/* * Activate this module whenever necessary. * This must happen during postmaster or standalong-backend startup, * or during WAL replay anytime the track_commit_timestamp setting is * changed in the master. * * The reason why this SLRU needs separate activation/deactivation functions is * that it can be enabled/disabled during start and the activation/deactivation * on master is propagated to slave via replay. Other SLRUs don't have this * property and they can be just initialized during normal startup. * * This is in charge of creating the currently active segment, if it's not * already there. The reason for this is that the server might have been * running with this module disabled for a while and thus might have skipped * the normal creation point. */ static void ActivateCommitTs(void) { TransactionId xid = ShmemVariableCache->nextXid; int pageno = TransactionIdToCTsPage(xid); /* * Re-Initialize our idea of the latest page number. */ LWLockAcquire(CommitTsControlLock, LW_EXCLUSIVE); CommitTsCtl->shared->latest_page_number = pageno; LWLockRelease(CommitTsControlLock); /* * If CommitTs is enabled, but it wasn't in the previous server run, we * need to set the oldest and newest values to the next Xid; that way, we * will not try to read data that might not have been set. * * XXX does this have a problem if a server is started with commitTs * enabled, then started with commitTs disabled, then restarted with it * enabled again? It doesn't look like it does, because there should be a * checkpoint that sets the value to InvalidTransactionId at end of * recovery; and so any chance of injecting new transactions without * CommitTs values would occur after the oldestCommitTs has been set to * Invalid temporarily. */ LWLockAcquire(CommitTsLock, LW_EXCLUSIVE); if (ShmemVariableCache->oldestCommitTs == InvalidTransactionId) { ShmemVariableCache->oldestCommitTs = ShmemVariableCache->newestCommitTs = ReadNewTransactionId(); } LWLockRelease(CommitTsLock); /* Create the current segment file, if necessary */ if (!SimpleLruDoesPhysicalPageExist(CommitTsCtl, pageno)) { int slotno; LWLockAcquire(CommitTsControlLock, LW_EXCLUSIVE); slotno = ZeroCommitTsPage(pageno, false); SimpleLruWritePage(CommitTsCtl, slotno); Assert(!CommitTsCtl->shared->page_dirty[slotno]); LWLockRelease(CommitTsControlLock); } /* Change the activation status in shared memory. */ LWLockAcquire(CommitTsLock, LW_EXCLUSIVE); commitTsShared->commitTsActive = true; LWLockRelease(CommitTsLock); }
/* * _bt_pagedel() -- Delete a page from the b-tree, if legal to do so. * * This action unlinks the page from the b-tree structure, removing all * pointers leading to it --- but not touching its own left and right links. * The page cannot be physically reclaimed right away, since other processes * may currently be trying to follow links leading to the page; they have to * be allowed to use its right-link to recover. See nbtree/README. * * On entry, the target buffer must be pinned and locked (either read or write * lock is OK). This lock and pin will be dropped before exiting. * * The "stack" argument can be a search stack leading (approximately) to the * target page, or NULL --- outside callers typically pass NULL since they * have not done such a search, but internal recursion cases pass the stack * to avoid duplicated search effort. * * Returns the number of pages successfully deleted (zero if page cannot * be deleted now; could be more than one if parent pages were deleted too). * * NOTE: this leaks memory. Rather than trying to clean up everything * carefully, it's better to run it in a temp context that can be reset * frequently. */ int _bt_pagedel(Relation rel, Buffer buf, BTStack stack, bool vacuum_full) { int result; BlockNumber target, leftsib, rightsib, parent; OffsetNumber poffset, maxoff; uint32 targetlevel, ilevel; ItemId itemid; IndexTuple targetkey, itup; ScanKey itup_scankey; Buffer lbuf, rbuf, pbuf; bool parent_half_dead; bool parent_one_child; bool rightsib_empty; Buffer metabuf = InvalidBuffer; Page metapg = NULL; BTMetaPageData *metad = NULL; Page page; BTPageOpaque opaque; MIRROREDLOCK_BUFMGR_MUST_ALREADY_BE_HELD; // Fetch gp_persistent_relation_node information that will be added to XLOG record. RelationFetchGpRelationNodeForXLog(rel); /* * We can never delete rightmost pages nor root pages. While at it, check * that page is not already deleted and is empty. */ page = BufferGetPage(buf); opaque = (BTPageOpaque) PageGetSpecialPointer(page); if (P_RIGHTMOST(opaque) || P_ISROOT(opaque) || P_ISDELETED(opaque) || P_FIRSTDATAKEY(opaque) <= PageGetMaxOffsetNumber(page)) { /* Should never fail to delete a half-dead page */ Assert(!P_ISHALFDEAD(opaque)); _bt_relbuf(rel, buf); return 0; } /* * Save info about page, including a copy of its high key (it must have * one, being non-rightmost). */ target = BufferGetBlockNumber(buf); targetlevel = opaque->btpo.level; leftsib = opaque->btpo_prev; itemid = PageGetItemId(page, P_HIKEY); targetkey = CopyIndexTuple((IndexTuple) PageGetItem(page, itemid)); /* * To avoid deadlocks, we'd better drop the target page lock before going * further. */ _bt_relbuf(rel, buf); /* * We need an approximate pointer to the page's parent page. We use the * standard search mechanism to search for the page's high key; this will * give us a link to either the current parent or someplace to its left * (if there are multiple equal high keys). In recursion cases, the * caller already generated a search stack and we can just re-use that * work. */ if (stack == NULL) { if (!InRecovery) { /* we need an insertion scan key to do our search, so build one */ itup_scankey = _bt_mkscankey(rel, targetkey); /* find the leftmost leaf page containing this key */ stack = _bt_search(rel, rel->rd_rel->relnatts, itup_scankey, false, &lbuf, BT_READ); /* don't need a pin on that either */ _bt_relbuf(rel, lbuf); /* * If we are trying to delete an interior page, _bt_search did * more than we needed. Locate the stack item pointing to our * parent level. */ ilevel = 0; for (;;) { if (stack == NULL) elog(ERROR, "not enough stack items"); if (ilevel == targetlevel) break; stack = stack->bts_parent; ilevel++; } } else { /* * During WAL recovery, we can't use _bt_search (for one reason, * it might invoke user-defined comparison functions that expect * facilities not available in recovery mode). Instead, just set * up a dummy stack pointing to the left end of the parent tree * level, from which _bt_getstackbuf will walk right to the parent * page. Painful, but we don't care too much about performance in * this scenario. */ pbuf = _bt_get_endpoint(rel, targetlevel + 1, false); stack = (BTStack) palloc(sizeof(BTStackData)); stack->bts_blkno = BufferGetBlockNumber(pbuf); stack->bts_offset = InvalidOffsetNumber; /* bts_btentry will be initialized below */ stack->bts_parent = NULL; _bt_relbuf(rel, pbuf); } } /* * We cannot delete a page that is the rightmost child of its immediate * parent, unless it is the only child --- in which case the parent has to * be deleted too, and the same condition applies recursively to it. We * have to check this condition all the way up before trying to delete. We * don't need to re-test when deleting a non-leaf page, though. */ if (targetlevel == 0 && !_bt_parent_deletion_safe(rel, target, stack)) return 0; /* * We have to lock the pages we need to modify in the standard order: * moving right, then up. Else we will deadlock against other writers. * * So, we need to find and write-lock the current left sibling of the * target page. The sibling that was current a moment ago could have * split, so we may have to move right. This search could fail if either * the sibling or the target page was deleted by someone else meanwhile; * if so, give up. (Right now, that should never happen, since page * deletion is only done in VACUUM and there shouldn't be multiple VACUUMs * concurrently on the same table.) */ if (leftsib != P_NONE) { lbuf = _bt_getbuf(rel, leftsib, BT_WRITE); page = BufferGetPage(lbuf); opaque = (BTPageOpaque) PageGetSpecialPointer(page); while (P_ISDELETED(opaque) || opaque->btpo_next != target) { /* step right one page */ leftsib = opaque->btpo_next; _bt_relbuf(rel, lbuf); if (leftsib == P_NONE) { elog(LOG, "no left sibling (concurrent deletion?) in \"%s\"", RelationGetRelationName(rel)); return 0; } lbuf = _bt_getbuf(rel, leftsib, BT_WRITE); page = BufferGetPage(lbuf); opaque = (BTPageOpaque) PageGetSpecialPointer(page); } } else lbuf = InvalidBuffer; /* * Next write-lock the target page itself. It should be okay to take just * a write lock not a superexclusive lock, since no scans would stop on an * empty page. */ buf = _bt_getbuf(rel, target, BT_WRITE); page = BufferGetPage(buf); opaque = (BTPageOpaque) PageGetSpecialPointer(page); /* * Check page is still empty etc, else abandon deletion. The empty check * is necessary since someone else might have inserted into it while we * didn't have it locked; the others are just for paranoia's sake. */ if (P_RIGHTMOST(opaque) || P_ISROOT(opaque) || P_ISDELETED(opaque) || P_FIRSTDATAKEY(opaque) <= PageGetMaxOffsetNumber(page)) { _bt_relbuf(rel, buf); if (BufferIsValid(lbuf)) _bt_relbuf(rel, lbuf); return 0; } if (opaque->btpo_prev != leftsib) elog(ERROR, "left link changed unexpectedly in block %u of index \"%s\"", target, RelationGetRelationName(rel)); /* * And next write-lock the (current) right sibling. */ rightsib = opaque->btpo_next; rbuf = _bt_getbuf(rel, rightsib, BT_WRITE); page = BufferGetPage(rbuf); opaque = (BTPageOpaque) PageGetSpecialPointer(page); if (opaque->btpo_prev != target) elog(ERROR, "right sibling's left-link doesn't match: " "block %u links to %u instead of expected %u in index \"%s\"", rightsib, opaque->btpo_prev, target, RelationGetRelationName(rel)); /* * Next find and write-lock the current parent of the target page. This is * essentially the same as the corresponding step of splitting. */ ItemPointerSet(&(stack->bts_btentry.t_tid), target, P_HIKEY); pbuf = _bt_getstackbuf(rel, stack, BT_WRITE); if (pbuf == InvalidBuffer) elog(ERROR, "failed to re-find parent key in index \"%s\" for deletion target page %u", RelationGetRelationName(rel), target); parent = stack->bts_blkno; poffset = stack->bts_offset; /* * If the target is the rightmost child of its parent, then we can't * delete, unless it's also the only child --- in which case the parent * changes to half-dead status. The "can't delete" case should have been * detected by _bt_parent_deletion_safe, so complain if we see it now. */ page = BufferGetPage(pbuf); opaque = (BTPageOpaque) PageGetSpecialPointer(page); maxoff = PageGetMaxOffsetNumber(page); parent_half_dead = false; parent_one_child = false; if (poffset >= maxoff) { if (poffset == P_FIRSTDATAKEY(opaque)) parent_half_dead = true; else elog(ERROR, "failed to delete rightmost child %u of block %u in index \"%s\"", target, parent, RelationGetRelationName(rel)); } else { /* Will there be exactly one child left in this parent? */ if (OffsetNumberNext(P_FIRSTDATAKEY(opaque)) == maxoff) parent_one_child = true; } /* * If we are deleting the next-to-last page on the target's level, then * the rightsib is a candidate to become the new fast root. (In theory, it * might be possible to push the fast root even further down, but the odds * of doing so are slim, and the locking considerations daunting.) * * We don't support handling this in the case where the parent is becoming * half-dead, even though it theoretically could occur. * * We can safely acquire a lock on the metapage here --- see comments for * _bt_newroot(). */ if (leftsib == P_NONE && !parent_half_dead) { page = BufferGetPage(rbuf); opaque = (BTPageOpaque) PageGetSpecialPointer(page); Assert(opaque->btpo.level == targetlevel); if (P_RIGHTMOST(opaque)) { /* rightsib will be the only one left on the level */ metabuf = _bt_getbuf(rel, BTREE_METAPAGE, BT_WRITE); metapg = BufferGetPage(metabuf); metad = BTPageGetMeta(metapg); /* * The expected case here is btm_fastlevel == targetlevel+1; if * the fastlevel is <= targetlevel, something is wrong, and we * choose to overwrite it to fix it. */ if (metad->btm_fastlevel > targetlevel + 1) { /* no update wanted */ _bt_relbuf(rel, metabuf); metabuf = InvalidBuffer; } } } /* * Check that the parent-page index items we're about to delete/overwrite * contain what we expect. This can fail if the index has become * corrupt for some reason. We want to throw any error before entering * the critical section --- otherwise it'd be a PANIC. * * The test on the target item is just an Assert because _bt_getstackbuf * should have guaranteed it has the expected contents. The test on the * next-child downlink is known to sometimes fail in the field, though. */ page = BufferGetPage(pbuf); opaque = (BTPageOpaque) PageGetSpecialPointer(page); #ifdef USE_ASSERT_CHECKING itemid = PageGetItemId(page, poffset); itup = (IndexTuple) PageGetItem(page, itemid); Assert(ItemPointerGetBlockNumber(&(itup->t_tid)) == target); #endif if (!parent_half_dead) { OffsetNumber nextoffset; nextoffset = OffsetNumberNext(poffset); itemid = PageGetItemId(page, nextoffset); itup = (IndexTuple) PageGetItem(page, itemid); if (ItemPointerGetBlockNumber(&(itup->t_tid)) != rightsib) elog(ERROR, "right sibling %u of block %u is not next child %u of block %u in index \"%s\"", rightsib, target, ItemPointerGetBlockNumber(&(itup->t_tid)), parent, RelationGetRelationName(rel)); } /* * Here we begin doing the deletion. */ /* No ereport(ERROR) until changes are logged */ START_CRIT_SECTION(); /* * Update parent. The normal case is a tad tricky because we want to * delete the target's downlink and the *following* key. Easiest way is * to copy the right sibling's downlink over the target downlink, and then * delete the following item. */ if (parent_half_dead) { PageIndexTupleDelete(page, poffset); opaque->btpo_flags |= BTP_HALF_DEAD; } else { OffsetNumber nextoffset; itemid = PageGetItemId(page, poffset); itup = (IndexTuple) PageGetItem(page, itemid); ItemPointerSet(&(itup->t_tid), rightsib, P_HIKEY); nextoffset = OffsetNumberNext(poffset); PageIndexTupleDelete(page, nextoffset); } /* * Update siblings' side-links. Note the target page's side-links will * continue to point to the siblings. Asserts here are just rechecking * things we already verified above. */ if (BufferIsValid(lbuf)) { page = BufferGetPage(lbuf); opaque = (BTPageOpaque) PageGetSpecialPointer(page); Assert(opaque->btpo_next == target); opaque->btpo_next = rightsib; } page = BufferGetPage(rbuf); opaque = (BTPageOpaque) PageGetSpecialPointer(page); Assert(opaque->btpo_prev == target); opaque->btpo_prev = leftsib; rightsib_empty = (P_FIRSTDATAKEY(opaque) > PageGetMaxOffsetNumber(page)); /* * Mark the page itself deleted. It can be recycled when all current * transactions are gone; or immediately if we're doing VACUUM FULL. */ page = BufferGetPage(buf); opaque = (BTPageOpaque) PageGetSpecialPointer(page); opaque->btpo_flags &= ~BTP_HALF_DEAD; opaque->btpo_flags |= BTP_DELETED; opaque->btpo.xact = vacuum_full ? FrozenTransactionId : ReadNewTransactionId(); /* And update the metapage, if needed */ if (BufferIsValid(metabuf)) { metad->btm_fastroot = rightsib; metad->btm_fastlevel = targetlevel; MarkBufferDirty(metabuf); } /* Must mark buffers dirty before XLogInsert */ MarkBufferDirty(pbuf); MarkBufferDirty(rbuf); MarkBufferDirty(buf); if (BufferIsValid(lbuf)) MarkBufferDirty(lbuf); /* XLOG stuff */ if (!rel->rd_istemp) { xl_btree_delete_page xlrec; xl_btree_metadata xlmeta; uint8 xlinfo; XLogRecPtr recptr; XLogRecData rdata[5]; XLogRecData *nextrdata; xl_btreetid_set(&(xlrec.target), rel, parent, poffset); xlrec.deadblk = target; xlrec.leftblk = leftsib; xlrec.rightblk = rightsib; rdata[0].data = (char *) &xlrec; rdata[0].len = SizeOfBtreeDeletePage; rdata[0].buffer = InvalidBuffer; rdata[0].next = nextrdata = &(rdata[1]); if (BufferIsValid(metabuf)) { xlmeta.root = metad->btm_root; xlmeta.level = metad->btm_level; xlmeta.fastroot = metad->btm_fastroot; xlmeta.fastlevel = metad->btm_fastlevel; nextrdata->data = (char *) &xlmeta; nextrdata->len = sizeof(xl_btree_metadata); nextrdata->buffer = InvalidBuffer; nextrdata->next = nextrdata + 1; nextrdata++; xlinfo = XLOG_BTREE_DELETE_PAGE_META; } else if (parent_half_dead) xlinfo = XLOG_BTREE_DELETE_PAGE_HALF; else xlinfo = XLOG_BTREE_DELETE_PAGE; nextrdata->data = NULL; nextrdata->len = 0; nextrdata->next = nextrdata + 1; nextrdata->buffer = pbuf; nextrdata->buffer_std = true; nextrdata++; nextrdata->data = NULL; nextrdata->len = 0; nextrdata->buffer = rbuf; nextrdata->buffer_std = true; nextrdata->next = NULL; if (BufferIsValid(lbuf)) { nextrdata->next = nextrdata + 1; nextrdata++; nextrdata->data = NULL; nextrdata->len = 0; nextrdata->buffer = lbuf; nextrdata->buffer_std = true; nextrdata->next = NULL; } recptr = XLogInsert(RM_BTREE_ID, xlinfo, rdata); if (BufferIsValid(metabuf)) { PageSetLSN(metapg, recptr); PageSetTLI(metapg, ThisTimeLineID); } page = BufferGetPage(pbuf); PageSetLSN(page, recptr); PageSetTLI(page, ThisTimeLineID); page = BufferGetPage(rbuf); PageSetLSN(page, recptr); PageSetTLI(page, ThisTimeLineID); page = BufferGetPage(buf); PageSetLSN(page, recptr); PageSetTLI(page, ThisTimeLineID); if (BufferIsValid(lbuf)) { page = BufferGetPage(lbuf); PageSetLSN(page, recptr); PageSetTLI(page, ThisTimeLineID); } } END_CRIT_SECTION(); /* release metapage; send out relcache inval if metapage changed */ if (BufferIsValid(metabuf)) { CacheInvalidateRelcache(rel); _bt_relbuf(rel, metabuf); } /* can always release leftsib immediately */ if (BufferIsValid(lbuf)) _bt_relbuf(rel, lbuf); /* * If parent became half dead, recurse to delete it. Otherwise, if right * sibling is empty and is now the last child of the parent, recurse to * try to delete it. (These cases cannot apply at the same time, though * the second case might itself recurse to the first.) * * When recursing to parent, we hold the lock on the target page until * done. This delays any insertions into the keyspace that was just * effectively reassigned to the parent's right sibling. If we allowed * that, and there were enough such insertions before we finish deleting * the parent, page splits within that keyspace could lead to inserting * out-of-order keys into the grandparent level. It is thought that that * wouldn't have any serious consequences, but it still seems like a * pretty bad idea. */ if (parent_half_dead) { /* recursive call will release pbuf */ _bt_relbuf(rel, rbuf); result = _bt_pagedel(rel, pbuf, stack->bts_parent, vacuum_full) + 1; _bt_relbuf(rel, buf); } else if (parent_one_child && rightsib_empty) { _bt_relbuf(rel, pbuf); _bt_relbuf(rel, buf); /* recursive call will release rbuf */ result = _bt_pagedel(rel, rbuf, stack, vacuum_full) + 1; } else { _bt_relbuf(rel, pbuf); _bt_relbuf(rel, buf); _bt_relbuf(rel, rbuf); result = 1; } return result; }
/*---------- * GetSnapshotData -- returns information about running transactions. * * The returned snapshot includes xmin (lowest still-running xact ID), * xmax (next xact ID to be assigned), and a list of running xact IDs * in the range xmin <= xid < xmax. It is used as follows: * All xact IDs < xmin are considered finished. * All xact IDs >= xmax are considered still running. * For an xact ID xmin <= xid < xmax, consult list to see whether * it is considered running or not. * This ensures that the set of transactions seen as "running" by the * current xact will not change after it takes the snapshot. * * We also compute the current global xmin (oldest xmin across all running * transactions) and save it in RecentGlobalXmin. This is the same * computation done by GetOldestXmin(TRUE). The xmin value is also stored * into RecentXmin. *---------- */ Snapshot GetSnapshotData(Snapshot snapshot, bool serializable) { SISeg *segP = shmInvalBuffer; ProcState *stateP = segP->procState; TransactionId xmin; TransactionId xmax; TransactionId globalxmin; int index; int count = 0; Assert(snapshot != NULL); /* * Allocating space for MaxBackends xids is usually overkill; * lastBackend would be sufficient. But it seems better to do the * malloc while not holding the lock, so we can't look at lastBackend. * * This does open a possibility for avoiding repeated malloc/free: * since MaxBackends does not change at runtime, we can simply reuse * the previous xip array if any. (This relies on the fact that all * calls pass static SnapshotData structs.) */ if (snapshot->xip == NULL) { /* * First call for this snapshot */ snapshot->xip = (TransactionId *) malloc(MaxBackends * sizeof(TransactionId)); if (snapshot->xip == NULL) ereport(ERROR, (errcode(ERRCODE_OUT_OF_MEMORY), errmsg("out of memory"))); } globalxmin = xmin = GetCurrentTransactionId(); /* * If we are going to set MyProc->xmin then we'd better get exclusive * lock; if not, this is a read-only operation so it can be shared. */ LWLockAcquire(SInvalLock, serializable ? LW_EXCLUSIVE : LW_SHARED); /*-------------------- * Unfortunately, we have to call ReadNewTransactionId() after acquiring * SInvalLock above. It's not good because ReadNewTransactionId() does * LWLockAcquire(XidGenLock), but *necessary*. We need to be sure that * no transactions exit the set of currently-running transactions * between the time we fetch xmax and the time we finish building our * snapshot. Otherwise we could have a situation like this: * * 1. Tx Old is running (in Read Committed mode). * 2. Tx S reads new transaction ID into xmax, then * is swapped out before acquiring SInvalLock. * 3. Tx New gets new transaction ID (>= S' xmax), * makes changes and commits. * 4. Tx Old changes some row R changed by Tx New and commits. * 5. Tx S finishes getting its snapshot data. It sees Tx Old as * done, but sees Tx New as still running (since New >= xmax). * * Now S will see R changed by both Tx Old and Tx New, *but* does not * see other changes made by Tx New. If S is supposed to be in * Serializable mode, this is wrong. * * By locking SInvalLock before we read xmax, we ensure that TX Old * cannot exit the set of running transactions seen by Tx S. Therefore * both Old and New will be seen as still running => no inconsistency. *-------------------- */ xmax = ReadNewTransactionId(); for (index = 0; index < segP->lastBackend; index++) { SHMEM_OFFSET pOffset = stateP[index].procStruct; if (pOffset != INVALID_OFFSET) { PGPROC *proc = (PGPROC *) MAKE_PTR(pOffset); /* Fetch xid just once - see GetNewTransactionId */ TransactionId xid = proc->xid; /* * Ignore my own proc (dealt with my xid above), procs not * running a transaction, and xacts started since we read the * next transaction ID. There's no need to store XIDs above * what we got from ReadNewTransactionId, since we'll treat * them as running anyway. We also assume that such xacts * can't compute an xmin older than ours, so they needn't be * considered in computing globalxmin. */ if (proc == MyProc || !TransactionIdIsNormal(xid) || TransactionIdFollowsOrEquals(xid, xmax)) continue; if (TransactionIdPrecedes(xid, xmin)) xmin = xid; snapshot->xip[count] = xid; count++; /* Update globalxmin to be the smallest valid xmin */ xid = proc->xmin; if (TransactionIdIsNormal(xid)) if (TransactionIdPrecedes(xid, globalxmin)) globalxmin = xid; } } if (serializable) MyProc->xmin = xmin; LWLockRelease(SInvalLock); /* Serializable snapshot must be computed before any other... */ Assert(TransactionIdIsValid(MyProc->xmin)); /* * Update globalxmin to include actual process xids. This is a * slightly different way of computing it than GetOldestXmin uses, but * should give the same result. */ if (TransactionIdPrecedes(xmin, globalxmin)) globalxmin = xmin; /* Update globals for use by VACUUM */ RecentGlobalXmin = globalxmin; RecentXmin = xmin; snapshot->xmin = xmin; snapshot->xmax = xmax; snapshot->xcnt = count; snapshot->curcid = GetCurrentCommandId(); return snapshot; }
/* * AutoVacMain */ NON_EXEC_STATIC void AutoVacMain(int argc, char *argv[]) { ListCell *cell; List *dblist; autovac_dbase *db; TransactionId xidForceLimit; bool for_xid_wrap; sigjmp_buf local_sigjmp_buf; /* we are a postmaster subprocess now */ IsUnderPostmaster = true; am_autovacuum = true; /* MPP-4990: Autovacuum always runs as utility-mode */ Gp_role = GP_ROLE_UTILITY; /* reset MyProcPid */ MyProcPid = getpid(); /* record Start Time for logging */ MyStartTime = time(NULL); /* Identify myself via ps */ init_ps_display("autovacuum process", "", "", ""); SetProcessingMode(InitProcessing); /* * If possible, make this process a group leader, so that the postmaster * can signal any child processes too. (autovacuum probably never has * any child processes, but for consistency we make all postmaster * child processes do this.) */ #ifdef HAVE_SETSID if (setsid() < 0) elog(FATAL, "setsid() failed: %m"); #endif /* * Set up signal handlers. We operate on databases much like a regular * backend, so we use the same signal handling. See equivalent code in * tcop/postgres.c. * * Currently, we don't pay attention to postgresql.conf changes that * happen during a single daemon iteration, so we can ignore SIGHUP. */ pqsignal(SIGHUP, SIG_IGN); /* * SIGINT is used to signal cancelling the current table's vacuum; SIGTERM * means abort and exit cleanly, and SIGQUIT means abandon ship. */ pqsignal(SIGINT, StatementCancelHandler); pqsignal(SIGTERM, die); pqsignal(SIGQUIT, quickdie); pqsignal(SIGALRM, handle_sig_alarm); pqsignal(SIGPIPE, SIG_IGN); pqsignal(SIGUSR1, procsignal_sigusr1_handler); /* We don't listen for async notifies */ pqsignal(SIGUSR2, SIG_IGN); pqsignal(SIGFPE, FloatExceptionHandler); pqsignal(SIGCHLD, SIG_DFL); /* Early initialization */ BaseInit(); /* * Create a per-backend PGPROC struct in shared memory, except in the * EXEC_BACKEND case where this was done in SubPostmasterMain. We must do * this before we can use LWLocks (and in the EXEC_BACKEND case we already * had to do some stuff with LWLocks). */ #ifndef EXEC_BACKEND InitProcess(); #endif /* * If an exception is encountered, processing resumes here. * * See notes in postgres.c about the design of this coding. */ if (sigsetjmp(local_sigjmp_buf, 1) != 0) { /* Prevents interrupts while cleaning up */ HOLD_INTERRUPTS(); /* Report the error to the server log */ EmitErrorReport(); /* * We can now go away. Note that because we called InitProcess, a * callback was registered to do ProcKill, which will clean up * necessary state. */ proc_exit(0); } /* We can now handle ereport(ERROR) */ PG_exception_stack = &local_sigjmp_buf; PG_SETMASK(&UnBlockSig); /* * Force zero_damaged_pages OFF in the autovac process, even if it is set * in postgresql.conf. We don't really want such a dangerous option being * applied non-interactively. */ SetConfigOption("zero_damaged_pages", "false", PGC_SUSET, PGC_S_OVERRIDE); /* Get a list of databases */ dblist = autovac_get_database_list(); /* * Determine the oldest datfrozenxid/relfrozenxid that we will allow * to pass without forcing a vacuum. (This limit can be tightened for * particular tables, but not loosened.) */ recentXid = ReadNewTransactionId(); xidForceLimit = recentXid - autovacuum_freeze_max_age; /* ensure it's a "normal" XID, else TransactionIdPrecedes misbehaves */ if (xidForceLimit < FirstNormalTransactionId) xidForceLimit -= FirstNormalTransactionId; /* * Choose a database to connect to. We pick the database that was least * recently auto-vacuumed, or one that needs vacuuming to prevent Xid * wraparound-related data loss. If any db at risk of wraparound is * found, we pick the one with oldest datfrozenxid, * independently of autovacuum times. * * Note that a database with no stats entry is not considered, except for * Xid wraparound purposes. The theory is that if no one has ever * connected to it since the stats were last initialized, it doesn't need * vacuuming. * * XXX This could be improved if we had more info about whether it needs * vacuuming before connecting to it. Perhaps look through the pgstats * data for the database's tables? One idea is to keep track of the * number of new and dead tuples per database in pgstats. However it * isn't clear how to construct a metric that measures that and not cause * starvation for less busy databases. */ db = NULL; for_xid_wrap = false; foreach(cell, dblist) { autovac_dbase *tmp = lfirst(cell); /* Find pgstat entry if any */ tmp->entry = pgstat_fetch_stat_dbentry(tmp->oid); /* Check to see if this one is at risk of wraparound */ if (TransactionIdPrecedes(tmp->frozenxid, xidForceLimit)) { if (db == NULL || TransactionIdPrecedes(tmp->frozenxid, db->frozenxid)) db = tmp; for_xid_wrap = true; continue; } else if (for_xid_wrap) continue; /* ignore not-at-risk DBs */ /* * Otherwise, skip a database with no pgstat entry; it means it * hasn't seen any activity. */ if (!tmp->entry) continue; /* * Remember the db with oldest autovac time. (If we are here, * both tmp->entry and db->entry must be non-null.) */ if (db == NULL || tmp->entry->last_autovac_time < db->entry->last_autovac_time) db = tmp; }
/* * _bt_pagedel() -- Delete a page from the b-tree. * * This action unlinks the page from the b-tree structure, removing all * pointers leading to it --- but not touching its own left and right links. * The page cannot be physically reclaimed right away, since other processes * may currently be trying to follow links leading to the page; they have to * be allowed to use its right-link to recover. See nbtree/README. * * On entry, the target buffer must be pinned and read-locked. This lock and * pin will be dropped before exiting. * * Returns the number of pages successfully deleted (zero on failure; could * be more than one if parent blocks were deleted). * * NOTE: this leaks memory. Rather than trying to clean up everything * carefully, it's better to run it in a temp context that can be reset * frequently. */ int _bt_pagedel(Relation rel, Buffer buf, bool vacuum_full) { BlockNumber target, leftsib, rightsib, parent; OffsetNumber poffset, maxoff; uint32 targetlevel, ilevel; ItemId itemid; BTItem targetkey, btitem; ScanKey itup_scankey; BTStack stack; Buffer lbuf, rbuf, pbuf; bool parent_half_dead; bool parent_one_child; bool rightsib_empty; Buffer metabuf = InvalidBuffer; Page metapg = NULL; BTMetaPageData *metad = NULL; Page page; BTPageOpaque opaque; /* * We can never delete rightmost pages nor root pages. While at it, check * that page is not already deleted and is empty. */ page = BufferGetPage(buf); opaque = (BTPageOpaque) PageGetSpecialPointer(page); if (P_RIGHTMOST(opaque) || P_ISROOT(opaque) || P_ISDELETED(opaque) || P_FIRSTDATAKEY(opaque) <= PageGetMaxOffsetNumber(page)) { _bt_relbuf(rel, buf); return 0; } /* * Save info about page, including a copy of its high key (it must have * one, being non-rightmost). */ target = BufferGetBlockNumber(buf); targetlevel = opaque->btpo.level; leftsib = opaque->btpo_prev; itemid = PageGetItemId(page, P_HIKEY); targetkey = CopyBTItem((BTItem) PageGetItem(page, itemid)); /* * We need to get an approximate pointer to the page's parent page. Use * the standard search mechanism to search for the page's high key; this * will give us a link to either the current parent or someplace to its * left (if there are multiple equal high keys). To avoid deadlocks, we'd * better drop the target page lock first. */ _bt_relbuf(rel, buf); /* we need a scan key to do our search, so build one */ itup_scankey = _bt_mkscankey(rel, &(targetkey->bti_itup)); /* find the leftmost leaf page containing this key */ stack = _bt_search(rel, rel->rd_rel->relnatts, itup_scankey, false, &lbuf, BT_READ); /* don't need a pin on that either */ _bt_relbuf(rel, lbuf); /* * If we are trying to delete an interior page, _bt_search did more than * we needed. Locate the stack item pointing to our parent level. */ ilevel = 0; for (;;) { if (stack == NULL) elog(ERROR, "not enough stack items"); if (ilevel == targetlevel) break; stack = stack->bts_parent; ilevel++; } /* * We have to lock the pages we need to modify in the standard order: * moving right, then up. Else we will deadlock against other writers. * * So, we need to find and write-lock the current left sibling of the * target page. The sibling that was current a moment ago could have * split, so we may have to move right. This search could fail if either * the sibling or the target page was deleted by someone else meanwhile; * if so, give up. (Right now, that should never happen, since page * deletion is only done in VACUUM and there shouldn't be multiple VACUUMs * concurrently on the same table.) */ if (leftsib != P_NONE) { lbuf = _bt_getbuf(rel, leftsib, BT_WRITE); page = BufferGetPage(lbuf); opaque = (BTPageOpaque) PageGetSpecialPointer(page); while (P_ISDELETED(opaque) || opaque->btpo_next != target) { /* step right one page */ leftsib = opaque->btpo_next; _bt_relbuf(rel, lbuf); if (leftsib == P_NONE) { elog(LOG, "no left sibling (concurrent deletion?) in \"%s\"", RelationGetRelationName(rel)); return 0; } lbuf = _bt_getbuf(rel, leftsib, BT_WRITE); page = BufferGetPage(lbuf); opaque = (BTPageOpaque) PageGetSpecialPointer(page); } } else lbuf = InvalidBuffer; /* * Next write-lock the target page itself. It should be okay to take just * a write lock not a superexclusive lock, since no scans would stop on an * empty page. */ buf = _bt_getbuf(rel, target, BT_WRITE); page = BufferGetPage(buf); opaque = (BTPageOpaque) PageGetSpecialPointer(page); /* * Check page is still empty etc, else abandon deletion. The empty check * is necessary since someone else might have inserted into it while we * didn't have it locked; the others are just for paranoia's sake. */ if (P_RIGHTMOST(opaque) || P_ISROOT(opaque) || P_ISDELETED(opaque) || P_FIRSTDATAKEY(opaque) <= PageGetMaxOffsetNumber(page)) { _bt_relbuf(rel, buf); if (BufferIsValid(lbuf)) _bt_relbuf(rel, lbuf); return 0; } if (opaque->btpo_prev != leftsib) elog(ERROR, "left link changed unexpectedly in block %u of \"%s\"", target, RelationGetRelationName(rel)); /* * And next write-lock the (current) right sibling. */ rightsib = opaque->btpo_next; rbuf = _bt_getbuf(rel, rightsib, BT_WRITE); /* * Next find and write-lock the current parent of the target page. This is * essentially the same as the corresponding step of splitting. However, * it's possible for the search to fail (for reasons explained in README). * If that happens, we recover by searching the whole parent level, which * is a tad inefficient but doesn't happen often enough to be a problem. */ ItemPointerSet(&(stack->bts_btitem.bti_itup.t_tid), target, P_HIKEY); pbuf = _bt_getstackbuf(rel, stack, BT_WRITE); if (pbuf == InvalidBuffer) { /* Find the leftmost page in the parent level */ pbuf = _bt_get_endpoint(rel, opaque->btpo.level + 1, false); stack->bts_blkno = BufferGetBlockNumber(pbuf); stack->bts_offset = InvalidOffsetNumber; _bt_relbuf(rel, pbuf); /* and repeat search from there */ pbuf = _bt_getstackbuf(rel, stack, BT_WRITE); if (pbuf == InvalidBuffer) elog(ERROR, "failed to re-find parent key in \"%s\" for deletion target page %u", RelationGetRelationName(rel), target); } parent = stack->bts_blkno; poffset = stack->bts_offset; /* * If the target is the rightmost child of its parent, then we can't * delete, unless it's also the only child --- in which case the parent * changes to half-dead status. */ page = BufferGetPage(pbuf); opaque = (BTPageOpaque) PageGetSpecialPointer(page); maxoff = PageGetMaxOffsetNumber(page); parent_half_dead = false; parent_one_child = false; if (poffset >= maxoff) { if (poffset == P_FIRSTDATAKEY(opaque)) parent_half_dead = true; else { _bt_relbuf(rel, pbuf); _bt_relbuf(rel, rbuf); _bt_relbuf(rel, buf); if (BufferIsValid(lbuf)) _bt_relbuf(rel, lbuf); return 0; } } else { /* Will there be exactly one child left in this parent? */ if (OffsetNumberNext(P_FIRSTDATAKEY(opaque)) == maxoff) parent_one_child = true; } /* * If we are deleting the next-to-last page on the target's level, then * the rightsib is a candidate to become the new fast root. (In theory, it * might be possible to push the fast root even further down, but the odds * of doing so are slim, and the locking considerations daunting.) * * We can safely acquire a lock on the metapage here --- see comments for * _bt_newroot(). */ if (leftsib == P_NONE) { page = BufferGetPage(rbuf); opaque = (BTPageOpaque) PageGetSpecialPointer(page); Assert(opaque->btpo.level == targetlevel); if (P_RIGHTMOST(opaque)) { /* rightsib will be the only one left on the level */ metabuf = _bt_getbuf(rel, BTREE_METAPAGE, BT_WRITE); metapg = BufferGetPage(metabuf); metad = BTPageGetMeta(metapg); /* * The expected case here is btm_fastlevel == targetlevel+1; if * the fastlevel is <= targetlevel, something is wrong, and we * choose to overwrite it to fix it. */ if (metad->btm_fastlevel > targetlevel + 1) { /* no update wanted */ _bt_relbuf(rel, metabuf); metabuf = InvalidBuffer; } } } /* * Here we begin doing the deletion. */ /* No ereport(ERROR) until changes are logged */ START_CRIT_SECTION(); /* * Update parent. The normal case is a tad tricky because we want to * delete the target's downlink and the *following* key. Easiest way is * to copy the right sibling's downlink over the target downlink, and then * delete the following item. */ page = BufferGetPage(pbuf); opaque = (BTPageOpaque) PageGetSpecialPointer(page); if (parent_half_dead) { PageIndexTupleDelete(page, poffset); opaque->btpo_flags |= BTP_HALF_DEAD; } else { OffsetNumber nextoffset; itemid = PageGetItemId(page, poffset); btitem = (BTItem) PageGetItem(page, itemid); Assert(ItemPointerGetBlockNumber(&(btitem->bti_itup.t_tid)) == target); ItemPointerSet(&(btitem->bti_itup.t_tid), rightsib, P_HIKEY); nextoffset = OffsetNumberNext(poffset); /* This part is just for double-checking */ itemid = PageGetItemId(page, nextoffset); btitem = (BTItem) PageGetItem(page, itemid); if (ItemPointerGetBlockNumber(&(btitem->bti_itup.t_tid)) != rightsib) elog(PANIC, "right sibling is not next child in \"%s\"", RelationGetRelationName(rel)); PageIndexTupleDelete(page, nextoffset); } /* * Update siblings' side-links. Note the target page's side-links will * continue to point to the siblings. */ if (BufferIsValid(lbuf)) { page = BufferGetPage(lbuf); opaque = (BTPageOpaque) PageGetSpecialPointer(page); Assert(opaque->btpo_next == target); opaque->btpo_next = rightsib; } page = BufferGetPage(rbuf); opaque = (BTPageOpaque) PageGetSpecialPointer(page); Assert(opaque->btpo_prev == target); opaque->btpo_prev = leftsib; rightsib_empty = (P_FIRSTDATAKEY(opaque) > PageGetMaxOffsetNumber(page)); /* * Mark the page itself deleted. It can be recycled when all current * transactions are gone; or immediately if we're doing VACUUM FULL. */ page = BufferGetPage(buf); opaque = (BTPageOpaque) PageGetSpecialPointer(page); opaque->btpo_flags |= BTP_DELETED; opaque->btpo.xact = vacuum_full ? FrozenTransactionId : ReadNewTransactionId(); /* And update the metapage, if needed */ if (BufferIsValid(metabuf)) { metad->btm_fastroot = rightsib; metad->btm_fastlevel = targetlevel; } /* XLOG stuff */ if (!rel->rd_istemp) { xl_btree_delete_page xlrec; xl_btree_metadata xlmeta; uint8 xlinfo; XLogRecPtr recptr; XLogRecData rdata[5]; XLogRecData *nextrdata; xlrec.target.node = rel->rd_node; ItemPointerSet(&(xlrec.target.tid), parent, poffset); xlrec.deadblk = target; xlrec.leftblk = leftsib; xlrec.rightblk = rightsib; rdata[0].data = (char *) &xlrec; rdata[0].len = SizeOfBtreeDeletePage; rdata[0].buffer = InvalidBuffer; rdata[0].next = nextrdata = &(rdata[1]); if (BufferIsValid(metabuf)) { xlmeta.root = metad->btm_root; xlmeta.level = metad->btm_level; xlmeta.fastroot = metad->btm_fastroot; xlmeta.fastlevel = metad->btm_fastlevel; nextrdata->data = (char *) &xlmeta; nextrdata->len = sizeof(xl_btree_metadata); nextrdata->buffer = InvalidBuffer; nextrdata->next = nextrdata + 1; nextrdata++; xlinfo = XLOG_BTREE_DELETE_PAGE_META; } else xlinfo = XLOG_BTREE_DELETE_PAGE; nextrdata->data = NULL; nextrdata->len = 0; nextrdata->next = nextrdata + 1; nextrdata->buffer = pbuf; nextrdata->buffer_std = true; nextrdata++; nextrdata->data = NULL; nextrdata->len = 0; nextrdata->buffer = rbuf; nextrdata->buffer_std = true; nextrdata->next = NULL; if (BufferIsValid(lbuf)) { nextrdata->next = nextrdata + 1; nextrdata++; nextrdata->data = NULL; nextrdata->len = 0; nextrdata->buffer = lbuf; nextrdata->buffer_std = true; nextrdata->next = NULL; } recptr = XLogInsert(RM_BTREE_ID, xlinfo, rdata); if (BufferIsValid(metabuf)) { PageSetLSN(metapg, recptr); PageSetTLI(metapg, ThisTimeLineID); } page = BufferGetPage(pbuf); PageSetLSN(page, recptr); PageSetTLI(page, ThisTimeLineID); page = BufferGetPage(rbuf); PageSetLSN(page, recptr); PageSetTLI(page, ThisTimeLineID); page = BufferGetPage(buf); PageSetLSN(page, recptr); PageSetTLI(page, ThisTimeLineID); if (BufferIsValid(lbuf)) { page = BufferGetPage(lbuf); PageSetLSN(page, recptr); PageSetTLI(page, ThisTimeLineID); } } END_CRIT_SECTION(); /* Write and release buffers */ if (BufferIsValid(metabuf)) _bt_wrtbuf(rel, metabuf); _bt_wrtbuf(rel, pbuf); _bt_wrtbuf(rel, rbuf); _bt_wrtbuf(rel, buf); if (BufferIsValid(lbuf)) _bt_wrtbuf(rel, lbuf); /* * If parent became half dead, recurse to try to delete it. Otherwise, if * right sibling is empty and is now the last child of the parent, recurse * to try to delete it. (These cases cannot apply at the same time, * though the second case might itself recurse to the first.) */ if (parent_half_dead) { buf = _bt_getbuf(rel, parent, BT_READ); return _bt_pagedel(rel, buf, vacuum_full) + 1; } if (parent_one_child && rightsib_empty) { buf = _bt_getbuf(rel, rightsib, BT_READ); return _bt_pagedel(rel, buf, vacuum_full) + 1; } return 1; }