/*------------------------------------------------------------------------- * * postinit.c * postgres initialization utilities * * Copyright (c) 1994, Regents of the University of California * * * IDENTIFICATION * $Header: /cvsroot/pgsql/src/backend/utils/init/postinit.c,v 1.52 1999/10/25 03:07:51 tgl Exp $ * * NOTES * InitPostgres() is the function called from PostgresMain * which does all non-trival initialization, mainly by calling * all the other initialization functions. InitPostgres() * is only used within the "postgres" backend and so that routine * is in tcop/postgres.c InitPostgres() is needed in cinterface.a * because things like the bootstrap backend program need it. Hence * you find that in this file... * * If you feel the need to add more initialization code, it should be * done in InitPostgres() or someplace lower. Do not start * putting stuff in PostgresMain - if you do then someone * will have to clean it up later, and it's not going to be me! * -cim 10/3/90 * *------------------------------------------------------------------------- */ #include #include #include #include #include #include "postgres.h" #include "access/heapam.h" #include "catalog/catname.h" #include "catalog/pg_database.h" #include "libpq/libpq.h" #include "miscadmin.h" #include "storage/backendid.h" #include "storage/proc.h" #include "storage/sinval.h" #include "storage/smgr.h" #include "utils/inval.h" #include "utils/portal.h" #include "utils/relcache.h" #include "utils/syscache.h" #include "version.h" #ifdef MULTIBYTE #include "mb/pg_wchar.h" #endif void BaseInit(void); static void VerifySystemDatabase(void); static void VerifyMyDatabase(void); static void ReverifyMyDatabase(char *name); static void InitCommunication(void); static void InitMyDatabaseInfo(char *name); static void InitUserid(void); static IPCKey PostgresIpcKey; /* ---------------------------------------------------------------- * InitPostgres support * ---------------------------------------------------------------- */ /* -------------------------------- * InitMyDatabaseInfo() -- Find and record the OID of the database we are * to open. * * The database's oid forms half of the unique key for the system * caches and lock tables. We therefore want it initialized before * we open any relations, since opening relations puts things in the * cache. To get around this problem, this code opens and scans the * pg_database relation by hand. * * This algorithm relies on the fact that first attribute in the * pg_database relation schema is the database name. It also knows * about the internal format of tuples on disk and the length of * the datname attribute. It knows the location of the pg_database * file. * Actually, the code looks as though it is using the pg_database * tuple definition to locate the database name, so the above statement * seems to be no longer correct. - thomas 1997-11-01 * * This code is called from InitPostgres(), before we chdir() to the * local database directory and before we open any relations. * Used to be called after the chdir(), but we now want to confirm * the location of the target database using pg_database info. * - thomas 1997-11-01 * -------------------------------- */ static void InitMyDatabaseInfo(char *name) { char *path, myPath[MAXPGPATH]; SetDatabaseName(name); GetRawDatabaseInfo(name, &MyDatabaseId, myPath); if (!OidIsValid(MyDatabaseId)) elog(FATAL, "Database %s does not exist in %s", DatabaseName, DatabaseRelationName); path = ExpandDatabasePath(myPath); SetDatabasePath(path); } /* InitMyDatabaseInfo() */ /* * DoChdirAndInitDatabaseNameAndPath * Set current directory to the database directory for the database * named . * Also set global variables DatabasePath and DatabaseName to those * values. Also check for proper version of database system and * database. Exit program via elog() if anything doesn't check out. * * Arguments: * Path and name are invalid if it invalid as a string. * Path is "badly formatted" if it is not a string containing a path * to a writable directory. * Name is "badly formatted" if it contains more than 16 characters or if * it is a bad file name (e.g., it contains a '/' or an 8-bit character). * * Exceptions: * BadState if called more than once. * BadArg if both path and name are "badly formatted" or invalid. * BadArg if path and name are both "inconsistent" and valid. * * This routine is inappropriate in bootstrap mode, since the directories * and version files need not exist yet if we're in bootstrap mode. */ static void VerifySystemDatabase() { char *reason; /* Failure reason returned by some function. NULL if no failure */ int fd; char errormsg[MAXPGPATH+100]; errormsg[0] = '\0'; #ifndef __CYGWIN32__ if ((fd = open(DataDir, O_RDONLY, 0)) == -1) #else if ((fd = open(DataDir, O_RDONLY | O_DIROPEN, 0)) == -1) #endif snprintf(errormsg, sizeof(errormsg), "Database system does not exist. " "PGDATA directory '%s' not found.\n\tNormally, you " "create a database system by running initdb.", DataDir); else { close(fd); ValidatePgVersion(DataDir, &reason); if (reason != NULL) snprintf(errormsg, sizeof(errormsg), "InitPostgres could not validate that the database" " system version is compatible with this level of" " Postgres.\n\tYou may need to run initdb to create" " a new database system.\n\t%s", reason); } if (errormsg[0] != '\0') elog(FATAL, errormsg); /* Above does not return */ } /* VerifySystemDatabase() */ static void VerifyMyDatabase() { const char *name; const char *myPath; /* Failure reason returned by some function. NULL if no failure */ char *reason; int fd; char errormsg[MAXPGPATH+100]; name = DatabaseName; myPath = DatabasePath; #ifndef __CYGWIN32__ if ((fd = open(myPath, O_RDONLY, 0)) == -1) #else if ((fd = open(myPath, O_RDONLY | O_DIROPEN, 0)) == -1) #endif snprintf(errormsg, sizeof(errormsg), "Database '%s' does not exist." "\n\tWe know this because the directory '%s' does not exist." "\n\tYou can create a database with the SQL command" " CREATE DATABASE.\n\tTo see what databases exist," " look at the subdirectories of '%s/base/'.", name, myPath, DataDir); else { close(fd); ValidatePgVersion(myPath, &reason); if (reason != NULL) snprintf(errormsg, sizeof(errormsg), "InitPostgres could not validate that the database" " version is compatible with this level of Postgres" "\n\teven though the database system as a whole" " appears to be at a compatible level." "\n\tYou may need to recreate the database with SQL" " commands DROP DATABASE and CREATE DATABASE." "\n\t%s", reason); else { /* * The directories and PG_VERSION files are in order. */ int rc; /* return code from some function we call */ #ifdef FILEDEBUG printf("Try changing directory for database %s to %s\n", name, myPath); #endif rc = chdir(myPath); if (rc < 0) snprintf(errormsg, sizeof(errormsg), "InitPostgres unable to change " "current directory to '%s', errno = %s (%d).", myPath, strerror(errno), errno); else errormsg[0] = '\0'; } } if (errormsg[0] != '\0') elog(FATAL, errormsg); /* Above does not return */ } /* VerifyMyDatabase() */ /* -------------------------------- * ReverifyMyDatabase * * Since we are forced to fetch the database OID out of pg_database without * benefit of locking or transaction ID checking (see utils/misc/database.c), * we might have gotten a wrong answer. Or, we might have attached to a * database that's in process of being destroyed by destroydb(). This * routine is called after we have all the locking and other infrastructure * running --- now we can check that we are really attached to a valid * database. * * In reality, if destroydb() is running in parallel with our startup, * it's pretty likely that we will have failed before now, due to being * unable to read some of the system tables within the doomed database. * This routine just exists to make *sure* we have not started up in an * invalid database. If we quit now, we should have managed to avoid * creating any serious problems. * * This is also a handy place to fetch the database encoding info out * of pg_database, if we are in MULTIBYTE mode. * -------------------------------- */ static void ReverifyMyDatabase(char *name) { Relation pgdbrel; HeapScanDesc pgdbscan; ScanKeyData key; HeapTuple tup; /* * Because we grab AccessShareLock here, we can be sure that * destroydb is not running in parallel with us (any more). */ pgdbrel = heap_openr(DatabaseRelationName, AccessShareLock); ScanKeyEntryInitialize(&key, 0, Anum_pg_database_datname, F_NAMEEQ, NameGetDatum(name)); pgdbscan = heap_beginscan(pgdbrel, 0, SnapshotNow, 1, &key); tup = heap_getnext(pgdbscan, 0); if (!HeapTupleIsValid(tup) || tup->t_data->t_oid != MyDatabaseId) { /* OOPS */ heap_close(pgdbrel, AccessShareLock); /* * The only real problem I could have created is to load dirty * buffers for the dead database into shared buffer cache; * if I did, some other backend will eventually try to write * them and die in mdblindwrt. Flush any such pages to forestall * trouble. */ DropBuffers(MyDatabaseId); /* Now I can commit hara-kiri with a clear conscience... */ elog(FATAL, "Database '%s', OID %u, has disappeared from pg_database", name, MyDatabaseId); } /* * OK, we're golden. Only other to-do item is to save the MULTIBYTE * encoding info out of the pg_database tuple. Note we also set the * "template encoding", which is the default encoding for any * CREATE DATABASE commands executed in this backend; essentially, * you get the same encoding of the database you connected to as * the default. (This replaces code that unreliably grabbed * template1's encoding out of pg_database. We could do an extra * scan to find template1's tuple, but for 99.99% of all backend * startups it'd be wasted cycles --- and the 'createdb' script * connects to template1 anyway, so there's no difference.) */ #ifdef MULTIBYTE SetDatabaseEncoding(((Form_pg_database) GETSTRUCT(tup))->encoding); SetTemplateEncoding(((Form_pg_database) GETSTRUCT(tup))->encoding); #endif heap_endscan(pgdbscan); heap_close(pgdbrel, AccessShareLock); } /* -------------------------------- * InitUserid * * initializes crap associated with the user id. * -------------------------------- */ static void InitUserid() { setuid(geteuid()); SetUserId(); } /* -------------------------------- * InitCommunication * * This routine initializes stuff needed for ipc, locking, etc. * it should be called something more informative. * * Note: * This does not set MyBackendId. MyBackendTag is set, however. * -------------------------------- */ static void InitCommunication() { char *postid; /* value of environment variable */ char *postport; /* value of environment variable */ char *ipc_key; /* value of environemnt variable */ IPCKey key = 0; /* ---------------- * try and get the backend tag from POSTID * ---------------- */ MyBackendId = -1; postid = getenv("POSTID"); if (!PointerIsValid(postid)) MyBackendTag = -1; else { MyBackendTag = atoi(postid); Assert(MyBackendTag >= 0); } ipc_key = getenv("IPC_KEY"); if (!PointerIsValid(ipc_key)) key = -1; else { key = atoi(ipc_key); Assert(MyBackendTag >= 0); } postport = getenv("POSTPORT"); if (PointerIsValid(postport)) { if (MyBackendTag == -1) elog(FATAL, "InitCommunication: missing POSTID"); } else if (IsUnderPostmaster) { elog(FATAL, "InitCommunication: under postmaster and POSTPORT not set"); } else { /* ---------------- * assume we're running a postgres backend by itself with * no front end or postmaster. * ---------------- */ if (MyBackendTag == -1) MyBackendTag = 1; key = PrivateIPCKey; } /* ---------------- * initialize shared memory and semaphores appropriately. * ---------------- */ if (!IsUnderPostmaster) /* postmaster already did this */ { PostgresIpcKey = key; AttachSharedMemoryAndSemaphores(key); } } /* -------------------------------- * InitPostgres * Initialize POSTGRES. * * Note: * Be very careful with the order of calls in the InitPostgres function. * -------------------------------- */ extern int NBuffers; int lockingOff = 0; /* backend -L switch */ /* */ void InitPostgres(char *name) /* database name */ { bool bootstrap; /* true if BootstrapProcessing */ /* * See if we're running in BootstrapProcessing mode */ bootstrap = IsBootstrapProcessingMode(); /* ---------------- * initialize the backend local portal stack used by * internal PQ function calls. see src/lib/libpq/be-dumpdata.c * This is different from the "portal manager" so this goes here. * -cim 2/12/91 * ---------------- */ be_portalinit(); /* * initialize the local buffer manager */ InitLocalBuffer(); #ifndef XLOG if (!TransactionFlushEnabled()) on_shmem_exit(FlushBufferPool, (caddr_t) NULL); #endif /* ---------------- * initialize the database id used for system caches and lock tables * ---------------- */ if (bootstrap) { SetDatabasePath(ExpandDatabasePath(name)); SetDatabaseName(name); LockDisable(true); } else { VerifySystemDatabase(); InitMyDatabaseInfo(name); VerifyMyDatabase(); } /* * Code after this point assumes we are in the proper directory! * * So, how do we implement alternate locations for databases? There are * two possible locations for tables and we need to look in * DataDir/pg_database to find the true location of an individual * database. We can brute-force it as done in InitMyDatabaseInfo(), or * we can be patient and wait until we open pg_database gracefully. * Will try that, but may not work... - thomas 1997-11-01 */ /* * Initialize the transaction system and the relation descriptor cache. * Note we have to make certain the lock manager is off while we do this. */ AmiTransactionOverride(IsBootstrapProcessingMode()); LockDisable(true); /* * Part of the initialization processing done here sets a read lock on * pg_log. Since locking is disabled the set doesn't have intended * effect of locking out writers, but this is ok, since we only lock * it to examine AMI transaction status, and this is never written * after initdb is done. -mer 15 June 1992 */ RelationInitialize(); /* pre-allocated reldescs created here */ InitializeTransactionSystem(); /* pg_log,etc init/crash recovery * here */ LockDisable(false); /* * Set up my per-backend PROC struct in shared memory. */ InitProcess(PostgresIpcKey); /* * Initialize my entry in the shared-invalidation manager's * array of per-backend data. (Formerly this came before * InitProcess, but now it must happen after, because it uses * MyProc.) Once I have done this, I am visible to other backends! * * Sets up MyBackendId, a unique backend identifier. */ InitSharedInvalidationState(); if (MyBackendId > MAXBACKENDS || MyBackendId <= 0) { elog(FATAL, "cinit2: bad backend id %d (%d)", MyBackendTag, MyBackendId); } /* * Initialize the access methods. * Does not touch files (?) - thomas 1997-11-01 */ initam(); /* * Initialize all the system catalog caches. */ zerocaches(); /* * Does not touch files since all routines are builtins (?) - thomas * 1997-11-01 */ InitCatalogCache(); /* * Set ourselves to the proper user id and figure out our postgres * user id. If we ever add security so that we check for valid * postgres users, we might do it here. */ InitUserid(); /* * Initialize local data in cache invalidation stuff */ if (!bootstrap) InitLocalInvalidateData(); if (lockingOff) LockDisable(true); /* * Unless we are bootstrapping, double-check that InitMyDatabaseInfo() * got a correct result. We can't do this until essentially all the * infrastructure is up, so just do it at the end. */ if (!bootstrap) ReverifyMyDatabase(name); } void BaseInit(void) { /* * Turn on the exception handler. Note: we cannot use elog, Assert, * AssertState, etc. until after exception handling is on. */ EnableExceptionHandling(true); /* * Memory system initialization - we may call palloc after * EnableMemoryContext()). Note that EnableMemoryContext() * must happen before EnablePortalManager(). */ EnableMemoryContext(true); /* initializes the "top context" */ EnablePortalManager(true); /* memory for portal/transaction stuff */ /* * Attach to shared memory and semaphores, and initialize our * input/output/debugging file descriptors. */ InitCommunication(); DebugFileOpen(); smgrinit(); }