X-Git-Url: https://git.openstreetmap.org./nominatim.git/blobdiff_plain/cf35f1931887e8d3548393d2c04e1bb8cf643e0c..baefb4062f31ebf5668abc0736b19bd8d010e4ed:/utils/setup.php diff --git a/utils/setup.php b/utils/setup.php index 8d0a20b8..eff7b71b 100755 --- a/utils/setup.php +++ b/utils/setup.php @@ -22,6 +22,7 @@ array('create-functions', '', 0, 1, 0, 0, 'bool', 'Create functions'), array('enable-diff-updates', '', 0, 1, 0, 0, 'bool', 'Turn on the code required to make diff updates work'), array('enable-debug-statements', '', 0, 1, 0, 0, 'bool', 'Include debug warning statements in pgsql commands'), + array('ignore-errors', '', 0, 1, 0, 0, 'bool', 'Continue import even when errors in SQL are present (EXPERT)'), array('create-minimal-tables', '', 0, 1, 0, 0, 'bool', 'Create minimal main tables'), array('create-tables', '', 0, 1, 0, 0, 'bool', 'Create main tables'), array('create-partition-tables', '', 0, 1, 0, 0, 'bool', 'Create required partition tables'), @@ -32,13 +33,13 @@ array('disable-token-precalc', '', 0, 1, 0, 0, 'bool', 'Disable name precalculation (EXPERT)'), array('import-tiger-data', '', 0, 1, 0, 0, 'bool', 'Import tiger data (not included in \'all\')'), array('calculate-postcodes', '', 0, 1, 0, 0, 'bool', 'Calculate postcode centroids'), - array('create-roads', '', 0, 1, 0, 0, 'bool', ''), array('osmosis-init', '', 0, 1, 0, 0, 'bool', 'Generate default osmosis configuration'), array('index', '', 0, 1, 0, 0, 'bool', 'Index the data'), array('index-noanalyse', '', 0, 1, 0, 0, 'bool', 'Do not perform analyse operations during index (EXPERT)'), array('index-output', '', 0, 1, 1, 1, 'string', 'File to dump index information to'), array('create-search-indices', '', 0, 1, 0, 0, 'bool', 'Create additional indices required for search and update'), array('create-website', '', 0, 1, 1, 1, 'realpath', 'Create symlinks to setup web directory'), + array('drop', '', 0, 1, 0, 0, 'bool', 'Drop tables needed for updates, making the database readonly (EXPERIMENTAL)'), ); getCmdOpt($_SERVER['argv'], $aCMDOptions, $aCMDResult, true, true); @@ -78,16 +79,20 @@ } // Assume we can steal all the cache memory in the box (unless told otherwise) - $iCacheMemory = (isset($aCMDResult['osm2pgsql-cache'])?$aCMDResult['osm2pgsql-cache']:getCacheMemoryMB()); - if ($iCacheMemory > getTotalMemoryMB()) + if (isset($aCMDResult['osm2pgsql-cache'])) + { + $iCacheMemory = $aCMDResult['osm2pgsql-cache']; + } + else { $iCacheMemory = getCacheMemoryMB(); - echo "WARNING: resetting cache memory to $iCacheMemory\n"; } $aDSNInfo = DB::parseDSN(CONST_Database_DSN); if (!isset($aDSNInfo['port']) || !$aDSNInfo['port']) $aDSNInfo['port'] = 5432; + $fPostgisVersion = (float) CONST_Postgis_Version; + if ($aCMDResult['create-db'] || $aCMDResult['all']) { echo "Create DB\n"; @@ -109,7 +114,7 @@ $oDB =& getDB(); $sVersionString = $oDB->getOne('select version()'); - preg_match('#PostgreSQL ([0-9]+)[.]([0-9]+)[.]([0-9]+) #', $sVersionString, $aMatches); + preg_match('#PostgreSQL ([0-9]+)[.]([0-9]+)[^0-9]#', $sVersionString, $aMatches); if (CONST_Postgresql_Version != $aMatches[1].'.'.$aMatches[2]) { echo "ERROR: PostgreSQL version is not correct. Expected ".CONST_Postgresql_Version." found ".$aMatches[1].'.'.$aMatches[2]."\n"; @@ -125,12 +130,11 @@ pgsqlRunScript('CREATE EXTENSION hstore'); } - $fPostgisVersion = (float) CONST_Postgis_Version; if ($fPostgisVersion < 2.0) { pgsqlRunScriptFile(CONST_Path_Postgresql_Postgis.'/postgis.sql'); pgsqlRunScriptFile(CONST_Path_Postgresql_Postgis.'/spatial_ref_sys.sql'); } else { - pgsqlRunScript('CREATE EXTENSION postgis'); + pgsqlRunScript('CREATE EXTENSION IF NOT EXISTS postgis'); } if ($fPostgisVersion < 2.1) { // Function was renamed in 2.1 and throws an annoying deprecation warning @@ -164,6 +168,13 @@ { pgsqlRunScript('update country_name set partition = 0'); } + + // the following will be needed by create_functions later but + // is only defined in the subsequently called create_tables. + // Create dummies here that will be overwritten by the proper + // versions in create-tables. + pgsqlRunScript('CREATE TABLE place_boundingbox ()'); + pgsqlRunScript('create type wikipedia_article_match as ()'); } if ($aCMDResult['import-data'] || $aCMDResult['all']) @@ -182,7 +193,15 @@ { $osm2pgsql .= ' --flat-nodes '.CONST_Osm2pgsql_Flatnode_File; } - $osm2pgsql .= ' -lsc -O gazetteer --hstore'; + if (CONST_Tablespace_Osm2pgsql_Data) + $osm2pgsql .= ' --tablespace-slim-data '.CONST_Tablespace_Osm2pgsql_Data; + if (CONST_Tablespace_Osm2pgsql_Index) + $osm2pgsql .= ' --tablespace-slim-index '.CONST_Tablespace_Osm2pgsql_Index; + if (CONST_Tablespace_Place_Data) + $osm2pgsql .= ' --tablespace-main-data '.CONST_Tablespace_Place_Data; + if (CONST_Tablespace_Place_Index) + $osm2pgsql .= ' --tablespace-main-index '.CONST_Tablespace_Place_Index; + $osm2pgsql .= ' -lsc -O gazetteer --hstore --number-processes 1'; $osm2pgsql .= ' -C '.$iCacheMemory; $osm2pgsql .= ' -P '.$aDSNInfo['port']; $osm2pgsql .= ' -d '.$aDSNInfo['database'].' '.$aCMDResult['osm-file']; @@ -207,6 +226,15 @@ if ($aCMDResult['enable-debug-statements']) $sTemplate = str_replace('--DEBUG:', '', $sTemplate); if (CONST_Limit_Reindexing) $sTemplate = str_replace('--LIMIT INDEXING:', '', $sTemplate); pgsqlRunScript($sTemplate); + + if ($fPostgisVersion < 2.0) { + echo "Helper functions for postgis < 2.0\n"; + $sTemplate = file_get_contents(CONST_BasePath.'/sql/postgis_15_aux.sql'); + } else { + echo "Helper functions for postgis >= 2.0\n"; + $sTemplate = file_get_contents(CONST_BasePath.'/sql/postgis_20_aux.sql'); + } + pgsqlRunScript($sTemplate); } if ($aCMDResult['create-minimal-tables']) @@ -241,13 +269,30 @@ if ($aCMDResult['create-tables'] || $aCMDResult['all']) { - echo "Tables\n"; $bDidSomething = true; - pgsqlRunScriptFile(CONST_BasePath.'/sql/tables.sql'); + + echo "Tables\n"; + $sTemplate = file_get_contents(CONST_BasePath.'/sql/tables.sql'); + $sTemplate = str_replace('{www-user}', CONST_Database_Web_User, $sTemplate); + $sTemplate = replace_tablespace('{ts:address-data}', + CONST_Tablespace_Address_Data, $sTemplate); + $sTemplate = replace_tablespace('{ts:address-index}', + CONST_Tablespace_Address_Index, $sTemplate); + $sTemplate = replace_tablespace('{ts:search-data}', + CONST_Tablespace_Search_Data, $sTemplate); + $sTemplate = replace_tablespace('{ts:search-index}', + CONST_Tablespace_Search_Index, $sTemplate); + $sTemplate = replace_tablespace('{ts:aux-data}', + CONST_Tablespace_Aux_Data, $sTemplate); + $sTemplate = replace_tablespace('{ts:aux-index}', + CONST_Tablespace_Aux_Index, $sTemplate); + pgsqlRunScript($sTemplate, false); // re-run the functions + echo "Functions\n"; $sTemplate = file_get_contents(CONST_BasePath.'/sql/functions.sql'); - $sTemplate = str_replace('{modulepath}',CONST_BasePath.'/module', $sTemplate); + $sTemplate = str_replace('{modulepath}', + CONST_BasePath.'/module', $sTemplate); pgsqlRunScript($sTemplate); } @@ -265,6 +310,18 @@ if (!$aCMDResult['no-partitions']) $aPartitions[] = 0; $sTemplate = file_get_contents(CONST_BasePath.'/sql/partition-tables.src.sql'); + $sTemplate = replace_tablespace('{ts:address-data}', + CONST_Tablespace_Address_Data, $sTemplate); + $sTemplate = replace_tablespace('{ts:address-index}', + CONST_Tablespace_Address_Index, $sTemplate); + $sTemplate = replace_tablespace('{ts:search-data}', + CONST_Tablespace_Search_Data, $sTemplate); + $sTemplate = replace_tablespace('{ts:search-index}', + CONST_Tablespace_Search_Index, $sTemplate); + $sTemplate = replace_tablespace('{ts:aux-data}', + CONST_Tablespace_Aux_Data, $sTemplate); + $sTemplate = replace_tablespace('{ts:aux-index}', + CONST_Tablespace_Aux_Index, $sTemplate); preg_match_all('#^-- start(.*?)^-- end#ms', $sTemplate, $aMatches, PREG_SET_ORDER); foreach($aMatches as $aMatch) { @@ -412,40 +469,17 @@ pgsqlRunScript('ANALYSE'); } - if ($aCMDResult['create-roads']) - { - $bDidSomething = true; - - $oDB =& getDB(); - $aDBInstances = array(); - for($i = 0; $i < $iInstances; $i++) - { - $aDBInstances[$i] =& getDB(true); - if (!pg_query($aDBInstances[$i]->connection, 'set enable_bitmapscan = off')) fail(pg_last_error($oDB->connection)); - $sSQL = 'select count(*) from (select insertLocationRoad(partition, place_id, calculated_country_code, geometry) from '; - $sSQL .= 'placex where osm_id % '.$iInstances.' = '.$i.' and rank_search between 26 and 27 and class = \'highway\') as x '; - if ($aCMDResult['verbose']) echo "$sSQL\n"; - if (!pg_send_query($aDBInstances[$i]->connection, $sSQL)) fail(pg_last_error($oDB->connection)); - } - $bAnyBusy = true; - while($bAnyBusy) - { - $bAnyBusy = false; - for($i = 0; $i < $iInstances; $i++) - { - if (pg_connection_busy($aDBInstances[$i]->connection)) $bAnyBusy = true; - } - sleep(1); - echo '.'; - } - echo "\n"; - } - if ($aCMDResult['import-tiger-data']) { $bDidSomething = true; - pgsqlRunScriptFile(CONST_BasePath.'/sql/tiger_import_start.sql'); + $sTemplate = file_get_contents(CONST_BasePath.'/sql/tiger_import_start.sql'); + $sTemplate = str_replace('{www-user}', CONST_Database_Web_User, $sTemplate); + $sTemplate = replace_tablespace('{ts:aux-data}', + CONST_Tablespace_Aux_Data, $sTemplate); + $sTemplate = replace_tablespace('{ts:aux-index}', + CONST_Tablespace_Aux_Index, $sTemplate); + pgsqlRunScript($sTemplate, false); $aDBInstances = array(); for($i = 0; $i < $iInstances; $i++) @@ -453,7 +487,7 @@ $aDBInstances[$i] =& getDB(true); } - foreach(glob(CONST_BasePath.'/data/tiger2011/*.sql') as $sFile) + foreach(glob(CONST_Tiger_Data_Path.'/*.sql') as $sFile) { echo $sFile.': '; $hFile = fopen($sFile, "r"); @@ -497,7 +531,13 @@ } echo "Creating indexes\n"; - pgsqlRunScriptFile(CONST_BasePath.'/sql/tiger_import_finish.sql'); + $sTemplate = file_get_contents(CONST_BasePath.'/sql/tiger_import_finish.sql'); + $sTemplate = str_replace('{www-user}', CONST_Database_Web_User, $sTemplate); + $sTemplate = replace_tablespace('{ts:aux-data}', + CONST_Tablespace_Aux_Data, $sTemplate); + $sTemplate = replace_tablespace('{ts:aux-index}', + CONST_Tablespace_Aux_Index, $sTemplate); + pgsqlRunScript($sTemplate, false); } if ($aCMDResult['calculate-postcodes'] || $aCMDResult['all']) @@ -518,7 +558,7 @@ if (!pg_query($oDB->connection, $sSQL)) fail(pg_last_error($oDB->connection)); } - if ($aCMDResult['osmosis-init'] || $aCMDResult['all']) + if ($aCMDResult['osmosis-init'] || ($aCMDResult['all'] && !$aCMDResult['drop'])) // no use doing osmosis-init when dropping update tables { $bDidSomething = true; $oDB =& getDB(); @@ -546,7 +586,7 @@ } // Find the last node in the DB - $iLastOSMID = $oDB->getOne("select max(id) from planet_osm_nodes"); + $iLastOSMID = $oDB->getOne("select max(osm_id) from place where osm_type = 'N'"); // Lookup the timestamp that node was created (less 3 hours for margin for changsets to be closed) $sLastNodeURL = 'http://www.openstreetmap.org/api/0.6/node/'.$iLastOSMID."/1"; @@ -556,49 +596,59 @@ // Search for the correct state file - uses file timestamps so need to sort by date descending $sRepURL = CONST_Replication_Url."/"; - $sRep = file_get_contents($sRepURL."?C=M;O=D"); + $sRep = file_get_contents($sRepURL."?C=M;O=D;F=1"); // download.geofabrik.de: 000/26-Feb-2013 11:53 - // planet.openstreetmap.org: 273/ 22-Mar-2013 07:41 - - preg_match_all('#([0-9]{3}/).*(([0-9]{2})-([A-z]{3})-([0-9]{4}) ([0-9]{2}):([0-9]{2}))#', $sRep, $aRepMatches, PREG_SET_ORDER); - $aPrevRepMatch = false; - foreach($aRepMatches as $aRepMatch) + // planet.openstreetmap.org: 273/ 2013-03-11 07:41 - + preg_match_all('#([0-9]{3}/)\s*([-0-9a-zA-Z]+ [0-9]{2}:[0-9]{2})#', $sRep, $aRepMatches, PREG_SET_ORDER); + if ($aRepMatches) { - if (strtotime($aRepMatch[2]) < $iLastNodeTimestamp) break; - $aPrevRepMatch = $aRepMatch; - } - if ($aPrevRepMatch) $aRepMatch = $aPrevRepMatch; + $aPrevRepMatch = false; + foreach($aRepMatches as $aRepMatch) + { + if (strtotime($aRepMatch[2]) < $iLastNodeTimestamp) break; + $aPrevRepMatch = $aRepMatch; + } + if ($aPrevRepMatch) $aRepMatch = $aPrevRepMatch; - $sRepURL .= $aRepMatch[1]; - $sRep = file_get_contents($sRepURL."?C=M;O=D"); - preg_match_all('#([0-9]{3}/).*(([0-9]{2})-([A-z]{3})-([0-9]{4}) ([0-9]{2}):([0-9]{2}))#', $sRep, $aRepMatches, PREG_SET_ORDER); - $aPrevRepMatch = false; - foreach($aRepMatches as $aRepMatch) - { - if (strtotime($aRepMatch[2]) < $iLastNodeTimestamp) break; - $aPrevRepMatch = $aRepMatch; - } - if ($aPrevRepMatch) $aRepMatch = $aPrevRepMatch; + $sRepURL .= $aRepMatch[1]; + $sRep = file_get_contents($sRepURL."?C=M;O=D;F=1"); + preg_match_all('#([0-9]{3}/)\s*([-0-9a-zA-Z]+ [0-9]{2}:[0-9]{2})#', $sRep, $aRepMatches, PREG_SET_ORDER); + $aPrevRepMatch = false; + foreach($aRepMatches as $aRepMatch) + { + if (strtotime($aRepMatch[2]) < $iLastNodeTimestamp) break; + $aPrevRepMatch = $aRepMatch; + } + if ($aPrevRepMatch) $aRepMatch = $aPrevRepMatch; - $sRepURL .= $aRepMatch[1]; - $sRep = file_get_contents($sRepURL."?C=M;O=D"); - preg_match_all('#([0-9]{3}).state.txt.*(([0-9]{2})-([A-z]{3})-([0-9]{4}) ([0-9]{2}):([0-9]{2}))#', $sRep, $aRepMatches, PREG_SET_ORDER); - $aPrevRepMatch = false; - foreach($aRepMatches as $aRepMatch) + $sRepURL .= $aRepMatch[1]; + $sRep = file_get_contents($sRepURL."?C=M;O=D;F=1"); + preg_match_all('#([0-9]{3}).state.txt\s*([-0-9a-zA-Z]+ [0-9]{2}:[0-9]{2})#', $sRep, $aRepMatches, PREG_SET_ORDER); + $aPrevRepMatch = false; + foreach($aRepMatches as $aRepMatch) + { + if (strtotime($aRepMatch[2]) < $iLastNodeTimestamp) break; + $aPrevRepMatch = $aRepMatch; + } + if ($aPrevRepMatch) $aRepMatch = $aPrevRepMatch; + + $sRepURL .= $aRepMatch[1].'.state.txt'; + echo "Getting state file: $sRepURL\n"; + $sStateFile = file_get_contents($sRepURL); + if (!$sStateFile || strlen($sStateFile) > 1000) fail("unable to obtain state file"); + file_put_contents(CONST_BasePath.'/settings/state.txt', $sStateFile); + echo "Updating DB status\n"; + pg_query($oDB->connection, 'TRUNCATE import_status'); + $sSQL = "INSERT INTO import_status VALUES('".$aRepMatch[2]."')"; + pg_query($oDB->connection, $sSQL); + } + else { - if (strtotime($aRepMatch[2]) < $iLastNodeTimestamp) break; - $aPrevRepMatch = $aRepMatch; + if (!$aCMDResult['all']) + { + fail("Cannot read state file directory."); + } } - if ($aPrevRepMatch) $aRepMatch = $aPrevRepMatch; - - $sRepURL .= $aRepMatch[1].'.state.txt'; - echo "Getting state file: $sRepURL\n"; - $sStateFile = file_get_contents($sRepURL); - if (!$sStateFile || strlen($sStateFile) > 1000) fail("unable to obtain state file"); - file_put_contents(CONST_BasePath.'/settings/state.txt', $sStateFile); - echo "Updating DB status\n"; - pg_query($oDB->connection, 'TRUNCATE import_status'); - $sSQL = "INSERT INTO import_status VALUES('".$aRepMatch[2]."')"; - pg_query($oDB->connection, $sSQL); } } @@ -629,6 +679,12 @@ if (!$aCMDResult['no-partitions']) $aPartitions[] = 0; $sTemplate = file_get_contents(CONST_BasePath.'/sql/indices.src.sql'); + $sTemplate = replace_tablespace('{ts:address-index}', + CONST_Tablespace_Address_Index, $sTemplate); + $sTemplate = replace_tablespace('{ts:search-index}', + CONST_Tablespace_Search_Index, $sTemplate); + $sTemplate = replace_tablespace('{ts:aux-index}', + CONST_Tablespace_Aux_Index, $sTemplate); preg_match_all('#^-- start(.*?)^-- end#ms', $sTemplate, $aMatches, PREG_SET_ORDER); foreach($aMatches as $aMatch) { @@ -657,6 +713,7 @@ @symlink(CONST_BasePath.'/website/reverse.php', $sTargetDir.'/reverse.php'); @symlink(CONST_BasePath.'/website/search.php', $sTargetDir.'/search.php'); @symlink(CONST_BasePath.'/website/search.php', $sTargetDir.'/index.php'); + @symlink(CONST_BasePath.'/website/lookup.php', $sTargetDir.'/lookup.php'); @symlink(CONST_BasePath.'/website/deletable.php', $sTargetDir.'/deletable.php'); @symlink(CONST_BasePath.'/website/polygons.php', $sTargetDir.'/polygons.php'); @symlink(CONST_BasePath.'/website/status.php', $sTargetDir.'/status.php'); @@ -673,10 +730,77 @@ } } + if ($aCMDResult['drop']) + { + // The implementation is potentially a bit dangerous because it uses + // a positive selection of tables to keep, and deletes everything else. + // Including any tables that the unsuspecting user might have manually + // created. USE AT YOUR OWN PERIL. + $bDidSomething = true; + + // tables we want to keep. everything else goes. + $aKeepTables = array( + "*columns", + "import_polygon_*", + "import_status", + "place_addressline", + "location_property*", + "placex", + "search_name", + "seq_*", + "word", + "query_log", + "new_query_log", + "gb_postcode", + "spatial_ref_sys", + "country_name", + "place_classtype_*" + ); + + $oDB =& getDB(); + $aDropTables = array(); + $aHaveTables = $oDB->getCol("SELECT tablename FROM pg_tables WHERE schemaname='public'"); + if (PEAR::isError($aHaveTables)) + { + fail($aPartitions->getMessage()); + } + foreach($aHaveTables as $sTable) + { + $bFound = false; + foreach ($aKeepTables as $sKeep) + { + if (fnmatch($sKeep, $sTable)) + { + $bFound = true; + break; + } + } + if (!$bFound) array_push($aDropTables, $sTable); + } + + foreach ($aDropTables as $sDrop) + { + if ($aCMDResult['verbose']) echo "dropping table $sDrop\n"; + @pg_query($oDB->connection, "DROP TABLE $sDrop CASCADE"); + // ignore warnings/errors as they might be caused by a table having + // been deleted already by CASCADE + } + + if (!is_null(CONST_Osm2pgsql_Flatnode_File)) + { + if ($aCMDResult['verbose']) echo "deleting ".CONST_Osm2pgsql_Flatnode_File."\n"; + unlink(CONST_Osm2pgsql_Flatnode_File); + } + } + if (!$bDidSomething) { showUsage($aCMDOptions, true); } + else + { + echo "Setup finished.\n"; + } function pgsqlRunScriptFile($sFilename) { @@ -723,7 +847,11 @@ } fclose($ahPipes[1]); - proc_close($hProcess); + $iReturn = proc_close($hProcess); + if ($iReturn > 0) + { + fail("pgsql returned with error code ($iReturn)"); + } if ($ahGzipPipes) { fclose($ahGzipPipes[1]); @@ -732,12 +860,15 @@ } - function pgsqlRunScript($sScript) + function pgsqlRunScript($sScript, $bfatal = true) { + global $aCMDResult; // Convert database DSN to psql parameters $aDSNInfo = DB::parseDSN(CONST_Database_DSN); if (!isset($aDSNInfo['port']) || !$aDSNInfo['port']) $aDSNInfo['port'] = 5432; $sCMD = 'psql -p '.$aDSNInfo['port'].' -d '.$aDSNInfo['database']; + if ($bfatal && !$aCMDResult['ignore-errors']) + $sCMD .= ' -v ON_ERROR_STOP=1'; $aDescriptors = array( 0 => array('pipe', 'r'), 1 => STDOUT, @@ -750,10 +881,15 @@ while(strlen($sScript)) { $written = fwrite($ahPipes[0], $sScript); + if ($written <= 0) break; $sScript = substr($sScript, $written); } fclose($ahPipes[0]); - proc_close($hProcess); + $iReturn = proc_close($hProcess); + if ($bfatal && $iReturn > 0) + { + fail("pgsql returned with error code ($iReturn)"); + } } function pgsqlRunRestoreData($sDumpFile) @@ -781,7 +917,7 @@ } fclose($ahPipes[1]); - proc_close($hProcess); + $iReturn = proc_close($hProcess); } function pgsqlRunDropAndRestore($sDumpFile) @@ -809,7 +945,7 @@ } fclose($ahPipes[1]); - proc_close($hProcess); + $iReturn = proc_close($hProcess); } function passthruCheckReturn($cmd) @@ -818,3 +954,15 @@ passthru($cmd, $result); if ($result != 0) fail('Error executing external command: '.$cmd); } + + function replace_tablespace($sTemplate, $sTablespace, $sSql) + { + if ($sTablespace) + $sSql = str_replace($sTemplate, 'TABLESPACE "'.$sTablespace.'"', + $sSql); + else + $sSql = str_replace($sTemplate, '', $sSql); + + return $sSql; + } +