBiosInfoGet: return $host because \$Conf{NmbLookupCmd}" . " is empty\n") if ( $bpc->{verbose} ); return ($host, undef); } my $args = { nmbLookupPath => $bpc->{Conf}{NmbLookupPath}, host => $host, }; $nmbCmd = $bpc->cmdVarSubstitute($bpc->{Conf}{NmbLookupCmd}, $args); foreach ( split(/[\n\r]+/, $bpc->cmdSystemOrEval($nmbCmd, undef, $args)) ) { # # skip and other non entries # next if ( /<\w{2}> - /i ); next if ( !/^\s*([\w\s-]+?)\s*<(\w{2})\> - .*/i ); $netBiosHostName ||= $1 if ( $2 eq "00" ); # host is first 00 $netBiosUserName = $1 if ( $2 eq "03" ); # user is last 03 } if ( !defined($netBiosHostName) ) { print(STDERR "NetBiosInfoGet: failed: can't parse return string\n") if ( $bpc->{verbose} ); return; } $netBiosHostName = lc($netBiosHostName); $netBiosUserName = lc($netBiosUserName); print(STDERR "NetBiosInfoGet: success, returning host $netBiosHostName," . " user $netBiosUserName\n") if ( $bpc->{verbose} ); return ($netBiosHostName, $netBiosUserName); } # # Given a NetBios name lookup the IP address via NetBios. # In the case of a host returning multiple interfaces we # return the first IP address that matches the subnet mask. # If none match the subnet mask (or nmblookup doesn't print # the subnet mask) then just the first IP address is returned. # sub NetBiosHostIPFind { my($bpc, $host) = @_; my($netBiosHostName, $netBiosUserName); my($s, $nmbCmd, $subnet, $ipAddr, $firstIpAddr); # # Skip NetBios lookup if NmbLookupFindHostCmd is emtpy # if ( $bpc->{Conf}{NmbLookupFindHostCmd} eq "" ) { print(STDERR "NetBiosHostIPFind: return $host because" . " \$Conf{NmbLookupFindHostCmd} is empty\n") if ( $bpc->{verbose} ); return $host; } my $args = { nmbLookupPath => $bpc->{Conf}{NmbLookupPath}, host => $host, }; $nmbCmd = $bpc->cmdVarSubstitute($bpc->{Conf}{NmbLookupFindHostCmd}, $args); foreach my $resp ( split(/[\n\r]+/, $bpc->cmdSystemOrEval($nmbCmd, undef, $args) ) ) { if ( $resp =~ /querying\s+\Q$host\E\s+on\s+(\d+\.\d+\.\d+\.\d+)/i ) { $subnet = $1; $subnet = $1 if ( $subnet =~ /^(.*?)(\.255)+$/ ); } elsif ( $resp =~ /^\s*(\d+\.\d+\.\d+\.\d+)\s+\Q$host/ ) { my $ip = $1; $firstIpAddr = $ip if ( !defined($firstIpAddr) ); $ipAddr = $ip if ( !defined($ipAddr) && $ip =~ /^\Q$subnet/ ); } } $ipAddr = $firstIpAddr if ( !defined($ipAddr) ); if ( defined($ipAddr) ) { print(STDERR "NetBiosHostIPFind: found IP address $ipAddr for" . " host $host\n") if ( $bpc->{verbose} ); return $ipAddr; } else { print(STDERR "NetBiosHostIPFind: couldn't find IP address for" . " host $host\n") if ( $bpc->{verbose} ); return; } } sub fileNameEltMangle { my($bpc, $name) = @_; return "" if ( $name eq "" ); $name =~ s{([%/\n\r])}{sprintf("%%%02x", ord($1))}eg; return "f$name"; } # # We store files with every name preceded by "f". This # avoids possible name conflicts with other information # we store in the same directories (eg: attribute info). # The process of turning a normal path into one with each # node prefixed with "f" is called mangling. # sub fileNameMangle { my($bpc, $name) = @_; $name =~ s{/([^/]+)}{"/" . $bpc->fileNameEltMangle($1)}eg; $name =~ s{^([^/]+)}{$bpc->fileNameEltMangle($1)}eg; return $name; } # # This undoes FileNameMangle # sub fileNameUnmangle { my($bpc, $name) = @_; $name =~ s{/f}{/}g; $name =~ s{^f}{}; $name =~ s{%(..)}{chr(hex($1))}eg; return $name; } # # Escape shell meta-characters with backslashes. # This should be applied to each argument seperately, not an # entire shell command. # sub shellEscape { my($bpc, $cmd) = @_; $cmd =~ s/([][;&()<>{}|^\n\r\t *\$\\'"`?])/\\$1/g; return $cmd; } # # For printing exec commands (which don't use a shell) so they look like # a valid shell command this function should be called with the exec # args. The shell command string is returned. # sub execCmd2ShellCmd { my($bpc, @args) = @_; my $str; foreach my $a ( @args ) { $str .= " " if ( $str ne "" ); $str .= $bpc->shellEscape($a); } return $str; } # # Do a URI-style escape to protect/encode special characters # sub uriEsc { my($bpc, $s) = @_; $s =~ s{([^\w.\/-])}{sprintf("%%%02X", ord($1));}eg; return $s; } # # Do a URI-style unescape to restore special characters # sub uriUnesc { my($bpc, $s) = @_; $s =~ s{%(..)}{chr(hex($1))}eg; return $s; } # # Do variable substitution prior to execution of a command. # sub cmdVarSubstitute { my($bpc, $template, $vars) = @_; my(@cmd); # # Return without any substitution if the first entry starts with "&", # indicating this is perl code. # if ( (ref($template) eq "ARRAY" ? $template->[0] : $template) =~ /^\&/ ) { return $template; } if ( ref($template) ne "ARRAY" ) { # # Split at white space, except if escaped by \ # $template = [split(/(?{$1}) && ref($vars->{$1}) ne "ARRAY" ? ($2 eq "+" ? $bpc->shellEscape($vars->{$1}) : $vars->{$1}) : "\${$1}$2" }eg; # # Now replicate any array arguments; this just works for just one # array var in each argument. # if ( $arg =~ m[(.*)\${(\w+)}(\+?)(.*)] && ref($vars->{$2}) eq "ARRAY" ) { my $pre = $1; my $var = $2; my $esc = $3; my $post = $4; foreach my $v ( @{$vars->{$var}} ) { $v = $bpc->shellEscape($v) if ( $esc eq "+" ); push(@cmd, "$pre$v$post"); } } else { push(@cmd, $arg); } } return \@cmd; } # # Exec or eval a command. $cmd is either a string on an array ref. # # @args are optional arguments for the eval() case; they are not used # for exec(). # sub cmdExecOrEval { my($bpc, $cmd, @args) = @_; if ( (ref($cmd) eq "ARRAY" ? $cmd->[0] : $cmd) =~ /^\&/ ) { $cmd = join(" ", $cmd) if ( ref($cmd) eq "ARRAY" ); print(STDERR "cmdExecOrEval: about to eval perl code $cmd\n") if ( $bpc->{verbose} ); eval($cmd); print(STDERR "Perl code fragment for exec shouldn't return!!\n"); exit(1); } else { $cmd = [split(/\s+/, $cmd)] if ( ref($cmd) ne "ARRAY" ); print(STDERR "cmdExecOrEval: about to exec ", $bpc->execCmd2ShellCmd(@$cmd), "\n") if ( $bpc->{verbose} ); alarm(0); $cmd = [map { m/(.*)/ } @$cmd]; # untaint # # force list-form of exec(), ie: no shell even for 1 arg # exec { $cmd->[0] } @$cmd; print(STDERR "Exec failed for @$cmd\n"); exit(1); } } # # System or eval a command. $cmd is either a string on an array ref. # $stdoutCB is a callback for output generated by the command. If it # is undef then output is returned. If it is a code ref then the function # is called with each piece of output as an argument. If it is a scalar # ref the output is appended to this variable. # # @args are optional arguments for the eval() case; they are not used # for system(). # # Also, $? should be set when the CHILD pipe is closed. # sub cmdSystemOrEvalLong { my($bpc, $cmd, $stdoutCB, $ignoreStderr, $pidHandlerCB, @args) = @_; my($pid, $out, $allOut); local(*CHILD); $? = 0; if ( (ref($cmd) eq "ARRAY" ? $cmd->[0] : $cmd) =~ /^\&/ ) { $cmd = join(" ", $cmd) if ( ref($cmd) eq "ARRAY" ); print(STDERR "cmdSystemOrEval: about to eval perl code $cmd\n") if ( $bpc->{verbose} ); $out = eval($cmd); $$stdoutCB .= $out if ( ref($stdoutCB) eq 'SCALAR' ); &$stdoutCB($out) if ( ref($stdoutCB) eq 'CODE' ); print(STDERR "cmdSystemOrEval: finished: got output $out\n") if ( $bpc->{verbose} ); return $out if ( !defined($stdoutCB) ); return; } else { $cmd = [split(/\s+/, $cmd)] if ( ref($cmd) ne "ARRAY" ); print(STDERR "cmdSystemOrEval: about to system ", $bpc->execCmd2ShellCmd(@$cmd), "\n") if ( $bpc->{verbose} ); if ( !defined($pid = open(CHILD, "-|")) ) { my $err = "Can't fork to run @$cmd\n"; $? = 1; $$stdoutCB .= $err if ( ref($stdoutCB) eq 'SCALAR' ); &$stdoutCB($err) if ( ref($stdoutCB) eq 'CODE' ); return $err if ( !defined($stdoutCB) ); return; } binmode(CHILD); if ( !$pid ) { # # This is the child # close(STDERR); if ( $ignoreStderr ) { open(STDERR, ">", "/dev/null"); } else { open(STDERR, ">&STDOUT"); } alarm(0); $cmd = [map { m/(.*)/ } @$cmd]; # untaint # # force list-form of exec(), ie: no shell even for 1 arg # exec { $cmd->[0] } @$cmd; print(STDERR "Exec of @$cmd failed\n"); exit(1); } # # Notify caller of child's pid # &$pidHandlerCB($pid) if ( ref($pidHandlerCB) eq "CODE" ); # # The parent gathers the output from the child # while ( ) { $$stdoutCB .= $_ if ( ref($stdoutCB) eq 'SCALAR' ); &$stdoutCB($_) if ( ref($stdoutCB) eq 'CODE' ); $out .= $_ if ( !defined($stdoutCB) ); $allOut .= $_ if ( $bpc->{verbose} ); } $? = 0; close(CHILD); } print(STDERR "cmdSystemOrEval: finished: got output $allOut\n") if ( $bpc->{verbose} ); return $out; } # # The shorter version that sets $ignoreStderr = 0, ie: merges stdout # and stderr together. # sub cmdSystemOrEval { my($bpc, $cmd, $stdoutCB, @args) = @_; return $bpc->cmdSystemOrEvalLong($cmd, $stdoutCB, 0, undef, @args); } # # Promotes $conf->{BackupFilesOnly}, $conf->{BackupFilesExclude} # to hashes and $conf->{$shareName} to an array. # sub backupFileConfFix { my($bpc, $conf, $shareName) = @_; $conf->{$shareName} = [ $conf->{$shareName} ] if ( ref($conf->{$shareName}) ne "ARRAY" ); foreach my $param qw(BackupFilesOnly BackupFilesExclude) { next if ( !defined($conf->{$param}) ); if ( ref($conf->{$param}) eq "HASH" ) { # # A "*" entry means wildcard - it is the default for # all shares. Replicate the "*" entry for all shares, # but still allow override of specific entries. # next if ( !defined($conf->{$param}{"*"}) ); $conf->{$param} = { map({ $_ => $conf->{$param}{"*"} } @{$conf->{$shareName}}), %{$conf->{$param}} }; } else { $conf->{$param} = [ $conf->{$param} ] if ( ref($conf->{$param}) ne "ARRAY" ); $conf->{$param} = { map { $_ => $conf->{$param} } @{$conf->{$shareName}} }; } } } # # This is sort() compare function, used below. # # New client LOG names are LOG.MMYYYY. Old style names are # LOG, LOG.0, LOG.1 etc. Sort them so new names are # first, and newest to oldest. # sub compareLOGName { my $na = $1 if ( $a =~ /LOG\.(\d+)(\.z)?$/ ); my $nb = $1 if ( $b =~ /LOG\.(\d+)(\.z)?$/ ); $na = -1 if ( !defined($na) ); $nb = -1 if ( !defined($nb) ); if ( length($na) >= 5 && length($nb) >= 5 ) { # # Both new style: format is MMYYYY. Bigger dates are # more recent. # my $ma = $2 * 12 + $1 if ( $na =~ /(\d+)(\d{4})/ ); my $mb = $2 * 12 + $1 if ( $nb =~ /(\d+)(\d{4})/ ); return $mb - $ma; } elsif ( length($na) >= 5 && length($nb) < 5 ) { return -1; } elsif ( length($na) < 5 && length($nb) >= 5 ) { return 1; } else { # # Both old style. Smaller numbers are more recent. # return $na - $nb; } } # # Returns list of paths to a clients's (or main) LOG files, # most recent first. # sub sortedPCLogFiles { my($bpc, $host) = @_; my(@files, $dir); if ( $host ne "" ) { $dir = "$bpc->{TopDir}/pc/$host"; } else { $dir = "$bpc->{LogDir}"; } if ( opendir(DIR, $dir) ) { foreach my $file ( readdir(DIR) ) { next if ( !-f "$dir/$file" ); next if ( $file ne "LOG" && $file !~ /^LOG\.\d/ ); push(@files, "$dir/$file"); } closedir(DIR); } return sort compareLOGName @files; } # # converts a glob-style pattern into a perl regular expression. # sub glob2re { my ( $bpc, $glob ) = @_; my ( $char, $subst ); # $escapeChars escapes characters with no special glob meaning but # have meaning in regexps. my $escapeChars = [ '.', '/', ]; # $charMap is where we implement the special meaning of glob # patterns and translate them to regexps. my $charMap = { '?' => '[^/]', '*' => '[^/]*', }; # multiple forward slashes are equivalent to one slash. We should # never have to use this. $glob =~ s/\/+/\//; foreach $char (@$escapeChars) { $glob =~ s/\Q$char\E/\\$char/g; } while ( ( $char, $subst ) = each(%$charMap) ) { $glob =~ s/(?