#IBM(c) 2007 EPL license http://www.eclipse.org/legal/epl-v10.html #TODO: #MEMLEAK fix # see NodeRange.pm for notes about how to produce a memory leak # xCAT as it stands at this moment shouldn't leak anymore due to what is # described there, but that only hides from the real problem and the leak will # likely crop up if future architecture changes happen # in summary, a created Table object without benefit of db worker thread # to abstract its existance will consume a few kilobytes of memory # that never gets reused # just enough notes to remind me of the design that I think would allow for # -cache to persist so long as '_build_cache' calls concurrently stack (for NodeRange interpretation mainly) (done) # -Allow plugins to define a staleness threshold for getNodesAttribs freshness (complicated enough to postpone...) # so that actions requested by disparate managed nodes may aggregate in SQL calls # reference count managed cache lifetime, if clear_cache is called, and build_chache has been called twice, decrement the counter # if called again, decrement again and clear cache # for getNodesAttribs, we can put a parameter to request allowable staleneess # if the cachestamp is too old, build_cache is called # in this mode, 'use_cache' is temporarily set to 1, regardless of # potential other consumers (notably, NodeRange) #perl errors/and warnings are not currently wrapped. # This probably will be cleaned #up #Some known weird behaviors #creating new sqlite db files when only requested to read non-existant table, easy to fix, #class xcattable #FYI on emulated AutoCommit: #SQLite specific behavior has Table layer implementing AutoCommit. There #is a significant limitation, 'rollback' may not roll all the way back #if an intermediate transaction occured on the same table #TODO: short term, have tabutils implement it's own rollback (the only consumer) #TODO: longer term, either figure out a way to properly implement it or # document it as a limitation for SQLite configurations package xCAT::Table; use xCAT::MsgUtils; use Sys::Syslog; use Storable qw/freeze thaw/; use IO::Socket; #use Data::Dumper; use POSIX qw/WNOHANG/; use Time::HiRes qw (sleep); use Safe; my $evalcpt = new Safe; BEGIN { $::XCATROOT = $ENV{'XCATROOT'} ? $ENV{'XCATROOT'} : -d '/opt/xcat' ? '/opt/xcat' : '/usr'; } # if AIX - make sure we include perl 5.8.2 in INC path. # Needed to find perl dependencies shipped in deps tarball. if ($^O =~ /^aix/i) { use lib "/usr/opt/perl5/lib/5.8.2/aix-thread-multi"; use lib "/usr/opt/perl5/lib/5.8.2"; use lib "/usr/opt/perl5/lib/site_perl/5.8.2/aix-thread-multi"; use lib "/usr/opt/perl5/lib/site_perl/5.8.2"; } use lib "$::XCATROOT/lib/perl"; my $cachethreshold=16; #How many nodes in 'getNodesAttribs' before switching to full DB retrieval use DBI; $DBI::dbi_debug=9; # increase the debug output use strict; use Scalar::Util qw/weaken/; require xCAT::Schema; require xCAT::NodeRange; use Text::Balanced qw(extract_bracketed); require xCAT::NotifHandler; my $dbworkerpid; #The process id of the database worker my $dbworkersocket; my $dbsockpath = "/tmp/xcat/dbworker.sock.".$$; my $exitdbthread; my $dbobjsforhandle; my $intendedpid; sub dbc_call { my $self = shift; my $function = shift; my @args = @_; my $request = { function => $function, tablename => $self->{tabname}, autocommit => $self->{autocommit}, args=>\@args, }; return dbc_submit($request); } sub dbc_submit { my $request = shift; $request->{'wantarray'} = wantarray(); my $data = freeze($request); $data.= "\nENDOFFREEZEQFVyo4Cj6Q0v\n"; my $clisock; my $tries=300; while($tries and !($clisock = IO::Socket::UNIX->new(Peer => $dbsockpath, Type => SOCK_STREAM, Timeout => 120) ) ) { #print "waiting for clisock to be available\n"; sleep 0.1; $tries--; } unless ($clisock) { use Carp qw/cluck/; cluck(); } print $clisock $data; $data=""; my $lastline=""; while ($lastline ne "ENDOFFREEZEQFVyo4Cj6Q0j\n" and $lastline ne "*XCATBUGDETECTED*76e9b54341\n") { #index($lastline,"ENDOFFREEZEQFVyo4Cj6Q0j") < 0) { $lastline = <$clisock>; $data .= $lastline; } if ($lastline eq "*XCATBUGDETECTED*76e9b54341\n") { #if it was an error #in the midst of the operation, die like it used to die my $err; $data =~ /\*XCATBUGDETECTED\*:(.*):\*XCATBUGDETECTED\*/s; $err = $1; die $err; } my @returndata = @{thaw($data)}; if (wantarray) { return @returndata; } else { return $returndata[0]; } } sub shut_dbworker { $dbworkerpid = 0; #For now, just turn off usage of the db worker #This was created as the monitoring framework shutdown code otherwise seems to have a race condition #this may incur an extra db handle per service node to tolerate shutdown scenarios } sub init_dbworker { #create a db worker process #First, release all non-db-worker owned db handles (will recreate if we have to) foreach (values %{$::XCAT_DBHS}) { #@{$drh->{ChildHandles}}) { if ($_) { $_->disconnect(); } $_->{InactiveDestroy} = 1; undef $_; } $::XCAT_DBHS={}; $dbobjsforhandle={};#TODO: It's not said explicitly, but this means an #existing TABLE object is useless if going into db worker. Table objects #must be recreated after the transition. Only xcatd should have to #worry about it. This may warrant being done better, making a Table #object meaningfully survive in much the same way it survives a DB handle #migration in handle_dbc_request $dbworkerpid = fork; xCAT::NodeRange::reset_db(); #do in both processes, to be sure unless (defined $dbworkerpid) { die "Error spawining database worker"; } unless ($dbworkerpid) { $intendedpid = $$; $SIG{CHLD} = sub { while (waitpid(-1,WNOHANG) > 0) {}}; #avoid zombies from notification framework #This process is the database worker, it's job is to manage database queries to reduce required handles and to permit cross-process caching $0 = "xcatd: DB Access"; use File::Path; mkpath('/tmp/xcat/'); use IO::Socket; $SIG{TERM} = $SIG{INT} = sub { $exitdbthread=1; $SIG{ALRM} = sub { exit 0; }; alarm(10); }; unlink($dbsockpath); umask(0077); $dbworkersocket = IO::Socket::UNIX->new(Local => $dbsockpath, Type => SOCK_STREAM, Listen => 8192); unless ($dbworkersocket) { die $!; } my $currcon; my $clientset = new IO::Select; $clientset->add($dbworkersocket); #setup signal in NotifHandler so that the cache can be updated xCAT::NotifHandler::setup($$, 0); while (not $exitdbthread) { eval { my @ready_socks = $clientset->can_read; foreach $currcon (@ready_socks) { if ($currcon == $dbworkersocket) { #We have a new connection to register my $dbconn = $currcon->accept; if ($dbconn) { $clientset->add($dbconn); } } else { eval { handle_dbc_conn($currcon,$clientset); }; if ($@) { my $err=$@; xCAT::MsgUtils->message("S","xcatd: possible BUG encountered by xCAT DB worker ".$err); if ($currcon) { eval { #avoid hang by allowin client to die too print $currcon "*XCATBUGDETECTED*:$err:*XCATBUGDETECTED*\n"; print $currcon "*XCATBUGDETECTED*76e9b54341\n"; }; } } } } }; if ($@) { #this should never be reached, but leave it intact just in case my $err=$@; xCAT::MsgUtils->message("S","xcatd: possible BUG encountered by xCAT DB worker ".$err); } if ($intendedpid != $$) { #avoid redundant fork exit(0); } } close($dbworkersocket); unlink($dbsockpath); exit 0; } return $dbworkerpid; } sub handle_dbc_conn { my $client = shift; my $clientset = shift; my $data; if ($data = <$client>) { my $lastline; while ($lastline ne "ENDOFFREEZEQFVyo4Cj6Q0v\n") { #$data !~ /ENDOFFREEZEQFVyo4Cj6Q0v/) { $lastline = <$client>; $data .= $lastline; } my $request = thaw($data); my $response; my @returndata; if ($request->{'wantarray'}) { @returndata = handle_dbc_request($request); } else { @returndata = (scalar(handle_dbc_request($request))); } $response = freeze(\@returndata); $response .= "\nENDOFFREEZEQFVyo4Cj6Q0j\n"; print $client $response; } else { #Connection terminated, clean up $clientset->remove($client); close($client); } } my %opentables; #USED ONLY BY THE DB WORKER TO TRACK OPEN DATABASES sub handle_dbc_request { my $request = shift; my $functionname = $request->{function}; my $tablename = $request->{tablename}; my @args = @{$request->{args}}; my $autocommit = $request->{autocommit}; my $dbindex; foreach $dbindex (keys %{$::XCAT_DBHS}) { #Go through the current open DB handles unless ($::XCAT_DBHS->{$dbindex}) { next; } #If we have a stale dbindex entry skip it (should no longer happen with additions to init_dbworker unless ($::XCAT_DBHS->{$dbindex} and $::XCAT_DBHS->{$dbindex}->ping) { #We have a database that we were unable to reach, migrate database #handles out from under table objects my @afflictedobjs = (); #Get the list of objects whose database handle needs to be replaced if (defined $dbobjsforhandle->{$::XCAT_DBHS->{$dbindex}}) { @afflictedobjs = @{$dbobjsforhandle->{$::XCAT_DBHS->{$dbindex}}}; } else { die "DB HANDLE TRACKING CODE HAS A BUG"; } my $oldhandle = $::XCAT_DBHS->{$dbindex}; #store old handle off $::XCAT_DBHS->{$dbindex} = $::XCAT_DBHS->{$dbindex}->clone(); #replace broken db handle with nice, new, working one $dbobjsforhandle->{$::XCAT_DBHS->{$dbindex}} = $dbobjsforhandle->{$oldhandle}; #Move the map of depenednt objects to the new handle foreach (@afflictedobjs) { #migrate afflicted objects to the new DB handle $$_->{dbh} = $::XCAT_DBHS->{$dbindex}; } delete $dbobjsforhandle->{$oldhandle}; #remove the entry for the stale handle $oldhandle->disconnect(); #free resources associated with dead handle } } if ($functionname eq 'new') { unless ($opentables{$tablename}->{$autocommit}) { shift @args; #Strip repeat class stuff $opentables{$tablename}->{$autocommit} = xCAT::Table->new(@args); } if ($opentables{$tablename}->{$autocommit}) { return 1; } else { return 0; } } else { unless (defined $opentables{$tablename}->{$autocommit}) { #We are servicing a Table object that used to be #non data-worker. Create a new DB worker side Table like the one #that requests this $opentables{$tablename}->{$autocommit} = xCAT::Table->new($tablename,-create=>0,-autocommit=>$autocommit); unless ($opentables{$tablename}->{$autocommit}) { return undef; } } } if ($functionname eq 'getAllAttribs') { return $opentables{$tablename}->{$autocommit}->getAllAttribs(@args); } elsif ($functionname eq 'getAttribs') { return $opentables{$tablename}->{$autocommit}->getAttribs(@args); } elsif ($functionname eq 'getTable') { return $opentables{$tablename}->{$autocommit}->getTable(@args); } elsif ($functionname eq 'getAllNodeAttribs') { return $opentables{$tablename}->{$autocommit}->getAllNodeAttribs(@args); } elsif ($functionname eq 'getAllEntries') { return $opentables{$tablename}->{$autocommit}->getAllEntries(@args); } elsif ($functionname eq 'getAllAttribsWhere') { return $opentables{$tablename}->{$autocommit}->getAllAttribsWhere(@args); } elsif ($functionname eq 'addAttribs') { return $opentables{$tablename}->{$autocommit}->addAttribs(@args); } elsif ($functionname eq 'setAttribs') { return $opentables{$tablename}->{$autocommit}->setAttribs(@args); } elsif ($functionname eq 'setAttribsWhere') { return $opentables{$tablename}->{$autocommit}->setAttribsWhere(@args); } elsif ($functionname eq 'delEntries') { return $opentables{$tablename}->{$autocommit}->delEntries(@args); } elsif ($functionname eq 'commit') { return $opentables{$tablename}->{$autocommit}->commit(@args); } elsif ($functionname eq 'rollback') { return $opentables{$tablename}->{$autocommit}->rollback(@args); } elsif ($functionname eq 'getNodesAttribs') { return $opentables{$tablename}->{$autocommit}->getNodesAttribs(@args); } elsif ($functionname eq 'setNodesAttribs') { return $opentables{$tablename}->{$autocommit}->setNodesAttribs(@args); } elsif ($functionname eq 'getNodeAttribs') { return $opentables{$tablename}->{$autocommit}->getNodeAttribs(@args); } elsif ($functionname eq '_set_use_cache') { return $opentables{$tablename}->{$autocommit}->_set_use_cache(@args); } elsif ($functionname eq '_build_cache') { return $opentables{$tablename}->{$autocommit}->_build_cache(@args); } elsif ($functionname eq '_clear_cache') { return $opentables{$tablename}->{$autocommit}->_clear_cache(@args); } else { die "undefined function $functionname"; } } sub _set_use_cache { my $self = shift; if ($dbworkerpid) { return dbc_call($self,'_set_use_cache',@_); } $self->{_use_cache} = shift; } #-------------------------------------------------------------------------------- =head1 xCAT::Table xCAT::Table - Perl module for xCAT configuration access =head2 SYNOPSIS use xCAT::Table; my $table = xCAT::Table->new("tablename"); my $hashref=$table->getNodeAttribs("nodename","columname1","columname2"); printf $hashref->{columname1}; =head2 DESCRIPTION This module provides convenience methods that abstract the backend specific configuration to a common API. Currently implements the preferred SQLite backend, as well as a CSV backend, postgresql and MySQL, using their respective perl DBD modules. NOTES The CSV backend is really slow at scale. Room for optimization is likely, but in general DBD::CSV is slow, relative to xCAT 1.2.x. The SQLite backend, on the other hand, is significantly faster on reads than the xCAT 1.2.x way, so it is recommended. BUGS This module is not thread-safe, due to underlying DBD thread issues. Specifically in testing, SQLite DBD leaks scalars if a thread where a Table object exists spawns a child and that child exits. The recommended workaround for now is to spawn a thread to contain all Table objects if you intend to spawn threads from your main thread. As long as no thread in which the new method is called spawns child threads, it seems to work fine. AUTHOR Jarrod Johnson xCAT::Table is released under an IBM license.... =cut #-------------------------------------------------------------------------- =head2 Subroutines =cut #-------------------------------------------------------------------------- =head3 buildcreatestmt Description: Build create table statement ( see new) Arguments: Table name Table schema ( hash of column names) Returns: Table creation SQL Globals: Error: Example: my $str = buildcreatestmt($self->{tabname}, $xCAT::Schema::tabspec{$self->{tabname}}); =cut #-------------------------------------------------------------------------------- sub buildcreatestmt { my $tabn = shift; my $descr = shift; my $xcatcfg = shift; my $retv = "CREATE TABLE $tabn (\n "; my $col; my $types=$descr->{types}; my $delimitedcol; foreach $col (@{$descr->{cols}}) { my $datatype; if ($xcatcfg =~ /^DB2:/){ $datatype=get_datatype_string_db2($col, $types, $tabn,$descr); } else { $datatype=get_datatype_string($col,$xcatcfg, $types); } if ($datatype eq "TEXT") { if (isAKey(\@{$descr->{keys}}, $col)) { # keys need defined length $datatype = "VARCHAR(128) "; } } # delimit the columns of the table $delimitedcol= &delimitcol($col); $retv .= $delimitedcol . " $datatype"; # mysql change if (grep /^$col$/, @{$descr->{required}}) { # will have already put in NOT NULL, if DB2 and a key if (!($xcatcfg =~ /^DB2:/)){ # not a db2 key $retv .= " NOT NULL"; } else { # is DB2 if (!(isAKey(\@{$descr->{keys}}, $col))) { # not a key $retv .= " NOT NULL"; } } } $retv .= ",\n "; } if ($retv =~ /PRIMARY KEY/) { $retv =~ s/,\n $/\n)/; } else { $retv .= "PRIMARY KEY ("; foreach (@{$descr->{keys}}) { $delimitedcol= &delimitcol($_); $retv .= $delimitedcol . ","; } $retv =~ s/,$/)\n)/; } $retv =~ s/,$/)\n)/; # allow engine change for mysql if ($descr->{engine}) { if ($xcatcfg =~ /^mysql:/) { #for mysql $retv .= " ENGINE=$descr->{engine} "; } } #print "retv=$retv\n"; return $retv; } #-------------------------------------------------------------------------- =head3 Description: get_datatype_string ( for mysql,sqlite,postgresql) Arguments: Table column,database,types Returns: the datatype for the column being defined Globals: Error: Example: my $datatype=get_datatype_string($col,$xcatcfg, $types); =cut #-------------------------------------------------------------------------------- sub get_datatype_string { my $col=shift; #column name my $xcatcfg=shift; #db config string my $types=shift; #hash pointer my $ret; if (($types) && ($types->{$col})) { if ($types->{$col} =~ /INTEGER AUTO_INCREMENT/) { if ($xcatcfg =~ /^SQLite:/) { $ret = "INTEGER PRIMARY KEY AUTOINCREMENT"; } elsif ($xcatcfg =~ /^Pg:/) { $ret = "SERIAL"; } elsif ($xcatcfg =~ /^mysql:/){ $ret = "INTEGER AUTO_INCREMENT"; } else { } } else { $ret = $types->{$col}; } } else { $ret = "TEXT"; } return $ret; } #-------------------------------------------------------------------------- =head3 Description: get_datatype_string_db2 ( for DB2) Arguments: Table column,database,types,tablename,table schema Returns: the datatype for the column being defined Globals: Error: Example: my $datatype=get_datatype_string_db2($col, $types,$tablename,$descr); =cut #-------------------------------------------------------------------------------- sub get_datatype_string_db2 { my $col=shift; #column name my $types=shift; #types field (eventlog) my $tablename=shift; # tablename my $descr=shift; # table schema my $ret = "varchar(512)"; # default for most attributes if (($types) && ($types->{$col})) { if ($types->{$col} =~ /INTEGER AUTO_INCREMENT/) { $ret = "INTEGER GENERATED ALWAYS AS IDENTITY"; } else { # if the column is a key if (isAKey(\@{$descr->{keys}}, $col)) { $ret = $types->{$col}; $ret .= " NOT NULL "; } else { $ret = $types->{$col}; if ($ret eq "TEXT") { # text not in db2 $ret = "VARCHAR(512)"; } } } } else { # type not specifically define if (isAKey(\@{$descr->{keys}}, $col)) { $ret = "VARCHAR(128) NOT NULL "; } } if ($col eq "disable") { $ret = "varchar(8)"; } if ($col eq "rawdata") { # from eventlog table $ret = "varchar(4098)"; } return $ret; } #-------------------------------------------------------------------------- =head3 Description: get_xcatcfg Arguments: none Returns: the database name from /etc/xcat/cfgloc or sqlite Globals: Error: Example: my $xcatcfg =get_xcatcfg(); =cut #-------------------------------------------------------------------------------- sub get_xcatcfg { my $xcatcfg = (defined $ENV{'XCATCFG'} ? $ENV{'XCATCFG'} : ''); unless ($xcatcfg) { if (-r "/etc/xcat/cfgloc") { my $cfgl; open($cfgl,"<","/etc/xcat/cfgloc"); $xcatcfg = <$cfgl>; close($cfgl); chomp($xcatcfg); $ENV{'XCATCFG'}=$xcatcfg; #Store it in env to avoid many file reads } } if ($xcatcfg =~ /^$/) { if (-d "/opt/xcat/cfg") { $xcatcfg = "SQLite:/opt/xcat/cfg"; } else { if (-d "/etc/xcat") { $xcatcfg = "SQLite:/etc/xcat"; } } } ($xcatcfg =~ /^$/) && die "Can't locate xCAT configuration"; unless ($xcatcfg =~ /:/) { $xcatcfg = "SQLite:" . $xcatcfg; } return $xcatcfg; } #-------------------------------------------------------------------------- =head3 new Description: Constructor: Connects to or Creates Database Table Arguments: Table name 0 = Connect to table 1 = Create table Returns: Hash: Database Handle, Statement Handle, nodelist Globals: Error: Example: $nodelisttab = xCAT::Table->new("nodelist"); Comments: none =cut #-------------------------------------------------------------------------------- sub new { #Constructor takes table name as argument #Also takes a true/false value, or assumes 0. If something true is passed, create table #is requested my @args = @_; my $self = {}; my $proto = shift; $self->{tabname} = shift; unless (defined($xCAT::Schema::tabspec{$self->{tabname}})) { return undef; } $self->{schema} = $xCAT::Schema::tabspec{$self->{tabname}}; $self->{colnames} = \@{$self->{schema}->{cols}}; $self->{descriptions} = \%{$self->{schema}->{descriptions}}; my %otherargs = @_; my $create = 1; if (exists($otherargs{'-create'}) && ($otherargs{'-create'}==0)) {$create = 0;} $self->{autocommit} = $otherargs{'-autocommit'}; unless (defined($self->{autocommit})) { $self->{autocommit} = 1; } $self->{realautocommit} = $self->{autocommit}; #Assume requester autocommit behavior maps directly to DBI layer autocommit my $class = ref($proto) || $proto; if ($dbworkerpid) { my $request = { function => "new", tablename => $self->{tabname}, autocommit => $self->{autocommit}, args=>\@args, }; unless (dbc_submit($request)) { return undef; } } else { #direct db access mode $self->{dbuser}=""; $self->{dbpass}=""; my $xcatcfg =get_xcatcfg(); my $xcatdb2schema; if ($xcatcfg =~ /^DB2:/) { # for DB2 , get schema name my @parts = split ( '\|', $xcatcfg); $xcatdb2schema = $parts[1]; $xcatdb2schema =~ tr/a-z/A-Z/; # convert to upper } if ($xcatcfg =~ /^SQLite:/) { $self->{backend_type} = 'sqlite'; $self->{realautocommit} = 1; #Regardless of autocommit semantics, only electively do autocommit due to SQLite locking difficulties #for SQLite, we cannot open the same db twice without deadlock risk, so we cannot have both autocommit on and off via #different handles, so we pick one #previously, Table.pm tried to imitate autocommit, but evidently was problematic, so now SQLite is just almost always #autocommit, turned off very selectively #so realautocommit is a hint to say that the handle needs to be set back to autocommit as soon as possible my @path = split(':', $xcatcfg, 2); unless (-e $path[1] . "/" . $self->{tabname} . ".sqlite" || $create) { return undef; } $self->{connstring} = "dbi:" . $xcatcfg . "/" . $self->{tabname} . ".sqlite"; } elsif ($xcatcfg =~ /^CSV:/) { $self->{backend_type} = 'csv'; $xcatcfg =~ m/^.*?:(.*)$/; my $path = $1; $self->{connstring} = "dbi:CSV:f_dir=" . $path; } else #Generic DBI { ($self->{connstring},$self->{dbuser},$self->{dbpass}) = split(/\|/,$xcatcfg); $self->{connstring} =~ s/^dbi://; $self->{connstring} =~ s/^/dbi:/; #return undef; } if ($xcatcfg =~ /^DB2:/) { # for DB2 ,export the INSTANCE name $ENV{'DB2INSTANCE'} = $self->{dbuser}; } my $oldumask= umask 0077; unless ($::XCAT_DBHS->{$self->{connstring},$self->{dbuser},$self->{dbpass},$self->{realautocommit}}) { #= $self->{tabname}; $::XCAT_DBHS->{$self->{connstring},$self->{dbuser},$self->{dbpass},$self->{realautocommit}} = DBI->connect($self->{connstring}, $self->{dbuser}, $self->{dbpass}, {AutoCommit => $self->{realautocommit}}); } umask $oldumask; $self->{dbh} = $::XCAT_DBHS->{$self->{connstring},$self->{dbuser},$self->{dbpass},$self->{realautocommit}}; #Store the Table object reference as afflicted by changes to the DBH #This for now is ok, as either we aren't in DB worker mode, in which case this structure would be short lived... #or we are in db worker mode, in which case Table objects live indefinitely #TODO: be able to reap these objects sanely, just in case push @{$dbobjsforhandle->{$::XCAT_DBHS->{$self->{connstring},$self->{dbuser},$self->{dbpass},$self->{realautocommit}}}},\$self; #DBI->connect($self->{connstring}, $self->{dbuser}, $self->{dbpass}, {AutoCommit => $autocommit}); if ($xcatcfg =~ /^SQLite:/) { my $dbexistq = "SELECT name from sqlite_master WHERE type='table' and name = ?"; my $sth = $self->{dbh}->prepare($dbexistq); $sth->execute($self->{tabname}); my $result = $sth->fetchrow(); $sth->finish; unless (defined $result) { if ($create) { my $str = buildcreatestmt($self->{tabname}, $xCAT::Schema::tabspec{$self->{tabname}}, $xcatcfg); $self->{dbh}->do($str); if (!$self->{dbh}->{AutoCommit}) { $self->{dbh}->commit; } } else { return undef; } } } elsif ($xcatcfg =~ /^CSV:/) { $self->{dbh}->{'csv_tables'}->{$self->{tabname}} = {'file' => $self->{tabname} . ".csv"}; $xcatcfg =~ m/^.*?:(.*)$/; my $path = $1; if (!-e $path . "/" . $self->{tabname} . ".csv") { unless ($create) { return undef; } my $str = buildcreatestmt($self->{tabname}, $xCAT::Schema::tabspec{$self->{tabname}}, $xcatcfg); $self->{dbh}->do($str); } } else { #generic DBI if (!$self->{dbh}) { xCAT::MsgUtils->message("S", "Could not connect to the database. Database handle not defined."); return undef; } my $tbexistq; my $dbtablename=$self->{tabname}; my $found = 0; if ($xcatcfg =~ /^DB2:/) { # for DB2 $dbtablename =~ tr/a-z/A-Z/; # convert to upper $tbexistq = $self->{dbh}->table_info(undef,$xcatdb2schema,$dbtablename,'TABLE'); } else { $tbexistq = $self->{dbh}->table_info('','',$self->{tabname},'TABLE'); } while (my $data = $tbexistq->fetchrow_hashref) { if ($data->{'TABLE_NAME'} =~ /^\"?$dbtablename\"?\z/) { if ($xcatcfg =~ /^DB2:/) { # for DB2 if ($data->{'TABLE_SCHEM'} =~ /^\"?$xcatdb2schema\"?\z/) { # must check schema also with db2 $found = 1; last; } } else { # not db2 $found = 1; last; } } } unless ($found) { unless ($create) { return undef; } my $str = buildcreatestmt($self->{tabname}, $xCAT::Schema::tabspec{$self->{tabname}}, $xcatcfg); $self->{dbh}->do($str); if (!$self->{dbh}->{AutoCommit}) { $self->{dbh}->commit; # commit the create } } } # end Generic DBI } #END DB ACCESS SPECIFIC SECTION if ($self->{tabname} eq 'nodelist') { weaken($self->{nodelist} = $self); } else { $self->{nodelist} = xCAT::Table->new('nodelist',-create=>1); } bless($self, $class); return $self; } #-------------------------------------------------------------------------- =head3 updateschema Description: Alters table info in the database based on Schema changes Handles adding attributes Handles removing attributes but does not really remove them from the database. Handles adding keys Arguments: Hash containing Database and Table Handle and schema Returns: None Globals: Error: Example: my $nodelisttab=xCAT::Table->new('nodelist'); $nodelisttab->updateschema(); $nodelisttab->close(); Comments: none =cut #-------------------------------------------------------------------------------- sub updateschema { #This determines alter table statements required.. my $self = shift; my $descr=$xCAT::Schema::tabspec{$self->{tabname}}; my $tn=$self->{tabname}; my $xcatdb2schema; my $xcatcfg=get_xcatcfg(); my $rc=0; my $msg; if ($xcatcfg =~ /^DB2:/) { # for DB2 , get schema name my @parts = split ( '\|', $xcatcfg); $xcatdb2schema = $parts[1]; $xcatdb2schema =~ tr/a-z/A-Z/; # convert to upper } my @columns; my %dbkeys; if ($self->{backend_type} eq 'sqlite') { my $dbexistq = "PRAGMA table_info('$tn')"; my $sth = $self->{dbh}->prepare($dbexistq); $sth->execute; my $tn=$self->{tabname}; while ( my $col_info = $sth->fetchrow_hashref ) { #print Dumper($col_info); my $tmp_col=$col_info->{name}; $tmp_col =~ s/"//g; push @columns, $tmp_col; if ($col_info->{pk}) { $dbkeys{$tmp_col}=1; } } $sth->finish; } else { #Attempt generic dbi.. #my $sth = $self->{dbh}->column_info('','',$self->{tabname},''); my $sth; if ($xcatcfg =~ /^DB2:/) { # for DB2 my $db2table = $self->{tabname}; $db2table =~ tr/a-z/A-Z/; # convert to upper for db2 $sth = $self->{dbh}->column_info(undef,$xcatdb2schema,$db2table,'%'); } else { $sth = $self->{dbh}->column_info(undef,undef,$self->{tabname},'%'); } while (my $cd = $sth->fetchrow_hashref) { #print Dumper($cd); push @columns,$cd->{'COLUMN_NAME'}; #special code for old version of perl-DBD-mysql if (defined($cd->{mysql_is_pri_key}) && ($cd->{mysql_is_pri_key}==1)) { my $tmp_col=$cd->{'COLUMN_NAME'}; $tmp_col =~ s/"//g; $dbkeys{$tmp_col}=1; } } foreach (@columns) { #Column names may end up quoted by database engin s/"//g; } #get primary keys if ($xcatcfg =~ /^DB2:/) { # for DB2 my $db2table = $self->{tabname}; $db2table =~ tr/a-z/A-Z/; # convert to upper for db2 $sth = $self->{dbh}->primary_key_info(undef,$xcatdb2schema,$db2table); } else { $sth = $self->{dbh}->primary_key_info(undef,undef,$self->{tabname}); } if ($sth) { my $data = $sth->fetchall_arrayref; #print "data=". Dumper($data); foreach my $cd (@$data) { my $tmp_col=$cd->[3]; $tmp_col =~ s/"//g; $dbkeys{$tmp_col}=1; } } } #Now @columns reflects the *actual* columns in the database my $dcol; my $types=$descr->{types}; foreach $dcol (@{$self->{colnames}}) { unless (grep /^$dcol$/, @columns) { #TODO: log/notify of schema upgrade? my $datatype; if ($xcatcfg =~ /^DB2:/){ $datatype=get_datatype_string_db2($dcol, $types, $tn,$descr); } else{ $datatype=get_datatype_string($dcol, $xcatcfg, $types); } if ($datatype eq "TEXT") { # keys cannot be TEXT if (isAKey(\@{$descr->{keys}}, $dcol)) { # keys $datatype = "VARCHAR(128) "; } } if (grep /^$dcol$/, @{$descr->{required}}) { $datatype .= " NOT NULL"; } # delimit the columns of the table my $tmpcol= &delimitcol($dcol); my $tablespace; my $stmt = "ALTER TABLE " . $self->{tabname} . " ADD $tmpcol $datatype"; $self->{dbh}->do($stmt); $msg="updateschema: Running $stmt"; xCAT::MsgUtils->message("S", $msg); if ($self->{dbh}->errstr) { xCAT::MsgUtils->message("S", "Error adding columns for table " . $self->{tabname} .":" . $self->{dbh}->errstr); if ($xcatcfg =~ /^DB2:/){ # see if table space error my $error = $self->{dbh}->errstr; if ($error =~ /54010/) { # move the table to the next tablespace if ($error =~ /USERSPACE1/) { $tablespace="XCATTBS16K"; } else { $tablespace="XCATTBS32K"; } my $tablename=$self->{tabname}; $tablename=~ tr/a-z/A-Z/; # convert to upper $msg="Moving table $self->{tabname} to $tablespace"; xCAT::MsgUtils->message("S", $msg); my $stmt2="Call sysproc.admin_move_table('XCATDB',\'$tablename\',\'$tablespace\',\'$tablespace\',\'$tablespace\','','','','','','MOVE')"; $self->{dbh}->do($stmt2); if ($self->{dbh}->errstr) { xCAT::MsgUtils->message("S", "Error on tablespace move for table " . $self->{tabname} .":" . $self->{dbh}->errstr); } else { # tablespace move try column add again if (!$self->{dbh}->{AutoCommit}) { # commit tbsp move $self->{dbh}->commit; } $self->{dbh}->do($stmt); } } # if tablespace error } # if db2 } # error on add column if (!$self->{dbh}->{AutoCommit}) { # commit add column $self->{dbh}->commit; } } } #for existing columns that are new keys now # note new keys can only be created from existing columns # since keys cannot be null, the copy from the backup table will fail if # the old value was null. my @new_dbkeys=@{$descr->{keys}}; my @old_dbkeys=keys %dbkeys; #print "new_dbkeys=@new_dbkeys; old_dbkeys=@old_dbkeys; columns=@columns\n"; my $change_keys=0; #Add the new key columns to the table foreach my $dbkey (@new_dbkeys) { if (! exists($dbkeys{$dbkey})) { $change_keys=1; # Call tabdump plugin to create a CSV file # can be used in case the restore fails # put in /tmp/ my $backuptable="/tmp/$tn.csv.$$"; my $cmd="$::XCATROOT/sbin/tabdump $tn > $backuptable"; `$cmd`; $msg="updateschema: Backing up table before key change, $cmd"; xCAT::MsgUtils->message("S", $msg); #for my sql, we do not have to recreate table, but we have to make sure the type is correct, my $datatype; if ($xcatcfg =~ /^mysql:/) { $datatype=get_datatype_string($dbkey, $xcatcfg, $types); } else { # db2 $datatype=get_datatype_string_db2($dbkey, $types, $tn,$descr); } if ($datatype eq "TEXT") { if (isAKey(\@{$descr->{keys}}, $dbkey)) { # keys need defined length $datatype = "VARCHAR(128) "; } } # delimit the columns my $tmpkey= &delimitcol($dbkey); if (($xcatcfg =~ /^DB2:/) || ($xcatcfg =~ /^Pg:/)) { # get rid of NOT NULL, cannot modify with NOT NULL my ($tmptype,$nullvalue)= split('NOT NULL',$datatype ); $datatype=$tmptype; } my $stmt; if ($xcatcfg =~ /^DB2:/){ $stmt = "ALTER TABLE " . $self->{tabname} . " ALTER COLUMN $tmpkey SET DATA TYPE $datatype"; } else { $stmt = "ALTER TABLE " . $self->{tabname} . " MODIFY COLUMN $tmpkey $datatype"; } $msg="updateschema: Running $stmt"; xCAT::MsgUtils->message("S", $msg); #print "stmt=$stmt\n"; $self->{dbh}->do($stmt); if ($self->{dbh}->errstr) { xCAT::MsgUtils->message("S", "Error changing the keys for table " . $self->{tabname} .":" . $self->{dbh}->errstr); } } } #finally add the new keys if ($change_keys) { if ($xcatcfg =~ /^mysql:/) { #for mysql, just alter the table my $tmp=join(',',@new_dbkeys); my $stmt = "ALTER TABLE " . $self->{tabname} . " DROP PRIMARY KEY, ADD PRIMARY KEY ($tmp)"; #print "stmt=$stmt\n"; $self->{dbh}->do($stmt); $msg="updateschema: Running $stmt"; xCAT::MsgUtils->message("S", $msg); if ($self->{dbh}->errstr) { xCAT::MsgUtils->message("S", "Error changing the keys for table " . $self->{tabname} .":" . $self->{dbh}->errstr); } } else { #for the rest, recreate the table #print "need to change keys\n"; my $btn=$tn . "_xcatbackup"; #remove the backup table just in case; # gets error if not there #my $str="DROP TABLE $btn"; #$self->{dbh}->do($str); #rename the table name to name_xcatbackup my $str; if ($xcatcfg =~ /^DB2:/){ $str = "RENAME TABLE $tn TO $btn"; } else { $str = "ALTER TABLE $tn RENAME TO $btn"; } $self->{dbh}->do($str); if (!$self->{dbh}->{AutoCommit}) { $self->{dbh}->commit; } $msg="updateschema: Running $str"; xCAT::MsgUtils->message("S", $msg); if ($self->{dbh}->errstr) { xCAT::MsgUtils->message("S", "Error renaming the table from $tn to $btn:" . $self->{dbh}->errstr); } if (!$self->{dbh}->{AutoCommit}) { $self->{dbh}->commit; } #create the table again $str = buildcreatestmt($tn, $descr, $xcatcfg); $self->{dbh}->do($str); if ($self->{dbh}->errstr) { xCAT::MsgUtils->message("S", "Error recreating table $tn:" . $self->{dbh}->errstr); } if (!$self->{dbh}->{AutoCommit}) { $self->{dbh}->commit; } #copy the data from backup to the table $str = "INSERT INTO $tn SELECT * FROM $btn"; $self->{dbh}->do($str); $msg="updateschema: Running $str"; xCAT::MsgUtils->message("S", $msg); if ($self->{dbh}->errstr) { xCAT::MsgUtils->message("S", "Error copying data from table $btn to $tn:" . $self->{dbh}->errstr); } else { #drop the backup table $str = "DROP TABLE $btn"; $self->{dbh}->do($str); } if (!$self->{dbh}->{AutoCommit}) { $self->{dbh}->commit; } } } return $rc; } #-------------------------------------------------------------------------- =head3 setNodeAttribs Description: Set attributes values on the node input to the routine Arguments: Hash: Database Handle, Statement Handle, nodelist Node name Attribute hash Returns: Globals: Error: Example: my $mactab = xCAT::Table->new('mac',-create=>1); $mactab->setNodeAttribs($node,{mac=>$mac}); $mactab->close(); Comments: none =cut #-------------------------------------------------------------------------------- sub setNodeAttribs { my $self = shift; my $node = shift; my $nodekey = "node"; if (defined $xCAT::Schema::tabspec{$self->{tabname}}->{nodecol}) { $nodekey = $xCAT::Schema::tabspec{$self->{tabname}}->{nodecol} }; return $self->setAttribs({$nodekey => $node}, @_); } #-------------------------------------------------------------------------- =head3 addNodeAttribs (not supported) Description: Add new attributes input to the routine to the nodes Arguments: Hash of new attributes Returns: Globals: Error: Example: Comments: none =cut #-------------------------------------------------------------------------------- sub addNodeAttribs { my $self = shift; xCAT::MsgUtils->message("S","addNodeAttribs is not supported"); die "addNodeAttribs is not supported"; return $self->addAttribs('node', @_); } #-------------------------------------------------------------------------- =head3 addAttribs (not supported) Description: add new attributes Arguments: Hash: Database Handle, Statement Handle, nodelist Key name Key value Hash reference of column-value pairs to set Returns: Globals: Error: Example: Comments: none =cut #-------------------------------------------------------------------------------- sub addAttribs { my $self = shift; xCAT::MsgUtils->message("S","addAttribs is not supported"); die "addAttribs is not supported"; if ($dbworkerpid) { return dbc_call($self,'addAttribs',@_); } if (not $self->{intransaction} and not $self->{autocommit} and $self->{realautocommit}) { #I have realized that this needs some comment work #so if we are not in a transaction, *but* the current caller context expects autocommit to be off (i.e. fast performance and rollback #however, the DBI layer is set to stick to autocommit on as much as possible (pretty much SQLite) because the single #handle is shared, we disable autocommit on that handle until commit or rollback is called #yes, that means some table calls coming in with expectation of autocommit during this hopefully short interval #could get rolled back along with this transaction, but it's unlikely and moving to a more robust DB solves it #so it is intentional that autocommit is left off because it is expected that a commit will come along one day and fix it right up #TODO: if caller crashes after inducing a 'begin transaction' without commit or rollback, this could be problematic. #calling commit on all table handles if a client goes away uncleanly may be a good enough solution. $self->{intransaction}=1; $self->{dbh}->{AutoCommit}=0; } my $key = shift; my $keyval = shift; my $elems = shift; my $cols = ""; my @bind = (); @bind = ($keyval); $cols = "$key,"; for my $col (keys %$elems) { $cols = $cols . $col . ","; if (ref($$elems{$col})) { push @bind, ${$elems}{$col}->[0]; } else { push @bind, $$elems{$col}; } } chop($cols); my $qstring = 'INSERT INTO ' . $self->{tabname} . " ($cols) VALUES ("; for (@bind) { $qstring = $qstring . "?,"; } $qstring =~ s/,$/)/; my $sth = $self->{dbh}->prepare($qstring); $sth->execute(@bind); #$self->{dbh}->commit; #notify the interested parties my $notif = xCAT::NotifHandler->needToNotify($self->{tabname}, 'a'); if ($notif == 1) { my %new_notif_data; $new_notif_data{$key} = $keyval; foreach (keys %$elems) { $new_notif_data{$_} = $$elems{$_}; } xCAT::NotifHandler->notify("a", $self->{tabname}, [0], \%new_notif_data); } $sth->finish(); } #-------------------------------------------------------------------------- =head3 rollback Description: rollback changes Arguments: Database Handle Returns: none Globals: Error: Example: my $tab = xCAT::Table->new($table,-create =>1,-autocommit=>0); $tab->rollback(); Comments: none =cut #-------------------------------------------------------------------------------- sub rollback { my $self = shift; if ($dbworkerpid) { return dbc_call($self,'rollback',@_); } $self->{dbh}->rollback; if ($self->{intransaction} and not $self->{autocommit} and $self->{realautocommit}) { #on rollback, if sharing a DB handle for autocommit and non-autocommit, put the handle back to autocommit $self->{intransaction}=0; $self->{dbh}->{AutoCommit}=1; } } #-------------------------------------------------------------------------- =head3 commit Description: Commit changes Arguments: Database Handle Returns: none Globals: Error: Example: my $tab = xCAT::Table->new($table,-create =>1,-autocommit=>0); $tab->commit(); Comments: none =cut #-------------------------------------------------------------------------------- sub commit { my $self = shift; if ($dbworkerpid) { return dbc_call($self,'commit',@_); } unless ($self->{dbh}->{AutoCommit}) { #caller can now hammer commit function with impunity, even when it makes no sense $self->{dbh}->commit; } if ($self->{intransaction} and not $self->{autocommit} and $self->{realautocommit}) { #if realautocommit indicates shared DBH between manual and auto commit, put the handle back to autocommit if a transaction #is committed (aka ended) $self->{intransaction}=0; $self->{dbh}->{AutoCommit}=1; } } #-------------------------------------------------------------------------- =head3 setAttribs Description: Arguments: Key name Key value Hash reference of column-value pairs to set Returns: None Globals: Error: Example: my $tab = xCAT::Table->new( 'ppc', -create=>1, -autocommit=>0 ); $keyhash{'node'} = $name; $updates{'type'} = lc($type); $updates{'id'} = $lparid; $updates{'hcp'} = $server; $updates{'profile'} = $prof; $updates{'frame'} = $frame; $updates{'mtms'} = "$model*$serial"; $tab->setAttribs( \%keyhash,\%updates ); $tab->commit; Comments: none =cut #-------------------------------------------------------------------------------- sub setAttribs { #Takes three arguments: #-Key name #-Key value #-Hash reference of column-value pairs to set my $xcatcfg =get_xcatcfg(); my $self = shift; if ($dbworkerpid) { return dbc_call($self,'setAttribs',@_); } my $pKeypairs=shift; my %keypairs = (); if ($pKeypairs != undef) { %keypairs = %{$pKeypairs}; } #my $key = shift; #my $keyval=shift; my $elems = shift; my $cols = ""; my @bind = (); my $action; my @notif_data; my $qstring; $qstring = "SELECT * FROM " . $self->{tabname} . " WHERE "; my @qargs = (); my $query; my $data; if (not $self->{intransaction} and not $self->{autocommit} and $self->{realautocommit}) { #search this code for the other if statement just like it for an explanation of why I do this $self->{intransaction}=1; $self->{dbh}->{AutoCommit}=0; } if (($pKeypairs != undef) && (keys(%keypairs)>0)) { foreach (keys %keypairs) { # delimit the columns of the table my $delimitedcol = &delimitcol($_); if ($xcatcfg =~ /^DB2:/) { # for DB2 $qstring .= $delimitedcol . " LIKE ? AND "; } else { $qstring .= $delimitedcol . " = ? AND "; } push @qargs, $keypairs{$_}; } $qstring =~ s/ AND \z//; #print "this is qstring1: $qstring\n"; $query = $self->{dbh}->prepare($qstring); $query->execute(@qargs); #get the first row $data = $query->fetchrow_arrayref(); if (defined $data) { $action = "u"; } else { $action = "a"; } } else { $action = "a";} #prepare the notification data my $notif = xCAT::NotifHandler->needToNotify($self->{tabname}, $action); if ($notif == 1) { if ($action eq "u") { #put the column names at the very front push(@notif_data, $query->{NAME}); #copy the data out because fetchall_arrayref overrides the data. my @first_row = @$data; push(@notif_data, \@first_row); #get the rest of the rows my $temp_data = $query->fetchall_arrayref(); foreach (@$temp_data) { push(@notif_data, $_); } } } if ($query) { $query->finish(); } if ($action eq "u") { #update the rows $action = "u"; for my $col (keys %$elems) { # delimit the columns of the table my $delimitedcol = &delimitcol($col); $cols = $cols . $delimitedcol . " = ?,"; push @bind, (($$elems{$col} =~ /NULL/) ? undef: $$elems{$col}); } chop($cols); my $cmd ; $cmd = "UPDATE " . $self->{tabname} . " set $cols where "; foreach (keys %keypairs) { if (ref($keypairs{$_})) { # delimit the columns my $delimitedcol = &delimitcol($_); $cmd .= $delimitedcol . " = '" . $keypairs{$_}->[0] . "' AND "; } else { my $delimitedcol = &delimitcol($_); $cmd .= $delimitedcol . " = '" . $keypairs{$_} . "' AND "; } } $cmd =~ s/ AND \z//; my $sth = $self->{dbh}->prepare($cmd); unless ($sth) { return (undef,"Error attempting requested DB operation"); } my $err = $sth->execute(@bind); if (not defined($err)) { return (undef, $sth->errstr); } $sth->finish; } else { #insert the rows $action = "a"; @bind = (); $cols = ""; my %newpairs; #first, merge the two structures to a single hash foreach (keys %keypairs) { $newpairs{$_} = $keypairs{$_}; } my $needinsert=0; for my $col (keys %$elems) { $newpairs{$col} = $$elems{$col}; if (defined $newpairs{$col} and not $newpairs{$col} eq "") { $needinsert=1; } } unless ($needinsert) { #Don't bother inserting truly blank lines return; } foreach (keys %newpairs) { my $delimitedcol = &delimitcol($_); $cols .= $delimitedcol . ","; push @bind, $newpairs{$_}; } chop($cols); my $qstring = 'INSERT INTO ' . $self->{tabname} . " ($cols) VALUES ("; for (@bind) { $qstring = $qstring . "?,"; } $qstring =~ s/,$/)/; my $sth = $self->{dbh}->prepare($qstring); my $err = $sth->execute(@bind); if (not defined($err)) { return (undef, $sth->errstr); } $sth->finish; } $self->_refresh_cache(); #cache is invalid, refresh #notify the interested parties if ($notif == 1) { #create new data ref my %new_notif_data = %keypairs; foreach (keys %$elems) { $new_notif_data{$_} = $$elems{$_}; } xCAT::NotifHandler->notify($action, $self->{tabname}, \@notif_data, \%new_notif_data); } return 0; } #-------------------------------------------------------------------------- =head3 setAttribsWhere Description: This function sets the attributes for the rows selected by the where clause. Warning, because we support mulitiple databases (SQLite,MySQL and DB2) that require different syntax. Any code using this routine, must call the Utils->getDBName routine and code the where clause that is appropriate for each supported database. Arguments: Where clause. Note: if the Where clause contains any reserved keywords like keys from the site table, then you will have to code them in backticks for MySQL and not in backticks for Postgresql. Hash reference of column-value pairs to set Returns: None Globals: Error: Example: my $tab = xCAT::Table->new( 'ppc', -create=>1, -autocommit=>1 ); $updates{'type'} = lc($type); $updates{'id'} = $lparid; $updates{'hcp'} = $server; $updates{'profile'} = $prof; $updates{'frame'} = $frame; $updates{'mtms'} = "$model*$serial"; $tab->setAttribsWhere( "node in ('node1', 'node2', 'node3')", \%updates ); Comments: none =cut #-------------------------------------------------------------------------------- sub setAttribsWhere { #Takes three arguments: #-Where clause #-Hash reference of column-value pairs to set my $self = shift; if ($dbworkerpid) { return dbc_call($self,'setAttribsWhere',@_); } my $where_clause = shift; my $elems = shift; my $cols = ""; my @bind = (); my $action; my @notif_data; if (not $self->{intransaction} and not $self->{autocommit} and $self->{realautocommit}) { #search this code for the other if statement just like it for an explanation of why I do this $self->{intransaction}=1; $self->{dbh}->{AutoCommit}=0; } my $qstring = "SELECT * FROM " . $self->{tabname} . " WHERE " . $where_clause; my @qargs = (); my $query = $self->{dbh}->prepare($qstring); $query->execute(@qargs); #get the first row my $data = $query->fetchrow_arrayref(); if (defined $data){ $action = "u";} else { return (0, "no rows selected."); } #prepare the notification data my $notif = xCAT::NotifHandler->needToNotify($self->{tabname}, $action); if ($notif == 1) { #put the column names at the very front push(@notif_data, $query->{NAME}); #copy the data out because fetchall_arrayref overrides the data. my @first_row = @$data; push(@notif_data, \@first_row); #get the rest of the rows my $temp_data = $query->fetchall_arrayref(); foreach (@$temp_data) { push(@notif_data, $_); } } $query->finish(); #update the rows for my $col (keys %$elems) { # delimit the columns of the table my $delimitedcol = &delimitcol($col); $cols = $cols . $delimitedcol . " = ?,"; push @bind, (($$elems{$col} =~ /NULL/) ? undef: $$elems{$col}); } chop($cols); my $cmd = "UPDATE " . $self->{tabname} . " set $cols where " . $where_clause; my $sth = $self->{dbh}->prepare($cmd); my $err = $sth->execute(@bind); if (not defined($err)) { return (undef, $sth->errstr); } #notify the interested parties if ($notif == 1) { #create new data ref my %new_notif_data = (); foreach (keys %$elems) { $new_notif_data{$_} = $$elems{$_}; } xCAT::NotifHandler->notify($action, $self->{tabname}, \@notif_data, \%new_notif_data); } $sth->finish; return 0; } #-------------------------------------------------------------------------- =head3 setNodesAttribs Description: Unconditionally assigns the requested values to tables for a list of nodes Arguments: 'self' (implicit in OO style call) A reference to a two-level hash similar to: { 'n1' => { comments => 'foo', data => 'foo2' }, 'n2' => { comments => 'bar', data => 'bar2' } } Alternative arguments (same set of data to be applied to multiple nodes): 'self' Reference to a list of nodes (no noderanges, just nodes) A hash of attributes to set, like in 'setNodeAttribs' Returns: =cut #-------------------------------------------------------------------------- sub setNodesAttribs { #This is currently a stub to be filled out with at scale enhancements. It will be a touch more complex than getNodesAttribs, due to the notification #The three steps should be: #-Query table and divide nodes into list to update and list to insert #-Update intelligently with respect to scale #-Insert intelligently with respect to scale (prepare one statement, execute many times, other syntaxes not universal) #Intelligently in this case means folding them to some degree. Update where clauses will be longer, but must be capped to avoid exceeding SQL statement length restrictions on some DBs. Restricting even all the way down to 256 could provide better than an order of magnitude better performance though my $self = shift; if ($dbworkerpid) { return dbc_call($self,'setNodesAttribs',@_); } my $nodelist = shift; my $keyset = shift; my %cols = (); my @orderedcols=(); my $oldac = $self->{dbh}->{AutoCommit}; #save autocommit state $self->{dbh}->{AutoCommit}=0; #turn off autocommit for performance my $hashrec; my $colsmatch=1; if (ref $nodelist eq 'HASH') { # argument of the form { n001 => { groups => 'something' }, n002 => { groups => 'other' } } $hashrec = $nodelist; my @nodes = keys %$nodelist; $nodelist = \@nodes; my $firstpass=1; foreach my $node (keys %$hashrec) { #Determine whether the passed structure is trying to set the same columns #for every node to determine if the short path can work or not if ($firstpass) { $firstpass=0; foreach (keys %{$hashrec->{$node}}) { $cols{$_}=1; } } else { foreach (keys %{$hashrec->{$node}}) { #make sure all columns in this entry are in the first unless (defined $cols{$_}) { $colsmatch=0; last; } } foreach my $col (keys %cols) { #make sure this entry does not lack any columns from the first unless (defined $hashrec->{$node}->{$col}) { $colsmatch=0; last; } } } } } else { #the legacy calling style with a list reference and a single hash reference of col=>va/ue pairs $hashrec = {}; foreach (@$nodelist) { $hashrec->{$_}=$keyset; } foreach (keys %$keyset) { $cols{$_}=1; } } #revert to the old way if notification is required or asymettric setNodesAttribs was requested with different columns #for different nodes if (not $colsmatch or xCAT::NotifHandler->needToNotify($self->{tabname}, 'u') or xCAT::NotifHandler->needToNotify($self->{tabname}, 'a')) { #TODO: enhance performance of this case too, for now just call the notification-capable code per node foreach (keys %$hashrec) { $self->setNodeAttribs($_,$hashrec->{$_}); } $self->{dbh}->commit; #commit pending transactions $self->{dbh}->{AutoCommit}=$oldac;#restore autocommit semantics return; } #this code currently is notification incapable. It enhances scaled setting by: #-turning off autocommit if on (done for above code too, but to be clear document the fact here too #-executing one select statement per set of nodes instead of per node (chopping into 1,000 node chunks for SQL statement length #-aggregating update statements #-preparing one insert statement and re-execing it (SQL-92 multi-row insert isn't ubiquitous enough) my $nodekey = "node"; if (defined $xCAT::Schema::tabspec{$self->{tabname}}->{nodecol}) { $nodekey = $xCAT::Schema::tabspec{$self->{tabname}}->{nodecol} }; @orderedcols = keys %cols; #pick a specific column ordering explicitly to assure consistency my @sqlorderedcols=(); # must quote to protect from reserved DB keywords foreach my $col (@orderedcols) { my $delimitedcol = &delimitcol($col); push @sqlorderedcols, $delimitedcol; } #use Data::Dumper; my $nodesatatime = 999; #the update case statement will consume '?' of which we are allowed 999 in the most restricted DB we support #ostensibly, we could do 999 at a time for the select statement, and subsequently limit the update aggregation only #to get fewer sql statements, but the code is probably more complex than most people want to read #at the moment anyway my @currnodes = splice(@$nodelist,0,$nodesatatime); #Do a few at a time to stay under max sql statement length and max variable count my $insertsth; #if insert is needed, this will hold the single prepared insert statement my $upsth; my $dnodekey = &delimitcol($nodekey); while (scalar @currnodes) { my %updatenodes=(); my %insertnodes=(); my $qstring; #sort nodes into inserts and updates $qstring = "SELECT * FROM " . $self->{tabname} . " WHERE $dnodekey in ("; $qstring .= '?, ' x scalar(@currnodes); $qstring =~ s/, $/)/; my $query = $self->{dbh}->prepare($qstring); $query->execute(@currnodes); my $rec; while ($rec = $query->fetchrow_hashref()) { $updatenodes{$rec->{$nodekey}}=1; } if (scalar keys %updatenodes < scalar @currnodes) { foreach (@currnodes) { unless ($updatenodes{$_}) { $insertnodes{$_}=1; } } } my $havenodecol; #whether to put node first in execute arguments or let it go naturally if (not $insertsth and keys %insertnodes) { #prepare an insert statement since one will be needed my $columns=""; my $bindhooks=""; $havenodecol = defined $cols{$nodekey}; unless ($havenodecol) { $columns = "$dnodekey, "; $bindhooks="?, "; } $columns .= join(", ",@sqlorderedcols); $bindhooks .= "?, " x scalar @sqlorderedcols; $bindhooks =~ s/, $//; $columns =~ s/, $//; my $instring = "INSERT INTO ".$self->{tabname}." ($columns) VALUES ($bindhooks)"; #print $instring; $insertsth = $self->{dbh}->prepare($instring); } foreach my $node (keys %insertnodes) { my @args = (); unless ($havenodecol) { @args = ($node); } foreach my $col (@orderedcols) { push @args,$hashrec->{$node}->{$col}; } $insertsth->execute(@args); } if (not $upsth and keys %updatenodes) { #prepare an insert statement since one will be needed my $upstring = "UPDATE ".$self->{tabname}." set "; foreach my $col (@sqlorderedcols) { #try aggregating requests. Could also see about single prepare, multiple executes instead $upstring .= "$col = ?, "; } if (grep { $_ eq $nodekey } @orderedcols) { $upstring =~ s/, \z//; } else { $upstring =~ s/, \z/ where $dnodekey = ?/; } $upsth = $self->{dbh}->prepare($upstring); } if (scalar keys %updatenodes) { my $upstring = "UPDATE ".$self->{tabname}." set "; foreach my $node (keys %updatenodes) { my @args=(); foreach my $col (@orderedcols) { #try aggregating requests. Could also see about single prepare, multiple executes instead push @args,$hashrec->{$node}->{$col}; } push @args,$node; $upsth->execute(@args); } } @currnodes = splice(@$nodelist,0,$nodesatatime); } $self->{dbh}->commit; #commit pending transactions $self->{dbh}->{AutoCommit}=$oldac;#restore autocommit semantics $self->_refresh_cache(); #cache is invalid, refresh } #-------------------------------------------------------------------------- =head3 getNodesAttribs Description: Retrieves the requested attributes for a node list Arguments: Table handle ('self') List ref of nodes Attribute type array Returns: two layer hash reference (->{nodename}->{attrib} Globals: Error: Example: my $ostab = xCAT::Table->new('nodetype'); my $ent = $ostab->getNodesAttribs(\@nodes,['profile','os','arch']); if ($ent) { print $ent->{n1}->{profile} Comments: Using this function will clue the table layer into the atomic nature of the request, and allow shortcuts to be taken as appropriate to fulfill the request at scale. =cut #-------------------------------------------------------------------------------- sub getNodesAttribs { my $self = shift; if ($dbworkerpid) { return dbc_call($self,'getNodesAttribs',@_); } my $nodelist = shift; my %options=(); my @attribs; if (ref $_[0]) { @attribs = @{shift()}; %options = @_; } else { @attribs = @_; } if (scalar($nodelist) > $cachethreshold) { $self->{_use_cache} = 0; $self->{nodelist}->{_use_cache}=0; if ($self->{tabname} eq 'nodelist') { #a sticky situation my @locattribs=@attribs; unless (grep(/^node$/,@locattribs)) { push @locattribs,'node'; } unless (grep(/^groups$/,@locattribs)) { push @locattribs,'node'; } $self->_build_cache(\@locattribs); } else { $self->_build_cache(\@attribs); $self->{nodelist}->_build_cache(['node','groups']); } $self->{_use_cache} = 1; $self->{nodelist}->{_use_cache}=1; } my $rethash; foreach (@$nodelist) { my @nodeentries=$self->getNodeAttribs($_,\@attribs,%options); $rethash->{$_} = \@nodeentries; #$self->getNodeAttribs($_,\@attribs); } $self->_clear_cache; $self->{_use_cache} = 0; $self->{nodelist}->_clear_cache; $self->{nodelist}->{_use_cache} = 0; return $rethash; } sub _refresh_cache { #if cache exists, force a rebuild, leaving reference counts alone my $self = shift; #dbworker check not currently required if ($self->{_use_cache}) { #only do things if cache is set up $self->_build_cache(1); #for now, rebuild the whole thing. #in the future, a faster cache update may be possible #however, the payoff may not be worth it #as this case is so rare #the only known case that trips over this is: #1st noderange starts being expanded #the nodelist is updated by another process #2nd noderange starts being expanded (sharing first cache) # (uses stale nodelist data and misses new nodes, the error) #1st noderange finishes #2nd noderange finishes } return; } sub _clear_cache { #PRIVATE FUNCTION TO EXPIRE CACHED DATA EXPLICITLY #This is no longer sufficient to do at destructor time, as Table objects actually live an indeterminite amount of time now #TODO: only clear cache if ref count mentioned in build_cache is 1, otherwise decrement ref count my $self = shift; if ($dbworkerpid) { return dbc_call($self,'_clear_cache',@_); } if ($self->{_cache_ref} > 1) { #don't clear the cache if there are still live references $self->{_cache_ref} -= 1; return; } elsif ($self->{_cache_ref} == 1) { #If it is 1, decrement to zero and carry on $self->{_cache_ref} = 0; } #it shouldn't have been zero, but whether it was 0 or 1, ensure that the cache is gone $self->{_use_cache}=0; # Signal slow operation to any in-flight operations that may fail with empty cache undef $self->{_tablecache}; undef $self->{_nodecache}; } sub _build_cache { #PRIVATE FUNCTION, PLEASE DON'T CALL DIRECTLY #TODO: increment a reference counter type thing to preserve current cache #Also, if ref count is 1 or greater, and the current cache is less than 3 seconds old, reuse the cache? my $self = shift; if ($dbworkerpid) { return dbc_call($self,'_build_cache',@_); } my $attriblist = shift; my $refresh = not ref $attriblist; #if attriblist is not a reference, it is a refresh request if (not $refresh and $self->{_cache_ref}) { #we have active cache reference, increment counter and return #TODO: ensure that the cache isn't somehow still ludirously old $self->{_cache_ref} += 1; return; } #If here, _cache_ref indicates no cache if (not $refresh) { $self->{_cache_ref} = 1; } my $oldusecache = $self->{_use_cache}; #save previous 'use_cache' setting $self->{_use_cache} = 0; #This function must disable cache #to function my $nodekey = "node"; if (defined $xCAT::Schema::tabspec{$self->{tabname}}->{nodecol}) { $nodekey = $xCAT::Schema::tabspec{$self->{tabname}}->{nodecol} }; unless (grep /^$nodekey$/,@$attriblist) { push @$attriblist,$nodekey; } my @tabcache = $self->getAllAttribs(@$attriblist); $self->{_tablecache} = \@tabcache; $self->{_nodecache} = {}; if ($tabcache[0]->{$nodekey}) { foreach(@tabcache) { push @{$self->{_nodecache}->{$_->{$nodekey}}},$_; } } $self->{_use_cache} = $oldusecache; #Restore setting to previous value $self->{_cachestamp} = time; } # This is a utility function to create a number out of a string, useful for things like round robin algorithms on unnumbered nodes sub mknum { my $string = shift; my $number=0; foreach (unpack("C*",$string)) { #do big endian, then it would make 'fred' and 'free' be one number apart $number += $_; } return $number; } $evalcpt->share('&mknum'); $evalcpt->permit('require'); #-------------------------------------------------------------------------- =head3 getNodeAttribs Description: Retrieves the requested attribute Arguments: Table handle Noderange Attribute type array Returns: Attribute hash ( key attribute type) Globals: Error: Example: my $ostab = xCAT::Table->new('nodetype'); my $ent = $ostab->getNodeAttribs($node,['profile','os','arch']); Comments: none =cut #-------------------------------------------------------------------------------- sub getNodeAttribs { my $self = shift; if ($dbworkerpid) { return dbc_call($self,'getNodeAttribs',@_); } my $node = shift; my @attribs; my %options = (); if (ref $_[0]) { @attribs = @{shift()}; %options = @_; } else { @attribs = @_; } my $datum; my @data = $self->getNodeAttribs_nosub_returnany($node, \@attribs,%options); #my ($datum, $extra) = $self->getNodeAttribs_nosub($node, \@attribs); #if ($extra) { return undef; } # return (undef,"Ambiguous query"); } defined($data[0]) || return undef; #(undef,"No matching entry found in configuration"); unless (scalar keys %{$data[0]}) { return undef; } my $attrib; foreach $datum (@data) { foreach $attrib (@attribs) { unless (defined $datum->{$attrib}) { #skip undefined values, save time next; } if ($datum->{$attrib} =~ /^\/[^\/]*\/[^\/]*\/$/) { my $exp = substr($datum->{$attrib}, 1); chop $exp; my @parts = split('/', $exp, 2); $node =~ s/$parts[0]/$parts[1]/; $datum->{$attrib} = $node; } elsif ($datum->{$attrib} =~ /^\|.*\|.*\|$/) { #Perform arithmetic and only arithmetic operations in bracketed issues on the right. #Tricky part: don't allow potentially dangerous code, only eval if #to-be-evaled expression is only made up of ()\d+-/%$ #Futher paranoia? use Safe module to make sure I'm good my $exp = substr($datum->{$attrib}, 1); chop $exp; my @parts = split('\|', $exp, 2); my $curr; my $next; my $prev; my $retval = $parts[1]; ($curr, $next, $prev) = extract_bracketed($retval, '()', qr/[^()]*/); unless($curr) { #If there were no paramaters to save, treat this one like a plain regex undef $@; #extract_bracketed would have set $@ if it didn't return, undef $@ $retval = $node; $retval =~ s/$parts[0]/$parts[1]/; $datum->{$attrib} = $retval; if ($datum->{$attrib} =~ /^$/) { #If regex forces a blank, act like a normal blank does delete $datum->{$attrib}; } next; #skip the redundancy that follows otherwise } while ($curr) { #my $next = $comps[0]; my $value = $node; $value =~ s/$parts[0]/$curr/; $value = $evalcpt->reval('use integer;'.$value); $retval = $prev . $value . $next; ($curr, $next, $prev) = extract_bracketed($retval, '()', qr/[^()]*/); } undef $@; #At this point, $retval is the expression after being arithmetically contemplated, a generated regex, and therefore #must be applied in total my $answval = $node; $answval =~ s/$parts[0]/$retval/; $datum->{$attrib} = $answval; #$retval; #print Data::Dumper::Dumper(extract_bracketed($parts[1],'()',qr/[^()]*/)); #use text::balanced extract_bracketed to parse earch atom, make sure nothing but arith operators, parans, and numbers are in it to guard against code execution } if ($datum->{$attrib} =~ /^$/) { #If regex forces a blank, act like a normal blank does delete $datum->{$attrib}; } } } return wantarray ? @data : $data[0]; } #-------------------------------------------------------------------------- =head3 getNodeAttribs_nosub Description: Arguments: Returns: Globals: Error: Example: Comments: none =cut #-------------------------------------------------------------------------------- sub getNodeAttribs_nosub_old { my $self = shift; my $node = shift; my $attref = shift; my %options = @_; my @data; my $datum; my @tents; my $return = 0; @tents = $self->getNodeAttribs_nosub_returnany($node, $attref,%options); my $nodekey = "node"; if (defined $xCAT::Schema::tabspec{$self->{tabname}}->{nodecol}) { $nodekey = $xCAT::Schema::tabspec{$self->{tabname}}->{nodecol} } foreach my $tent (@tents) { $datum={}; foreach (@$attref) { if ($tent and defined($tent->{$_})) { $return = 1; $datum->{$_} = $tent->{$_}; if ($options{withattribution} and $_ ne $nodekey) { $datum->{'!!xcatgroupattribution!!'}->{$_} = $tent->{'!!xcatsourcegroup!!'}; } } #else { #attempt to fill in gapped attributes #unless (scalar(@$attref) <= 1) { #my $sent = $self->getNodeAttribs_nos($node, [$_],%options); #if ($sent and defined($sent->{$_})) { #$return = 1; #$datum->{$_} = $sent->{$_}; #if ($options{withattribution} and $_ ne $nodekey) { #$datum->{'!!xcatgroupattribution!!'}->{$_} = $sent->{'!!xcatgroupattribution!!'}->{$_}; #} #} #} #} } push(@data,$datum); } if ($return) { return wantarray ? @data : $data[0]; } else { return undef; } } #-------------------------------------------------------------------------- =head3 getNodeAttribs_nosub_returnany Description: not used, kept for reference Arguments: Returns: Globals: Error: Example: Comments: none =cut #-------------------------------------------------------------------------------- sub getNodeAttribs_nosub_returnany_old { #This is the original function my $self = shift; my $node = shift; my @attribs = @{shift()}; my %options = @_; my @results; #my $recurse = ((scalar(@_) == 1) ? shift : 1); my $nodekey = "node"; if (defined $xCAT::Schema::tabspec{$self->{tabname}}->{nodecol}) { $nodekey = $xCAT::Schema::tabspec{$self->{tabname}}->{nodecol} }; @results = $self->getAttribs({$nodekey => $node}, @attribs); my $data = $results[0]; if (!defined($data)) { my ($nodeghash) = $self->{nodelist}->getAttribs({node => $node}, 'groups'); unless (defined($nodeghash) && defined($nodeghash->{groups})) { return undef; } my @nodegroups = split(/,/, $nodeghash->{groups}); my $group; foreach $group (@nodegroups) { @results = $self->getAttribs({$nodekey => $group}, @attribs); $data = $results[0]; if ($data != undef) { foreach (@results) { if ($_->{$nodekey}) { $_->{$nodekey} = $node; } if ($options{withattribution}) { $_->{'!!xcatgroupattribution!!'} = $group; } }; return @results; } } } else { #Don't need to 'correct' node attribute, considering result of the if that governs this code block? return @results; } return undef; #Made it here, config has no good answer } my $nextRecordAtEnd = qr/\+=NEXTRECORD$/; my $nextRecord = qr/\+=NEXTRECORD/; #this evolved a bit and i intend to rewrite it into something a bit cleaner at some point - cjhardee #looks for all of the requested attributes, looking into the groups of the node if needed sub getNodeAttribs_nosub_returnany { my $self = shift; my $node = shift; my @attribs = @{shift()}; my %options = @_; my @results; my $nodekey = "node"; if (defined $xCAT::Schema::tabspec{$self->{tabname}}->{nodecol}) { $nodekey = $xCAT::Schema::tabspec{$self->{tabname}}->{nodecol} }; @results = $self->getAttribs({$nodekey => $node}, @attribs); my %attribsToDo; for(@attribs) { $attribsToDo{$_} = 0 }; my $attrib; my $result; my $data = $results[0]; if(defined{$data}) { #if there was some data for the node, loop through and check it foreach $result (@results) { foreach $attrib (keys %attribsToDo) { #check each item in the results to see which attributes were satisfied if(defined($result) && defined($result->{$attrib}) && $result->{$attrib} !~ $nextRecordAtEnd) { delete $attribsToDo{$attrib}; } } } } if((keys (%attribsToDo)) == 0) { #if all of the attributes are satisfied, don't look at the groups return @results; } #find the groups for this node my ($nodeghash) = $self->{nodelist}->getAttribs({node => $node}, 'groups'); #no groups for the node, we are done unless (defined($nodeghash) && defined($nodeghash->{groups})) { return @results; } my @nodegroups = split(/,/, $nodeghash->{groups}); my $group; my @groupResults; my $groupResult; my %attribsDone; #print "After node results, still missing ".Dumper(\%attribsToDo)."\n"; #print "groups are ".Dumper(\@nodegroups); foreach $group (@nodegroups) { use Storable qw(dclone); my @prevResCopy = @{dclone(\@results)}; my @expandedResults; @groupResults = $self->getAttribs({$nodekey => $group}, keys (%attribsToDo)); #print "group results for $group are ".Dumper(\@groupResults)."\n"; $data = $groupResults[0]; if (defined($data)) { #if some attributes came back from the query for this group foreach $groupResult (@groupResults) { my %toPush; foreach $attrib (keys %attribsToDo) { #check each unfinished attribute against the results for this group #print "looking for attrib $attrib\n"; if(defined($groupResult->{$attrib})){ $attribsDone{$attrib} = 0; #print "found attArib $attrib = $groupResult->{$attrib}\n"; #print "and results look like this: \n".Dumper(\@results)."\n\n\n"; foreach $result (@results){ #loop through our existing results to add or modify the value for this attribute if(defined($result)) { if(defined($result->{$attrib})) { if($result->{$attrib} =~$nextRecordAtEnd){ #if the attribute value should be added $result->{$attrib} =~ s/$nextRecordAtEnd//; #pull out the existing next record string $result->{$attrib} .= $groupResult->{$attrib}; #add the group result onto the end of the existing value if($groupResult->{$attrib} =~ $nextRecordAtEnd && defined($attribsDone{$attrib})){ delete $attribsDone{$attrib}; } if($options{withattribution} && $attrib ne $nodekey) { if(defined($result->{'!!xcatgroupattribution!!'})) { if(defined($result->{'!!xcatgroupattribution!!'}->{$attrib})) { $result->{'!!xcatgroupattribution!!'}->{$attrib} .= "," . $group; } else { $result->{'!!xcatgroupattribution!!'}->{$attrib} = $node.",".$group; } } else { $result->{'!!xcatgroupattribution!!'}->{$attrib} = $node.",".$group; } } } } else {#attribute did not already have an entry #print "attrib $attrib was added with value $groupResult->{$attrib}\n"; $result->{$attrib} = $groupResult->{$attrib}; if($options{withattribution} && $attrib ne $nodekey){ $result->{'!!xcatgroupattribution!!'}->{$attrib} = $group; } if($groupResult->{$attrib} =~ $nextRecordAtEnd && defined($attribsDone{$attrib})){ delete $attribsDone{$attrib}; } } } else {#no results in the array so far #print "pushing for the first time. attr=$attrib groupResults=$groupResult->{$attrib}\n"; $toPush{$attrib} = $groupResult->{$attrib}; if($options{withattribution} && $attrib ne $nodekey){ $toPush{'!!xcatgroupattribution!!'}->{$attrib} = $group; } if($groupResult->{$nodekey}) { $toPush{$nodekey} = $node; } if($groupResult->{$attrib} =~ $nextRecordAtEnd && defined($attribsDone{$attrib})){ delete $attribsDone{$attrib}; } } } } } if(keys(%toPush) > 0) { #print "pushing ".Dumper(\%toPush)."\n"; if(!defined($results[0])) { shift(@results); } push(@results,\%toPush); } #print "pushing results into expanded results\n"; #print "results= ".Dumper(\@results)."\n"; push(@expandedResults, @results); #print "expandedResults= ".Dumper(\@expandedResults)."\n"; #print "setting results to previous:\n".Dumper(\@prevResCopy)."\n\n\n"; @results = @{dclone(\@prevResCopy)}; } @results = @expandedResults; foreach $attrib (keys %attribsDone) { if(defined($attribsToDo{$attrib})) { delete $attribsToDo{$attrib}; } } if((keys (%attribsToDo)) == 0) { #all of the attributes are satisfied, so stop looking at the groups last; } } } #print "results ".Dumper(\@results); #run through the results and remove any "+=NEXTRECORD" ocurrances for $result (@results) { for my $key (keys %$result) { $result->{$key} =~ s/\+=NEXTRECORD//g; } } return @results; } #-------------------------------------------------------------------------- =head3 getAllEntries Description: Read entire table Arguments: Table handle "all" return all lines ( even disabled) Default is to return only lines that have not been disabled Returns: Hash containing all rows in table Globals: Error: Example: my $tabh = xCAT::Table->new($table); my $recs=$tabh->getAllEntries(); # returns entries not disabled my $recs=$tabh->getAllEntries("all"); # returns all entries Comments: none =cut #-------------------------------------------------------------------------------- sub getAllEntries { my $self = shift; if ($dbworkerpid) { return dbc_call($self,'getAllEntries',@_); } my $allentries = shift; my @rets; my $query; # delimit the disable column based on the DB my $disable= &delimitcol("disable"); if ($allentries) { # get all lines $query = $self->{dbh}->prepare('SELECT * FROM ' . $self->{tabname}); } else { # get only enabled lines my $qstring = 'SELECT * FROM ' . $self->{tabname} . " WHERE " . $disable . " is NULL or " . $disable . " in ('0','no','NO','No','nO')"; $query = $self->{dbh}->prepare($qstring); } $query->execute(); while (my $data = $query->fetchrow_hashref()) { foreach (keys %$data) { if ($data->{$_} =~ /^$/) { $data->{$_} = undef; } } push @rets, $data; } $query->finish(); return \@rets; } #-------------------------------------------------------------------------- =head3 getAllAttribsWhere Description: Get all attributes with "where" clause When using a general Where clause with SQL statement then because we support mulitiple databases (SQLite,MySQL and DB2) that require different syntax. Any code using this routine, must call the Utils->getDBName routine and code the where clause that is appropriate for each supported database. When the input is the array of attr val strings, the routine will build the correct Where clause for the database we are running. Arguments: Database Handle Where clause or array of attrval strings to be build into a Where clause Returns: Array of attributes Globals: Error: Example: General Where clause: $nodelist->getAllAttribsWhere("groups like '%".$atom."%'",'node','group'); returns node and group attributes $nodelist->getAllAttribsWhere("groups like '%".$atom."%'",'ALL'); returns all attributes Input of attrval strings $nodelist->getAllAttribsWhere(array of attrval,'node','group'); returns node and group attributes $nodelist->getAllAttribsWhere(array of attrval,'ALL'); returns all attributes where operator can be (==,!=,=~,!~, >, <, >=,<=) Comments: none =cut #-------------------------------------------------------------------------------- sub getAllAttribsWhere { #Takes a list of attributes, returns all records in the table. my $self = shift; if ($dbworkerpid) { return dbc_call($self,'getAllAttribsWhere',@_); } my $clause = shift; my $whereclause; my @attribs = @_; my @results = (); my $query; my $query2; if(ref($clause) eq 'ARRAY'){ $whereclause = &buildWhereClause($clause); } else { $whereclause = $clause; } # delimit the disable column based on the DB my $disable= &delimitcol("disable"); $query2='SELECT * FROM ' . $self->{tabname} . ' WHERE (' . $whereclause . ") and ($disable is NULL or $disable in ('0','no','NO','No','nO'))"; $query = $self->{dbh}->prepare($query2); $query->execute(); while (my $data = $query->fetchrow_hashref()) { my %newrow = (); if ($attribs[0] eq "ALL") { # want all attributes foreach (keys %$data){ if ($data->{$_} =~ /^$/) { $data->{$_} = undef; } } push @results, $data; } else { # want specific attributes foreach (@attribs) { unless ($data->{$_} =~ /^$/ || !defined($data->{$_})) { #The reason we do this is to undef fields in rows that may still be returned.. $newrow{$_} = $data->{$_}; } } if (keys %newrow) { push(@results, \%newrow); } } } $query->finish(); return @results; } #-------------------------------------------------------------------------- =head3 getAllNodeAttribs Description: Get all the node attributes values for the input table on the attribute list Arguments: Table handle Attribute list Returns: Array of attribute values Globals: Error: Example: my @entries = $self->{switchtab}->getAllNodeAttribs(['port','switch']); Comments: none =cut #-------------------------------------------------------------------------------- sub getAllNodeAttribs { #Extract and substitute every node record, expanding groups and substituting as getNodeAttribs does my $self = shift; if ($dbworkerpid) { return dbc_call($self,'getAllNodeAttribs',@_); } my $attribq = shift; my $hashretstyle = shift; my $rethash; my @results = (); my %donenodes ; #Remember those that have been done once to not return same node multiple times my $query; my $nodekey = "node"; if (defined $xCAT::Schema::tabspec{$self->{tabname}}->{nodecol}) { $nodekey = $xCAT::Schema::tabspec{$self->{tabname}}->{nodecol} }; # delimit the disable column based on the DB my $disable= &delimitcol("disable"); my $dnodekey= &delimitcol($nodekey); my $qstring= 'SELECT '.$dnodekey.' FROM ' . $self->{tabname} . " WHERE " . $disable . " is NULL or " . $disable . " in ('0','no','NO','No','nO')"; $query = $self->{dbh}->prepare($qstring); $query->execute(); xCAT::NodeRange::retain_cache(1); $self->{_use_cache} = 0; $self->{nodelist}->{_use_cache}=0; $self->_clear_cache(); $self->{nodelist}->_clear_cache(); $self->_build_cache($attribq); $self->{nodelist}->_build_cache(['node','groups']); $self->{_use_cache} = 1; $self->{nodelist}->{_use_cache}=1; while (my $data = $query->fetchrow_hashref()) { unless ($data->{$nodekey} =~ /^$/ || !defined($data->{$nodekey})) { #ignore records without node attrib, not possible? my @nodes = xCAT::NodeRange::noderange($data->{$nodekey}) ; #expand node entry, to make groups expand unless (@nodes) { #in the event of an entry not in nodelist, use entry value verbatim @nodes = ($data->{$nodekey}); } #my $localhash = $self->getNodesAttribs(\@nodes,$attribq); #NOTE: This is stupid, rebuilds the cache for every entry, FIXME foreach (@nodes) { if ($donenodes{$_}) { next; } my $attrs; my $nde = $_; #if ($self->{giveand}) { #software requests each attribute be independently inherited # foreach (@attribs) { # my $attr = $self->getNodeAttribs($nde,$_); # $attrs->{$_}=$attr->{$_}; # } #} else { my @attrs = $self->getNodeAttribs($_, $attribq);#@{$localhash->{$_}} #$self->getNodeAttribs($_, $attribq) ; #Logic moves to getNodeAttribs #} #populate node attribute by default, this sort of expansion essentially requires it. #$attrs->{node} = $_; foreach my $att (@attrs) { $att->{$nodekey} = $_; } $donenodes{$_} = 1; if ($hashretstyle) { $rethash->{$_} = \@attrs; #$self->getNodeAttribs($_,\@attribs); } else { push @results, @attrs; #$self->getNodeAttribs($_,@attribs); } } } } $self->_clear_cache(); $self->{nodelist}->_clear_cache(); $self->{_use_cache} = 0; $self->{nodelist}->{_use_cache} = 0; xCAT::NodeRange::retain_cache(0); $query->finish(); if ($hashretstyle) { return $rethash; } else { return @results; } } #-------------------------------------------------------------------------- =head3 getAllAttribs Description: Returns a list of records in the input table for the input list of attributes. Arguments: Table handle List of attributes Returns: Array of attribute values Globals: Error: Example: $nodelisttab = xCAT::Table->new("nodelist"); my @attribs = ("node"); @nodes = $nodelisttab->getAllAttribs(@attribs); Comments: none =cut #-------------------------------------------------------------------------------- sub getAllAttribs { #Takes a list of attributes, returns all records in the table. my $self = shift; if ($dbworkerpid) { return dbc_call($self,'getAllAttribs',@_); } #print "Being asked to dump ".$self->{tabname}."for something\n"; my @attribs = @_; my @results = (); if ($self->{_use_cache}) { my @results; my $cacheline; CACHELINE: foreach $cacheline (@{$self->{_tablecache}}) { my $attrib; my %rethash; foreach $attrib (@attribs) { unless ($cacheline->{$attrib} =~ /^$/ || !defined($cacheline->{$attrib})) { #To undef fields in rows that may still be returned $rethash{$attrib} = $cacheline->{$attrib}; } } if (keys %rethash) { push @results, \%rethash; } } if (@results) { return @results; #return wantarray ? @results : $results[0]; } return undef; } # delimit the disable column based on the DB my $disable= &delimitcol("disable"); my $query; my $qstring = "SELECT * FROM " . $self->{tabname} . " WHERE " . $disable . " is NULL or " . $disable . " in ('0','no','NO','No','nO')"; $query = $self->{dbh}->prepare($qstring); #print $query; $query->execute(); while (my $data = $query->fetchrow_hashref()) { my %newrow = (); foreach (@attribs) { unless ($data->{$_} =~ /^$/ || !defined($data->{$_})) { #The reason we do this is to undef fields in rows that may still be returned.. $newrow{$_} = $data->{$_}; } } if (keys %newrow) { push(@results, \%newrow); } } $query->finish(); return @results; } #-------------------------------------------------------------------------- =head3 delEntries Description: Delete table entries Arguments: Table Handle Entry to delete Returns: Globals: Error: Example: my $table=xCAT::Table->new("notification", -create => 1,-autocommit => 0); my %key_col = (filename=>$fname); $table->delEntries(\%key_col); $table->commit; Comments: none =cut #-------------------------------------------------------------------------------- sub delEntries { my $self = shift; if ($dbworkerpid) { return dbc_call($self,'delEntries',@_); } my $keyref = shift; my @all_keyparis; my %keypairs; if (not $self->{intransaction} and not $self->{autocommit} and $self->{realautocommit}) { #search this code for the other if statement just like it for an explanation of why I do this $self->{intransaction}=1; $self->{dbh}->{AutoCommit}=0; } if (ref($keyref) eq 'ARRAY') { @all_keyparis = @{$keyref}; }else { push @all_keyparis, $keyref; } my $notif = xCAT::NotifHandler->needToNotify($self->{tabname}, 'd'); my $record_num = 100; my @pieces = splice(@all_keyparis, 0, $record_num); while (@pieces) { my @notif_data; if ($notif == 1) { my $qstring = "SELECT * FROM " . $self->{tabname}; if ($keyref) { $qstring .= " WHERE "; } my @qargs = (); foreach my $keypairs (@pieces) { $qstring .= "("; foreach my $keypair (keys %{$keypairs}) { # delimit the columns of the table my $dkeypair= &delimitcol($keypair); $qstring .= "$dkeypair = ? AND "; push @qargs, $keypairs->{$keypair}; } $qstring =~ s/ AND \z//; $qstring .= ") OR "; } $qstring =~ s/\(\)//; $qstring =~ s/ OR \z//; my $query = $self->{dbh}->prepare($qstring); $query->execute(@qargs); #prepare the notification data #put the column names at the very front push(@notif_data, $query->{NAME}); my $temp_data = $query->fetchall_arrayref(); foreach (@$temp_data) { push(@notif_data, $_); } $query->finish(); } my @stargs = (); my $delstring = 'DELETE FROM ' . $self->{tabname}; if ($keyref) { $delstring .= ' WHERE '; } foreach my $keypairs (@pieces) { $delstring .= "("; foreach my $keypair (keys %{$keypairs}) { my $dkeypair= &delimitcol($keypair); $delstring .= $dkeypair . ' = ? AND '; if (ref($keypairs->{$keypair})) { #XML transformed data may come in mangled unreasonably into listrefs push @stargs, $keypairs->{$keypair}->[0]; } else { push @stargs, $keypairs->{$keypair}; } } $delstring =~ s/ AND \z//; $delstring .= ") OR "; } $delstring =~ s/\(\)//; $delstring =~ s/ OR \z//; my $stmt = $self->{dbh}->prepare($delstring); $stmt->execute(@stargs); $stmt->finish; #notify the interested parties if ($notif == 1) { xCAT::NotifHandler->notify("d", $self->{tabname}, \@notif_data, {}); } @pieces = splice(@all_keyparis, 0, $record_num); } } #-------------------------------------------------------------------------- =head3 getAttribs Description: Arguments: key List of attributes Returns: Hash of requested attributes Globals: Error: Example: $table = xCAT::Table->new('passwd'); @tmp=$table->getAttribs({'key'=>'ipmi'},('username','password')); Comments: none =cut #-------------------------------------------------------------------------------- sub getAttribs { #Takes two arguments: #-Node name (will be compared against the 'Node' column) #-List reference of attributes for which calling code wants at least one of defined # (recurse argument intended only for internal use.) # Returns a hash reference with requested attributes defined. my $self = shift; if ($dbworkerpid) { return dbc_call($self,'getAttribs',@_); } #my $key = shift; #my $keyval = shift; my %keypairs = %{shift()}; my @attribs; if (ref $_[0]) { @attribs = @{shift()}; } else { @attribs = @_; } my @return; if ($self->{_use_cache}) { my @results; my $cacheline; if (scalar(keys %keypairs) == 1 and $keypairs{node}) { #99.9% of queries look like this, optimized case foreach $cacheline (@{$self->{_nodecache}->{$keypairs{node}}}) { my $attrib; my %rethash; foreach $attrib (@attribs) { unless ($cacheline->{$attrib} =~ /^$/ || !defined($cacheline->{$attrib})) { #To undef fields in rows that may still be returned $rethash{$attrib} = $cacheline->{$attrib}; } } if (keys %rethash) { push @results, \%rethash; } } } else { #SLOW WAY FOR GENERIC CASE CACHELINE: foreach $cacheline (@{$self->{_tablecache}}) { foreach (keys %keypairs) { if (not $keypairs{$_} and $keypairs{$_} ne 0 and $cacheline->{$_}) { next CACHELINE; } unless ($keypairs{$_} eq $cacheline->{$_}) { next CACHELINE; } } my $attrib; my %rethash; foreach $attrib (@attribs) { unless ($cacheline->{$attrib} =~ /^$/ || !defined($cacheline->{$attrib})) { #To undef fields in rows that may still be returned $rethash{$attrib} = $cacheline->{$attrib}; } } if (keys %rethash) { push @results, \%rethash; } } } if (@results) { return wantarray ? @results : $results[0]; } return undef; } #print "Uncached access to ".$self->{tabname}."\n"; my $statement = 'SELECT * FROM ' . $self->{tabname} . ' WHERE '; my @exeargs; foreach (keys %keypairs) { my $dkeypair= &delimitcol($_); if ($keypairs{$_}) { $statement .= $dkeypair . " = ? and "; if (ref($keypairs{$_})) { #correct for XML process mangling if occurred push @exeargs, $keypairs{$_}->[0]; } else { push @exeargs, $keypairs{$_}; } } else { $statement .= $dkeypair . " is NULL and " ; } } # delimit the disable column based on the DB my $disable= &delimitcol("disable"); $statement .= "(" . $disable . " is NULL or " . $disable . " in ('0','no','NO','No','nO'))"; #print "This is my statement: $statement \n"; my $query = $self->{dbh}->prepare($statement); unless (defined $query) { return undef; } $query->execute(@exeargs); my $data; while ($data = $query->fetchrow_hashref()) { my $attrib; my %rethash; foreach $attrib (@attribs) { unless ($data->{$attrib} =~ /^$/ || !defined($data->{$attrib})) { #To undef fields in rows that may still be returned $rethash{$attrib} = $data->{$attrib}; } } if (keys %rethash) { push @return, \%rethash; } } $query->finish(); if (@return) { return wantarray ? @return : $return[0]; } return undef; } #-------------------------------------------------------------------------- =head3 getTable Description: Read entire Table Arguments: Table Handle Returns: Array of table rows Globals: Error: Example: my $table=xCAT::Table->new("notification", -create =>0); my @row_array= $table->getTable; Comments: none =cut #-------------------------------------------------------------------------------- sub getTable { # Get contents of table # Takes no arguments # Returns an array of hashes containing the entire contents of this # table. Each array entry contains a pointer to a hash which is # one row of the table. The row hash is keyed by attribute name. my $self = shift; if ($dbworkerpid) { return dbc_call($self,'getTable',@_); } my @return; my $statement = 'SELECT * FROM ' . $self->{tabname}; my $query = $self->{dbh}->prepare($statement); $query->execute(); my $data; while ($data = $query->fetchrow_hashref()) { my $attrib; my %rethash; foreach $attrib (keys %{$data}) { $rethash{$attrib} = $data->{$attrib}; } if (keys %rethash) { push @return, \%rethash; } } $query->finish(); if (@return) { return @return; } return undef; } #-------------------------------------------------------------------------- =head3 close Description: Close out Table transaction Arguments: Table Handle Returns: Globals: Error: Example: my $mactab = xCAT::Table->new('mac'); $mactab->setNodeAttribs($macmap{$mac},{mac=>$mac}); $mactab->close(); Comments: none =cut #-------------------------------------------------------------------------------- sub close { my $self = shift; #if ($self->{dbh}) { $self->{dbh}->disconnect(); } #undef $self->{dbh}; if ($self->{tabname} eq 'nodelist') { undef $self->{nodelist}; } else { $self->{nodelist}->close(); } } #-------------------------------------------------------------------------- =head3 open Description: Connect to Database Arguments: Empty Hash Returns: Data Base Handle Globals: Error: Example: Comments: none =cut #-------------------------------------------------------------------------------- #UNSUED FUNCTION #sub open #{ # my $self = shift; # $self->{dbh} = DBI->connect($self->{connstring}, "", ""); #} #-------------------------------------------------------------------------- =head3 DESTROY Description: Disconnect from Database Arguments: Database Handle Returns: Globals: Error: Example: Comments: none =cut #-------------------------------------------------------------------------------- sub DESTROY { my $self = shift; $self->{dbh} = ''; undef $self->{dbh}; #if ($self->{dbh}) { $self->{dbh}->disconnect(); undef $self->{dbh};} undef $self->{nodelist}; #Could be circular } =head3 getTableList Description: Returns a list of the table names in the xCAT database. =cut sub getTableList { return keys %xCAT::Schema::tabspec; } =head3 getTableSchema Description: Returns the db schema for the specified table. Returns: A reference to a hash that contains the cols, keys, etc. for this table. (See Schema.pm for details.) =cut sub getTableSchema { return $xCAT::Schema::tabspec{$_[1]}; } =head3 getTableList Description: Returns a summary description for each table. Returns: A reference to a hash. Each key is the table name. Each value is the table description. =cut sub getDescriptions { my $classname = shift; # we ignore this because this function is static # List each table name and the value for table_desc. my $ret = {}; #my @a = keys %{$xCAT::Schema::tabspec{nodelist}}; print 'a=', @a, "\n"; foreach my $t (keys %xCAT::Schema::tabspec) { $ret->{$t} = $xCAT::Schema::tabspec{$t}->{table_desc}; } return $ret; } #-------------------------------------------------------------------------- =head3 isAKey Description: Checks to see if table field is a table key Arguments: Table field List of keys Returns: 1= is a key 0 = not a key Globals: Error: Example: if(isaKey($key_list, $col)); =cut #-------------------------------------------------------------------------------- sub isAKey { my ($keys,$col) = @_; my @key_list = @$keys; foreach my $key (@key_list) { if ( $col eq $key) { # it is a key return 1; } } return 0; } #-------------------------------------------------------------------------- =head3 getAutoIncrementColumns get a list of column names that are of type "INTEGER AUTO_INCREMENT". Returns: an array of column names that are auto increment. =cut #-------------------------------------------------------------------------------- sub getAutoIncrementColumns { my $self=shift; my $descr=$xCAT::Schema::tabspec{$self->{tabname}}; my $types=$descr->{types}; my @ret=(); foreach my $col (@{$descr->{cols}}) { if (($types) && ($types->{$col})) { if ($types->{$col} =~ /INTEGER AUTO_INCREMENT/) { push(@ret,$col); } } } return @ret; } #-------------------------------------------------------------------------- =head3 Description: get_filelist Arguments: directory,filelist,type Returns: The list of sql files to be processed which consists of all the files with .sql and _.sql or files with .pm and _.pm Globals: Error: Example: my @filelist =get_filelist($directory,$filelist,$type); where type = "sql" or "pm" Note either input a directory path in $directory of an array of full path to filenames in $filelist. See runsqlcmd for example. =cut #-------------------------------------------------------------------------------- sub get_filelist { use File::Basename; my $self=shift; my $directory = shift; my $files = shift; my $ext = shift; my $dbname = "sqlite"; my $xcatcfg = get_xcatcfg(); if ($xcatcfg =~ /^DB2:/) { $dbname = "db2"; } else { if ($xcatcfg =~ /^mysql:/) { $dbname = "mysql"; } else { if ($xcatcfg =~ /^Pg:/) { $dbname = "pgsql"; } } } $directory .= "/"; my @list; # check whether input files or a directory if (@$files) { @list=@$files; } else { @list = glob($directory . "*.$ext"); # all files } my @filelist = (); foreach my $file (@list) { my $filename= basename($file); # strip filename my($name,$ext1) = split '\.', $filename; #my($tmpname,$ext2) = split '\_', $name; my @parts = split '\_', $name; my $ext2 = $parts[-1]; # get last element if ($ext2 eq $dbname) { push @filelist, $file; } else { if ($ext2 eq "") { push @filelist, $file; } else { # if not one of the databases, they just have _ in # the file name if ($ext2 ne "db2" && $ext2 ne "mysql" && $ext2 ne "pgsql" && $ext2 ne "sqlite" ) { push @filelist, $file; } } } $ext2 = ""; $ext1 = ""; } return @filelist; } #-------------------------------------------------------------------------- =head3 Description: delimitcol Arguments: attribute name Returns: The attribute(column) delimited appropriately for the runnning Database Globals: Error: Example: my $delimitedcol=delimitcol($col); =cut #-------------------------------------------------------------------------------- sub delimitcol { my $attrin=shift; #input attribute name my $attrout; #output attribute name my $xcatcfg =get_xcatcfg(); # get database $attrout=$attrin; # for sqlite do nothing if (($xcatcfg =~ /^DB2:/) || ($xcatcfg =~ /^Pg:/)) { $attrout="\"$attrin\""; # use double quotes } else { if ($xcatcfg =~ /^mysql:/) { # use backtick $attrout="\`$attrin\`"; } } return $attrout; } #-------------------------------------------------------------------------- =head3 Description: buildwhereclause Arguments: Array of the following attr val where the operator can be the following: == != =~ !~ > < >= <= Returns: Where clause with SQL appropriate for the running DB Globals: Error: Example: my $whereclause=buildWhereClause(@array); =cut #-------------------------------------------------------------------------------- sub buildWhereClause { my $attrvalstr=shift; # array of atrval strings my $whereclause; # Where Clause my $firstpass=1; foreach my $m (@{$attrvalstr}) { my $attr; my $val; my $operator; if ($firstpass == 1) { # first pass no AND $firstpass = 0; } else { # add an AND $whereclause .=" AND "; } if ($m =~ /^[^=]*\==/) { #attr==val ($attr, $val) = split /==/,$m,2; $operator=' = '; } elsif ($m =~ /^[^=]*=~/) { #attr=~val ($attr, $val) = split /=~/,$m,2; $val =~ s/^\///; $val =~ s/\/$//; $operator=' like '; } elsif ($m =~ /^[^=]*\!=/) { #attr!=val ($attr,$val) = split /!=/,$m,2; $operator=' != '; } elsif ($m =~ /[^=]*!~/) { #attr!~val ($attr,$val) = split /!~/,$m,2; $val =~ s/^\///; $val =~ s/\/$//; $operator=' not like '; } elsif ($m =~ /^[^=]*\<=/) { #attr<=val ($attr, $val) = split /<=/,$m,2; $operator=' <= '; } elsif ($m =~ /^[^=]*\=/) { #attr>=val ($attr, $val) = split />=/,$m,2; $operator=' >= '; } elsif ($m =~ /^[^=]*\>/) { #attr>val ($attr, $val) = split />/,$m,2; $operator=' > '; } else { xCAT::MsgUtils->message("S", "Unsupported operator:$m on -w flag input, could not build a Where Clause."); $whereclause=""; return $whereclause; } my $delimitedattr= &delimitcol($attr); $whereclause .=$delimitedattr; $whereclause .=$operator; #$whereclause .="(\'"; $whereclause .="\'"; $whereclause .=$val; #$whereclause .="\')"; $whereclause .="\'"; } return $whereclause; } 1;