use Time::HiRes qw/time/;
use File::Pid;
use POSIX qw/strftime/;
+use BackupPC::SearchLib;
use constant BPC_FTYPE_DIR => 5;
-use constant EST_SYNC_EVERY => 10000;
+use constant EST_CHUNK => 100000;
my $debug = 0;
$|=1;
my $dsn = $Conf{SearchDSN} || die "Need SearchDSN in config.pl\n";
my $user = $Conf{SearchUser} || '';
-my $index_path = $Conf{HyperEstraierIndex};
-$index_path = $TopDir . '/' . $index_path;
-$index_path =~ s#//#/#g;
+my $use_hest = $Conf{HyperEstraierIndex};
+my ($index_path, $index_node_url) = BackupPC::SearchLib::getHyperEstraier_url($use_hest);
my $dbh = DBI->connect($dsn, $user, "", { RaiseError => 1, AutoCommit => 0 });
my %opt;
-if ( !getopts("cdm:v:i", \%opt ) ) {
+if ( !getopts("cdm:v:ij", \%opt ) ) {
print STDERR <<EOF;
usage: $0 [-c|-d] [-m num] [-v|-v level] [-i]
-d delete database before import
-m num import just num increments for one host
-v num set verbosity (debug) level (default $debug)
- -i update HyperEstraier full text index
+ -i update Hyper Estraier full text index
+ -j update full text, don't check existing files
+
+Option -j is variation on -i. It will allow faster initial creation
+of full-text index from existing database.
+
EOF
exit 1;
}
}
my $hest_db;
+my $hest_node;
sub signal {
my($sig) = @_;
my ($host_id, $share_id, $num) = @_;
- print curr_time," updating HyperEstraier: select files";
-
- my $t = time();
-
- my $where = '';
- if ($host_id && $share_id && $num) {
- $where = qq{
- WHERE
- hosts.id = ? AND
- shares.id = ? AND
- files.backupnum = ?
- };
- }
-
- my $sth = $dbh->prepare(qq{
- SELECT
- files.id AS fid,
- hosts.name AS hname,
- shares.name AS sname,
- -- shares.share AS sharename,
- files.backupnum AS backupnum,
- -- files.name AS filename,
- files.path AS filepath,
- files.date AS date,
- files.type AS type,
- files.size AS size,
- files.shareid AS shareid,
- backups.date AS backup_date
- FROM files
- INNER JOIN shares ON files.shareID=shares.ID
- INNER JOIN hosts ON hosts.ID = shares.hostID
- INNER JOIN backups ON backups.num = files.backupNum and backups.hostID = hosts.ID AND backups.shareID = shares.ID
- $where
- });
-
- $sth->execute(@_);
- my $results = $sth->rows;
+ my $skip_check = $opt{j} && print STDERR "Skipping check for existing files -- this should be used only with initital import\n";
- if ($results == 0) {
- print " - no files, skipping\n";
+ unless ($use_hest) {
+ print STDERR "HyperEstraier support not enabled in configuration\n";
return;
}
- my $dot = int($results / 15) || 1;
+ print curr_time," updating HyperEstraier:";
- print " $results ($dot/#)";
+ my $t = time();
- sub fmt_date {
- my $t = shift || return;
- my $iso = BackupPC::Lib::timeStamp($t);
- $iso =~ s/\s/T/;
- return $iso;
+ my $offset = 0;
+ my $added = 0;
+
+ print " opening index $use_hest";
+ if ($index_path) {
+ $hest_db = HyperEstraier::Database->new();
+ $hest_db->open($TopDir . $index_path, $HyperEstraier::Database::DBWRITER | $HyperEstraier::Database::DBCREAT);
+ print " directly";
+ } elsif ($index_node_url) {
+ $hest_node ||= HyperEstraier::Node->new($index_node_url);
+ $hest_node->set_auth('admin', 'admin');
+ print " via node URL";
+ } else {
+ die "don't know how to use HyperEstraier Index $use_hest";
}
+ print " increment is " . EST_CHUNK . " files:";
+
+ my $results = 0;
+
+ do {
+
+ my $where = '';
+ my @data;
+ if (defined($host_id) && defined($share_id) && defined($num)) {
+ $where = qq{
+ WHERE
+ hosts.id = ? AND
+ shares.id = ? AND
+ files.backupnum = ?
+ };
+ @data = ( $host_id, $share_id, $num );
+ }
- my $max = int($results / $dot);
+ my $limit = sprintf('LIMIT '.EST_CHUNK.' OFFSET %d', $offset);
+
+ my $sth = $dbh->prepare(qq{
+ SELECT
+ files.id AS fid,
+ hosts.name AS hname,
+ shares.name AS sname,
+ -- shares.share AS sharename,
+ files.backupnum AS backupnum,
+ -- files.name AS filename,
+ files.path AS filepath,
+ files.date AS date,
+ files.type AS type,
+ files.size AS size,
+ files.shareid AS shareid,
+ backups.date AS backup_date
+ FROM files
+ INNER JOIN shares ON files.shareID=shares.ID
+ INNER JOIN hosts ON hosts.ID = shares.hostID
+ INNER JOIN backups ON backups.num = files.backupNum and backups.hostID = hosts.ID AND backups.shareID = shares.ID
+ $where
+ $limit
+ });
+
+ $sth->execute(@data);
+ $results = $sth->rows;
+
+ if ($results == 0) {
+ print " - no new files\n";
+ last;
+ }
- print ", opening index $index_path...";
- use HyperEstraier;
- my $db = HyperEstraier::Database->new();
+ sub fmt_date {
+ my $t = shift || return;
+ my $iso = BackupPC::Lib::timeStamp($t);
+ $iso =~ s/\s/T/;
+ return $iso;
+ }
-# unless ($hest_db) {
-# print " open reader";
-# $hest_db = HyperEstraier::Database->new();
-#
-# }
+ while (my $row = $sth->fetchrow_hashref()) {
+ my $fid = $row->{'fid'} || die "no fid?";
+ my $uri = 'file:///' . $fid;
- $db->open($index_path, $HyperEstraier::Database::DBWRITER | $HyperEstraier::Database::DBCREAT);
+ unless ($skip_check) {
+ my $id = ($hest_db || $hest_node)->uri_to_id($uri);
+ next unless ($id == -1);
+ }
- my $added = 0;
+ # create a document object
+ my $doc = HyperEstraier::Document->new;
- while (my $row = $sth->fetchrow_hashref()) {
+ # add attributes to the document object
+ $doc->add_attr('@uri', $uri);
- my $fid = $row->{'fid'} || die "no fid?";
- my $uri = 'file:///' . $fid;
+ foreach my $c (@{ $sth->{NAME} }) {
+ $doc->add_attr($c, $row->{$c}) if ($row->{$c});
+ }
- my $id = $db->uri_to_id($uri);
- next unless ($id == -1);
+ #$doc->add_attr('@cdate', fmt_date($row->{'date'}));
- # create a document object
- my $doc = HyperEstraier::Document->new;
+ # add the body text to the document object
+ my $path = $row->{'filepath'};
+ $doc->add_text($path);
+ $path =~ s/(.)/$1 /g;
+ $doc->add_hidden_text($path);
- # add attributes to the document object
- $doc->add_attr('@uri', $uri);
+ print STDERR $doc->dump_draft,"\n" if ($debug > 1);
- foreach my $c (@{ $sth->{NAME} }) {
- $doc->add_attr($c, $row->{$c}) if ($row->{$c});
+ # register the document object to the database
+ if ($hest_db) {
+ $hest_db->put_doc($doc, $HyperEstraier::Database::PDCLEAN);
+ } elsif ($hest_node) {
+ $hest_node->put_doc($doc);
+ } else {
+ die "not supported";
+ }
+ $added++;
}
- #$doc->add_attr('@cdate', fmt_date($row->{'date'}));
+ print " $added";
+ $hest_db->sync() if ($index_path);
- # add the body text to the document object
- my $path = $row->{'filepath'};
- $doc->add_text($path);
- $path =~ s/(.)/$1 /g;
- $doc->add_hidden_text($path);
+ $offset += EST_CHUNK;
- print STDERR $doc->dump_draft,"\n" if ($debug > 1);
-
- # register the document object to the database
- $db->put_doc($doc, $HyperEstraier::Database::PDCLEAN);
-
- $added++;
- if ($added % $dot == 0) {
- print "$max ";
- $max--;
- }
-
- if ($added % EST_SYNC_EVERY == 0) {
- print "sync ";
- $db->sync();
- }
+ } while ($results == EST_CHUNK);
+ if ($index_path) {
+ print ", close";
+ $hest_db->close();
}
- print "sync $added new files";
- $db->sync();
- print ", close";
- $db->close();
-
my $dur = (time() - $t) || 1;
- printf(" [%.2f/s new %.2f/s dur: %s]\n",
- ( $results / $dur ),
+ printf(" [%.2f/s dur: %s]\n",
( $added / $dur ),
fmt_time($dur)
);
## update index ##
-if (($opt{i} || ($index_path && ! -e $index_path)) && !$opt{c}) {
+if (($opt{i} || $opt{j} || ($index_path && ! -e $index_path)) && !$opt{c}) {
# update all
print "force update of HyperEstraier index ";
print "importing existing data" unless (-e $index_path);
print "by -i flag" if ($opt{i});
+ print "by -j flag" if ($opt{j});
print "\n";
hest_update();
}
if ($opt{c}) {
sub do_index {
my $index = shift || return;
- my ($table,$col,$unique) = split(/_/, $index);
+ my ($table,$col,$unique) = split(/:/, $index);
$unique ||= '';
- $index =~ s/,/_/g;
+ $index =~ s/\W+/_/g;
+ print "$index on $table($col)" . ( $unique ? "u" : "" ) . " ";
$dbh->do(qq{ create $unique index $index on $table($col) });
}
print "creating tables...\n";
-
- $dbh->do(qq{
+
+ $dbh->do( qq{
create table hosts (
ID SERIAL PRIMARY KEY,
name VARCHAR(30) NOT NULL,
IP VARCHAR(15)
);
- });
-
- $dbh->do(qq{
+
create table shares (
ID SERIAL PRIMARY KEY,
hostID INTEGER NOT NULL references hosts(id),
name VARCHAR(30) NOT NULL,
- share VARCHAR(200) NOT NULL,
- localpath VARCHAR(200)
+ share VARCHAR(200) NOT NULL
);
- });
-
- $dbh->do(qq{
+
+ create table dvds (
+ ID SERIAL PRIMARY KEY,
+ num INTEGER NOT NULL,
+ name VARCHAR(255) NOT NULL,
+ mjesto VARCHAR(255)
+ );
+
create table backups (
+ id serial,
hostID INTEGER NOT NULL references hosts(id),
num INTEGER NOT NULL,
date integer NOT NULL,
type CHAR(4) not null,
shareID integer not null references shares(id),
- size integer not null,
- PRIMARY KEY(hostID, num, shareID)
+ size bigint not null,
+ inc_size bigint not null default -1,
+ inc_deleted boolean default false,
+ parts integer not null default 1,
+ PRIMARY KEY(id)
);
- });
- #do_index('backups_hostid,num_unique');
+ create table files (
+ ID SERIAL,
+ shareID INTEGER NOT NULL references shares(id),
+ backupNum INTEGER NOT NULL,
+ name VARCHAR(255) NOT NULL,
+ path VARCHAR(255) NOT NULL,
+ date integer NOT NULL,
+ type INTEGER NOT NULL,
+ size bigint NOT NULL,
+ primary key(id)
+ );
- $dbh->do(qq{
- create table dvds (
- ID SERIAL PRIMARY KEY,
- num INTEGER NOT NULL,
- name VARCHAR(255) NOT NULL,
- mjesto VARCHAR(255)
+ create table archive (
+ id serial,
+ dvd_nr int not null,
+ total_size bigint default -1,
+ note text,
+ username varchar(20) not null,
+ date timestamp default now(),
+ primary key(id)
+ );
+
+ create table archive_backup (
+ archive_id int not null references archive(id) on delete cascade,
+ backup_id int not null references backups(id),
+ primary key(archive_id, backup_id)
);
- });
- $dbh->do(qq{
- create table files (
- ID SERIAL PRIMARY KEY,
- shareID INTEGER NOT NULL references shares(id),
- backupNum INTEGER NOT NULL,
- name VARCHAR(255) NOT NULL,
- path VARCHAR(255) NOT NULL,
- date integer NOT NULL,
- type INTEGER NOT NULL,
- size INTEGER NOT NULL,
- dvdid INTEGER references dvds(id)
+ create table archive_burned (
+ archive_id int references archive(id),
+ date timestamp default now(),
+ part int not null default 1,
+ copy int not null default 1,
+ iso_size bigint default -1
);
+
});
- print "creating indexes:";
+ print "creating indexes: ";
foreach my $index (qw(
- hosts_name
- backups_hostID
- backups_num
- shares_hostID
- shares_name
- files_shareID
- files_path
- files_name
- files_date
- files_size
+ hosts:name
+ backups:hostID
+ backups:num
+ backups:shareID
+ shares:hostID
+ shares:name
+ files:shareID
+ files:path
+ files:name
+ files:date
+ files:size
+ archive:dvd_nr
+ archive_burned:archive_id
)) {
- print " $index";
do_index($index);
}
+
+ print " creating sequence: ";
+ foreach my $seq (qw/dvd_nr/) {
+ print "$seq ";
+ $dbh->do( qq{ CREATE SEQUENCE $seq } );
+ }
+
+
print "...\n";
$dbh->commit;
$hostID,
$backupNum,
$backup->{'endTime'},
- $backup->{'type'},
+ substr($backup->{'type'},0,4),
$shareID,
$size,
);
fmt_time($dur)
);
- hest_update($hostID, $shareID, $backupNum);
+ hest_update($hostID, $shareID, $backupNum) if ($nf + $nd > 0);
}
}
$sth->{insert_share} ||= $dbh->prepare(qq{
INSERT INTO shares
- (hostID,name,share,localpath)
- VALUES (?,?,?,?)
+ (hostID,name,share)
+ VALUES (?,?,?)
});
my $drop_down = $hostname . '/' . $share;
$drop_down =~ s#//+#/#g;
- $sth->{insert_share}->execute($hostID,$share, $drop_down ,undef);
+ $sth->{insert_share}->execute($hostID,$share, $drop_down);
return $dbh->last_insert_id(undef,undef,'shares',undef);
}
projects / BackupPC.git / blobdiff