VictoriaMetrics/lib/storage/storage.go

package storage

import (
	"fmt"
	"io"
	"io/ioutil"
	"math"
	"os"
	"path/filepath"
	"regexp"
	"runtime"
	"sort"
	"sync"
	"sync/atomic"
	"time"
	"unsafe"

	"github.com/VictoriaMetrics/VictoriaMetrics/lib/encoding"
	"github.com/VictoriaMetrics/VictoriaMetrics/lib/fasttime"
	"github.com/VictoriaMetrics/VictoriaMetrics/lib/fs"
	"github.com/VictoriaMetrics/VictoriaMetrics/lib/logger"
	"github.com/VictoriaMetrics/VictoriaMetrics/lib/memory"
	"github.com/VictoriaMetrics/VictoriaMetrics/lib/storagepacelimiter"
	"github.com/VictoriaMetrics/VictoriaMetrics/lib/timerpool"
	"github.com/VictoriaMetrics/VictoriaMetrics/lib/uint64set"
	"github.com/VictoriaMetrics/VictoriaMetrics/lib/workingsetcache"
	"github.com/VictoriaMetrics/fastcache"
)

const (
	msecsPerMonth     = 31 * 24 * 3600 * 1000
	maxRetentionMsecs = 100 * 12 * msecsPerMonth
)

// Storage represents TSDB storage.
type Storage struct {
	// Atomic counters must go at the top of the structure in order to properly align by 8 bytes on 32-bit archs.
	// See https://github.com/VictoriaMetrics/VictoriaMetrics/issues/212 .
	tooSmallTimestampRows uint64
	tooBigTimestampRows   uint64

	addRowsConcurrencyLimitReached uint64
	addRowsConcurrencyLimitTimeout uint64
	addRowsConcurrencyDroppedRows  uint64

	searchTSIDsConcurrencyLimitReached uint64
	searchTSIDsConcurrencyLimitTimeout uint64

	slowRowInserts         uint64
	slowPerDayIndexInserts uint64
	slowMetricNameLoads    uint64

	path           string
	cachePath      string
	retentionMsecs int64

	// lock file for exclusive access to the storage on the given path.
	flockF *os.File

	idbCurr atomic.Value

	tb *table

	// tsidCache is MetricName -> TSID cache.
	tsidCache *workingsetcache.Cache

	// metricIDCache is MetricID -> TSID cache.
	metricIDCache *workingsetcache.Cache

	// metricNameCache is MetricID -> MetricName cache.
	metricNameCache *workingsetcache.Cache

	// dateMetricIDCache is (Date, MetricID) cache.
	dateMetricIDCache *dateMetricIDCache

	// Fast cache for MetricID values occurred during the current hour.
	currHourMetricIDs atomic.Value

	// Fast cache for MetricID values occurred during the previous hour.
	prevHourMetricIDs atomic.Value

	// Fast cache for pre-populating per-day inverted index for the next day.
	// This is needed in order to remove CPU usage spikes at 00:00 UTC
	// due to creation of per-day inverted index for active time series.
	// See https://github.com/VictoriaMetrics/VictoriaMetrics/issues/430 for details.
	nextDayMetricIDs atomic.Value

	// Pending MetricID values to be added to currHourMetricIDs.
	pendingHourEntriesLock sync.Mutex
	pendingHourEntries     *uint64set.Set

	// Pending MetricIDs to be added to nextDayMetricIDs.
	pendingNextDayMetricIDsLock sync.Mutex
	pendingNextDayMetricIDs     *uint64set.Set

	// metricIDs for pre-fetched metricNames in the prefetchMetricNames function.
	prefetchedMetricIDs atomic.Value

	stop chan struct{}

	currHourMetricIDsUpdaterWG sync.WaitGroup
	nextDayMetricIDsUpdaterWG  sync.WaitGroup
	retentionWatcherWG         sync.WaitGroup

	// The snapshotLock prevents from concurrent creation of snapshots,
	// since this may result in snapshots without recently added data,
	// which may be in the process of flushing to disk by concurrently running
	// snapshot process.
	snapshotLock sync.Mutex
}

// OpenStorage opens storage on the given path with the given retentionMsecs.
func OpenStorage(path string, retentionMsecs int64) (*Storage, error) {
	path, err := filepath.Abs(path)
	if err != nil {
		return nil, fmt.Errorf("cannot determine absolute path for %q: %w", path, err)
	}
	if retentionMsecs <= 0 {
		retentionMsecs = maxRetentionMsecs
	}
	s := &Storage{
		path:           path,
		cachePath:      path + "/cache",
		retentionMsecs: retentionMsecs,

		stop: make(chan struct{}),
	}

	if err := fs.MkdirAllIfNotExist(path); err != nil {
		return nil, fmt.Errorf("cannot create a directory for the storage at %q: %w", path, err)
	}
	snapshotsPath := path + "/snapshots"
	if err := fs.MkdirAllIfNotExist(snapshotsPath); err != nil {
		return nil, fmt.Errorf("cannot create %q: %w", snapshotsPath, err)
	}

	// Protect from concurrent opens.
	flockF, err := fs.CreateFlockFile(path)
	if err != nil {
		return nil, err
	}
	s.flockF = flockF

	// Load caches.
	mem := memory.Allowed()
	s.tsidCache = s.mustLoadCache("MetricName->TSID", "metricName_tsid", mem/3)
	s.metricIDCache = s.mustLoadCache("MetricID->TSID", "metricID_tsid", mem/16)
	s.metricNameCache = s.mustLoadCache("MetricID->MetricName", "metricID_metricName", mem/8)
	s.dateMetricIDCache = newDateMetricIDCache()

	hour := fasttime.UnixHour()
	hmCurr := s.mustLoadHourMetricIDs(hour, "curr_hour_metric_ids")
	hmPrev := s.mustLoadHourMetricIDs(hour-1, "prev_hour_metric_ids")
	s.currHourMetricIDs.Store(hmCurr)
	s.prevHourMetricIDs.Store(hmPrev)
	s.pendingHourEntries = &uint64set.Set{}

	date := fasttime.UnixDate()
	nextDayMetricIDs := s.mustLoadNextDayMetricIDs(date)
	s.nextDayMetricIDs.Store(nextDayMetricIDs)
	s.pendingNextDayMetricIDs = &uint64set.Set{}

	s.prefetchedMetricIDs.Store(&uint64set.Set{})

	// Load indexdb
	idbPath := path + "/indexdb"
	idbSnapshotsPath := idbPath + "/snapshots"
	if err := fs.MkdirAllIfNotExist(idbSnapshotsPath); err != nil {
		return nil, fmt.Errorf("cannot create %q: %w", idbSnapshotsPath, err)
	}
	idbCurr, idbPrev, err := openIndexDBTables(idbPath, s.metricIDCache, s.metricNameCache, s.tsidCache)
	if err != nil {
		return nil, fmt.Errorf("cannot open indexdb tables at %q: %w", idbPath, err)
	}
	idbCurr.SetExtDB(idbPrev)
	s.idbCurr.Store(idbCurr)

	// Load data
	tablePath := path + "/data"
	tb, err := openTable(tablePath, s.getDeletedMetricIDs, retentionMsecs)
	if err != nil {
		s.idb().MustClose()
		return nil, fmt.Errorf("cannot open table at %q: %w", tablePath, err)
	}
	s.tb = tb

	s.startCurrHourMetricIDsUpdater()
	s.startNextDayMetricIDsUpdater()
	s.startRetentionWatcher()

	return s, nil
}

// DebugFlush flushes recently added storage data, so it becomes visible to search.
func (s *Storage) DebugFlush() {
	s.tb.flushRawRows()
	s.idb().tb.DebugFlush()
}

func (s *Storage) getDeletedMetricIDs() *uint64set.Set {
	return s.idb().getDeletedMetricIDs()
}

// CreateSnapshot creates snapshot for s and returns the snapshot name.
func (s *Storage) CreateSnapshot() (string, error) {
	logger.Infof("creating Storage snapshot for %q...", s.path)
	startTime := time.Now()

	s.snapshotLock.Lock()
	defer s.snapshotLock.Unlock()

	snapshotName := fmt.Sprintf("%s-%08X", time.Now().UTC().Format("20060102150405"), nextSnapshotIdx())
	srcDir := s.path
	dstDir := fmt.Sprintf("%s/snapshots/%s", srcDir, snapshotName)
	if err := fs.MkdirAllFailIfExist(dstDir); err != nil {
		return "", fmt.Errorf("cannot create dir %q: %w", dstDir, err)
	}
	dstDataDir := dstDir + "/data"
	if err := fs.MkdirAllFailIfExist(dstDataDir); err != nil {
		return "", fmt.Errorf("cannot create dir %q: %w", dstDataDir, err)
	}

	smallDir, bigDir, err := s.tb.CreateSnapshot(snapshotName)
	if err != nil {
		return "", fmt.Errorf("cannot create table snapshot: %w", err)
	}
	dstSmallDir := dstDataDir + "/small"
	if err := fs.SymlinkRelative(smallDir, dstSmallDir); err != nil {
		return "", fmt.Errorf("cannot create symlink from %q to %q: %w", smallDir, dstSmallDir, err)
	}
	dstBigDir := dstDataDir + "/big"
	if err := fs.SymlinkRelative(bigDir, dstBigDir); err != nil {
		return "", fmt.Errorf("cannot create symlink from %q to %q: %w", bigDir, dstBigDir, err)
	}
	fs.MustSyncPath(dstDataDir)

	idbSnapshot := fmt.Sprintf("%s/indexdb/snapshots/%s", s.path, snapshotName)
	idb := s.idb()
	currSnapshot := idbSnapshot + "/" + idb.name
	if err := idb.tb.CreateSnapshotAt(currSnapshot); err != nil {
		return "", fmt.Errorf("cannot create curr indexDB snapshot: %w", err)
	}
	ok := idb.doExtDB(func(extDB *indexDB) {
		prevSnapshot := idbSnapshot + "/" + extDB.name
		err = extDB.tb.CreateSnapshotAt(prevSnapshot)
	})
	if ok && err != nil {
		return "", fmt.Errorf("cannot create prev indexDB snapshot: %w", err)
	}
	dstIdbDir := dstDir + "/indexdb"
	if err := fs.SymlinkRelative(idbSnapshot, dstIdbDir); err != nil {
		return "", fmt.Errorf("cannot create symlink from %q to %q: %w", idbSnapshot, dstIdbDir, err)
	}

	fs.MustSyncPath(dstDir)
	fs.MustSyncPath(srcDir + "/snapshots")

	logger.Infof("created Storage snapshot for %q at %q in %.3f seconds", srcDir, dstDir, time.Since(startTime).Seconds())
	return snapshotName, nil
}

var snapshotNameRegexp = regexp.MustCompile("^[0-9]{14}-[0-9A-Fa-f]+$")

// ListSnapshots returns sorted list of existing snapshots for s.
func (s *Storage) ListSnapshots() ([]string, error) {
	snapshotsPath := s.path + "/snapshots"
	d, err := os.Open(snapshotsPath)
	if err != nil {
		return nil, fmt.Errorf("cannot open %q: %w", snapshotsPath, err)
	}
	defer fs.MustClose(d)

	fnames, err := d.Readdirnames(-1)
	if err != nil {
		return nil, fmt.Errorf("cannot read contents of %q: %w", snapshotsPath, err)
	}
	snapshotNames := make([]string, 0, len(fnames))
	for _, fname := range fnames {
		if !snapshotNameRegexp.MatchString(fname) {
			continue
		}
		snapshotNames = append(snapshotNames, fname)
	}
	sort.Strings(snapshotNames)
	return snapshotNames, nil
}

// DeleteSnapshot deletes the given snapshot.
func (s *Storage) DeleteSnapshot(snapshotName string) error {
	if !snapshotNameRegexp.MatchString(snapshotName) {
		return fmt.Errorf("invalid snapshotName %q", snapshotName)
	}
	snapshotPath := s.path + "/snapshots/" + snapshotName

	logger.Infof("deleting snapshot %q...", snapshotPath)
	startTime := time.Now()

	s.tb.MustDeleteSnapshot(snapshotName)
	idbPath := fmt.Sprintf("%s/indexdb/snapshots/%s", s.path, snapshotName)
	fs.MustRemoveAll(idbPath)
	fs.MustRemoveAll(snapshotPath)

	logger.Infof("deleted snapshot %q in %.3f seconds", snapshotPath, time.Since(startTime).Seconds())

	return nil
}

var snapshotIdx = uint64(time.Now().UnixNano())

func nextSnapshotIdx() uint64 {
	return atomic.AddUint64(&snapshotIdx, 1)
}

func (s *Storage) idb() *indexDB {
	return s.idbCurr.Load().(*indexDB)
}

// Metrics contains essential metrics for the Storage.
type Metrics struct {
	RowsAddedTotal    uint64
	DedupsDuringMerge uint64

	TooSmallTimestampRows uint64
	TooBigTimestampRows   uint64

	AddRowsConcurrencyLimitReached uint64
	AddRowsConcurrencyLimitTimeout uint64
	AddRowsConcurrencyDroppedRows  uint64
	AddRowsConcurrencyCapacity     uint64
	AddRowsConcurrencyCurrent      uint64

	SearchTSIDsConcurrencyLimitReached uint64
	SearchTSIDsConcurrencyLimitTimeout uint64
	SearchTSIDsConcurrencyCapacity     uint64
	SearchTSIDsConcurrencyCurrent      uint64

	SearchDelays uint64

	SlowRowInserts         uint64
	SlowPerDayIndexInserts uint64
	SlowMetricNameLoads    uint64

	TimestampsBlocksMerged uint64
	TimestampsBytesSaved   uint64

	TSIDCacheSize       uint64
	TSIDCacheSizeBytes  uint64
	TSIDCacheRequests   uint64
	TSIDCacheMisses     uint64
	TSIDCacheCollisions uint64

	MetricIDCacheSize       uint64
	MetricIDCacheSizeBytes  uint64
	MetricIDCacheRequests   uint64
	MetricIDCacheMisses     uint64
	MetricIDCacheCollisions uint64

	MetricNameCacheSize       uint64
	MetricNameCacheSizeBytes  uint64
	MetricNameCacheRequests   uint64
	MetricNameCacheMisses     uint64
	MetricNameCacheCollisions uint64

	DateMetricIDCacheSize        uint64
	DateMetricIDCacheSizeBytes   uint64
	DateMetricIDCacheSyncsCount  uint64
	DateMetricIDCacheResetsCount uint64

	HourMetricIDCacheSize      uint64
	HourMetricIDCacheSizeBytes uint64

	NextDayMetricIDCacheSize      uint64
	NextDayMetricIDCacheSizeBytes uint64

	PrefetchedMetricIDsSize      uint64
	PrefetchedMetricIDsSizeBytes uint64

	IndexDBMetrics IndexDBMetrics
	TableMetrics   TableMetrics
}

// Reset resets m.
func (m *Metrics) Reset() {
	*m = Metrics{}
}

// UpdateMetrics updates m with metrics from s.
func (s *Storage) UpdateMetrics(m *Metrics) {
	m.RowsAddedTotal = atomic.LoadUint64(&rowsAddedTotal)
	m.DedupsDuringMerge = atomic.LoadUint64(&dedupsDuringMerge)

	m.TooSmallTimestampRows += atomic.LoadUint64(&s.tooSmallTimestampRows)
	m.TooBigTimestampRows += atomic.LoadUint64(&s.tooBigTimestampRows)

	m.AddRowsConcurrencyLimitReached += atomic.LoadUint64(&s.addRowsConcurrencyLimitReached)
	m.AddRowsConcurrencyLimitTimeout += atomic.LoadUint64(&s.addRowsConcurrencyLimitTimeout)
	m.AddRowsConcurrencyDroppedRows += atomic.LoadUint64(&s.addRowsConcurrencyDroppedRows)
	m.AddRowsConcurrencyCapacity = uint64(cap(addRowsConcurrencyCh))
	m.AddRowsConcurrencyCurrent = uint64(len(addRowsConcurrencyCh))

	m.SearchTSIDsConcurrencyLimitReached += atomic.LoadUint64(&s.searchTSIDsConcurrencyLimitReached)
	m.SearchTSIDsConcurrencyLimitTimeout += atomic.LoadUint64(&s.searchTSIDsConcurrencyLimitTimeout)
	m.SearchTSIDsConcurrencyCapacity = uint64(cap(searchTSIDsConcurrencyCh))
	m.SearchTSIDsConcurrencyCurrent = uint64(len(searchTSIDsConcurrencyCh))

	m.SearchDelays = storagepacelimiter.Search.DelaysTotal()

	m.SlowRowInserts += atomic.LoadUint64(&s.slowRowInserts)
	m.SlowPerDayIndexInserts += atomic.LoadUint64(&s.slowPerDayIndexInserts)
	m.SlowMetricNameLoads += atomic.LoadUint64(&s.slowMetricNameLoads)

	m.TimestampsBlocksMerged = atomic.LoadUint64(&timestampsBlocksMerged)
	m.TimestampsBytesSaved = atomic.LoadUint64(&timestampsBytesSaved)

	var cs fastcache.Stats
	s.tsidCache.UpdateStats(&cs)
	m.TSIDCacheSize += cs.EntriesCount
	m.TSIDCacheSizeBytes += cs.BytesSize
	m.TSIDCacheRequests += cs.GetCalls
	m.TSIDCacheMisses += cs.Misses
	m.TSIDCacheCollisions += cs.Collisions

	cs.Reset()
	s.metricIDCache.UpdateStats(&cs)
	m.MetricIDCacheSize += cs.EntriesCount
	m.MetricIDCacheSizeBytes += cs.BytesSize
	m.MetricIDCacheRequests += cs.GetCalls
	m.MetricIDCacheMisses += cs.Misses
	m.MetricIDCacheCollisions += cs.Collisions

	cs.Reset()
	s.metricNameCache.UpdateStats(&cs)
	m.MetricNameCacheSize += cs.EntriesCount
	m.MetricNameCacheSizeBytes += cs.BytesSize
	m.MetricNameCacheRequests += cs.GetCalls
	m.MetricNameCacheMisses += cs.Misses
	m.MetricNameCacheCollisions += cs.Collisions

	m.DateMetricIDCacheSize += uint64(s.dateMetricIDCache.EntriesCount())
	m.DateMetricIDCacheSizeBytes += uint64(s.dateMetricIDCache.SizeBytes())
	m.DateMetricIDCacheSyncsCount += atomic.LoadUint64(&s.dateMetricIDCache.syncsCount)
	m.DateMetricIDCacheResetsCount += atomic.LoadUint64(&s.dateMetricIDCache.resetsCount)

	hmCurr := s.currHourMetricIDs.Load().(*hourMetricIDs)
	hmPrev := s.prevHourMetricIDs.Load().(*hourMetricIDs)
	hourMetricIDsLen := hmPrev.m.Len()
	if hmCurr.m.Len() > hourMetricIDsLen {
		hourMetricIDsLen = hmCurr.m.Len()
	}
	m.HourMetricIDCacheSize += uint64(hourMetricIDsLen)
	m.HourMetricIDCacheSizeBytes += hmCurr.m.SizeBytes()
	m.HourMetricIDCacheSizeBytes += hmPrev.m.SizeBytes()

	nextDayMetricIDs := &s.nextDayMetricIDs.Load().(*byDateMetricIDEntry).v
	m.NextDayMetricIDCacheSize += uint64(nextDayMetricIDs.Len())
	m.NextDayMetricIDCacheSizeBytes += nextDayMetricIDs.SizeBytes()

	prefetchedMetricIDs := s.prefetchedMetricIDs.Load().(*uint64set.Set)
	m.PrefetchedMetricIDsSize += uint64(prefetchedMetricIDs.Len())
	m.PrefetchedMetricIDsSizeBytes += uint64(prefetchedMetricIDs.SizeBytes())

	s.idb().UpdateMetrics(&m.IndexDBMetrics)
	s.tb.UpdateMetrics(&m.TableMetrics)
}

func (s *Storage) startRetentionWatcher() {
	s.retentionWatcherWG.Add(1)
	go func() {
		s.retentionWatcher()
		s.retentionWatcherWG.Done()
	}()
}

func (s *Storage) retentionWatcher() {
	retentionMonths := int((s.retentionMsecs + (msecsPerMonth - 1)) / msecsPerMonth)
	for {
		d := nextRetentionDuration(retentionMonths)
		select {
		case <-s.stop:
			return
		case <-time.After(d):
			s.mustRotateIndexDB()
		}
	}
}

func (s *Storage) startCurrHourMetricIDsUpdater() {
	s.currHourMetricIDsUpdaterWG.Add(1)
	go func() {
		s.currHourMetricIDsUpdater()
		s.currHourMetricIDsUpdaterWG.Done()
	}()
}

func (s *Storage) startNextDayMetricIDsUpdater() {
	s.nextDayMetricIDsUpdaterWG.Add(1)
	go func() {
		s.nextDayMetricIDsUpdater()
		s.nextDayMetricIDsUpdaterWG.Done()
	}()
}

var currHourMetricIDsUpdateInterval = time.Second * 10

func (s *Storage) currHourMetricIDsUpdater() {
	ticker := time.NewTicker(currHourMetricIDsUpdateInterval)
	defer ticker.Stop()
	for {
		select {
		case <-s.stop:
			s.updateCurrHourMetricIDs()
			return
		case <-ticker.C:
			s.updateCurrHourMetricIDs()
		}
	}
}

var nextDayMetricIDsUpdateInterval = time.Second * 11

func (s *Storage) nextDayMetricIDsUpdater() {
	ticker := time.NewTicker(nextDayMetricIDsUpdateInterval)
	defer ticker.Stop()
	for {
		select {
		case <-s.stop:
			s.updateNextDayMetricIDs()
			return
		case <-ticker.C:
			s.updateNextDayMetricIDs()
		}
	}
}

func (s *Storage) mustRotateIndexDB() {
	// Create new indexdb table.
	newTableName := nextIndexDBTableName()
	idbNewPath := s.path + "/indexdb/" + newTableName
	idbNew, err := openIndexDB(idbNewPath, s.metricIDCache, s.metricNameCache, s.tsidCache)
	if err != nil {
		logger.Panicf("FATAL: cannot create new indexDB at %q: %s", idbNewPath, err)
	}

	// Drop extDB
	idbCurr := s.idb()
	idbCurr.doExtDB(func(extDB *indexDB) {
		extDB.scheduleToDrop()
	})
	idbCurr.SetExtDB(nil)

	// Start using idbNew
	idbNew.SetExtDB(idbCurr)
	s.idbCurr.Store(idbNew)

	// Persist changes on the file system.
	fs.MustSyncPath(s.path)

	// Flush tsidCache, so idbNew can be populated with fresh data.
	s.tsidCache.Reset()

	// Flush dateMetricIDCache, so idbNew can be populated with fresh data.
	s.dateMetricIDCache.Reset()

	// Do not flush metricIDCache and metricNameCache, since all the metricIDs
	// from prev idb remain valid after the rotation.

	// There is no need in resetting nextDayMetricIDs, since it should be automatically reset every day.
}

// MustClose closes the storage.
func (s *Storage) MustClose() {
	close(s.stop)

	s.retentionWatcherWG.Wait()
	s.currHourMetricIDsUpdaterWG.Wait()
	s.nextDayMetricIDsUpdaterWG.Wait()

	s.tb.MustClose()
	s.idb().MustClose()

	// Save caches.
	s.mustSaveAndStopCache(s.tsidCache, "MetricName->TSID", "metricName_tsid")
	s.mustSaveAndStopCache(s.metricIDCache, "MetricID->TSID", "metricID_tsid")
	s.mustSaveAndStopCache(s.metricNameCache, "MetricID->MetricName", "metricID_metricName")

	hmCurr := s.currHourMetricIDs.Load().(*hourMetricIDs)
	s.mustSaveHourMetricIDs(hmCurr, "curr_hour_metric_ids")
	hmPrev := s.prevHourMetricIDs.Load().(*hourMetricIDs)
	s.mustSaveHourMetricIDs(hmPrev, "prev_hour_metric_ids")

	nextDayMetricIDs := s.nextDayMetricIDs.Load().(*byDateMetricIDEntry)
	s.mustSaveNextDayMetricIDs(nextDayMetricIDs)

	// Release lock file.
	if err := s.flockF.Close(); err != nil {
		logger.Panicf("FATAL: cannot close lock file %q: %s", s.flockF.Name(), err)
	}
}

func (s *Storage) mustLoadNextDayMetricIDs(date uint64) *byDateMetricIDEntry {
	e := &byDateMetricIDEntry{
		date: date,
	}
	name := "next_day_metric_ids"
	path := s.cachePath + "/" + name
	logger.Infof("loading %s from %q...", name, path)
	startTime := time.Now()
	if !fs.IsPathExist(path) {
		logger.Infof("nothing to load from %q", path)
		return e
	}
	src, err := ioutil.ReadFile(path)
	if err != nil {
		logger.Panicf("FATAL: cannot read %s: %s", path, err)
	}
	srcOrigLen := len(src)
	if len(src) < 16 {
		logger.Errorf("discarding %s, since it has broken header; got %d bytes; want %d bytes", path, len(src), 16)
		return e
	}

	// Unmarshal header
	dateLoaded := encoding.UnmarshalUint64(src)
	src = src[8:]
	if dateLoaded != date {
		logger.Infof("discarding %s, since it contains data for stale date; got %d; want %d", path, dateLoaded, date)
		return e
	}

	// Unmarshal uint64set
	m, tail, err := unmarshalUint64Set(src)
	if err != nil {
		logger.Infof("discarding %s because cannot load uint64set: %s", path, err)
		return e
	}
	if len(tail) > 0 {
		logger.Infof("discarding %s because non-empty tail left; len(tail)=%d", path, len(tail))
		return e
	}
	e.v = *m
	logger.Infof("loaded %s from %q in %.3f seconds; entriesCount: %d; sizeBytes: %d", name, path, time.Since(startTime).Seconds(), m.Len(), srcOrigLen)
	return e
}

func (s *Storage) mustLoadHourMetricIDs(hour uint64, name string) *hourMetricIDs {
	hm := &hourMetricIDs{
		hour: hour,
	}
	path := s.cachePath + "/" + name
	logger.Infof("loading %s from %q...", name, path)
	startTime := time.Now()
	if !fs.IsPathExist(path) {
		logger.Infof("nothing to load from %q", path)
		return hm
	}
	src, err := ioutil.ReadFile(path)
	if err != nil {
		logger.Panicf("FATAL: cannot read %s: %s", path, err)
	}
	srcOrigLen := len(src)
	if len(src) < 24 {
		logger.Errorf("discarding %s, since it has broken header; got %d bytes; want %d bytes", path, len(src), 24)
		return hm
	}

	// Unmarshal header
	isFull := encoding.UnmarshalUint64(src)
	src = src[8:]
	hourLoaded := encoding.UnmarshalUint64(src)
	src = src[8:]
	if hourLoaded != hour {
		logger.Infof("discarding %s, since it contains outdated hour; got %d; want %d", path, hourLoaded, hour)
		return hm
	}

	// Unmarshal uint64set
	m, tail, err := unmarshalUint64Set(src)
	if err != nil {
		logger.Infof("discarding %s because cannot load uint64set: %s", path, err)
		return hm
	}
	if len(tail) > 0 {
		logger.Infof("discarding %s because non-empty tail left; len(tail)=%d", path, len(tail))
		return hm
	}
	hm.m = m
	hm.isFull = isFull != 0
	logger.Infof("loaded %s from %q in %.3f seconds; entriesCount: %d; sizeBytes: %d", name, path, time.Since(startTime).Seconds(), m.Len(), srcOrigLen)
	return hm
}

func (s *Storage) mustSaveNextDayMetricIDs(e *byDateMetricIDEntry) {
	name := "next_day_metric_ids"
	path := s.cachePath + "/" + name
	logger.Infof("saving %s to %q...", name, path)
	startTime := time.Now()
	dst := make([]byte, 0, e.v.Len()*8+16)

	// Marshal header
	dst = encoding.MarshalUint64(dst, e.date)

	// Marshal e.v
	dst = marshalUint64Set(dst, &e.v)

	if err := ioutil.WriteFile(path, dst, 0644); err != nil {
		logger.Panicf("FATAL: cannot write %d bytes to %q: %s", len(dst), path, err)
	}
	logger.Infof("saved %s to %q in %.3f seconds; entriesCount: %d; sizeBytes: %d", name, path, time.Since(startTime).Seconds(), e.v.Len(), len(dst))
}

func (s *Storage) mustSaveHourMetricIDs(hm *hourMetricIDs, name string) {
	path := s.cachePath + "/" + name
	logger.Infof("saving %s to %q...", name, path)
	startTime := time.Now()
	dst := make([]byte, 0, hm.m.Len()*8+24)
	isFull := uint64(0)
	if hm.isFull {
		isFull = 1
	}

	// Marshal header
	dst = encoding.MarshalUint64(dst, isFull)
	dst = encoding.MarshalUint64(dst, hm.hour)

	// Marshal hm.m
	dst = marshalUint64Set(dst, hm.m)

	if err := ioutil.WriteFile(path, dst, 0644); err != nil {
		logger.Panicf("FATAL: cannot write %d bytes to %q: %s", len(dst), path, err)
	}
	logger.Infof("saved %s to %q in %.3f seconds; entriesCount: %d; sizeBytes: %d", name, path, time.Since(startTime).Seconds(), hm.m.Len(), len(dst))
}

func unmarshalUint64Set(src []byte) (*uint64set.Set, []byte, error) {
	mLen := encoding.UnmarshalUint64(src)
	src = src[8:]
	if uint64(len(src)) < 8*mLen {
		return nil, nil, fmt.Errorf("cannot unmarshal uint64set; got %d bytes; want at least %d bytes", len(src), 8*mLen)
	}
	m := &uint64set.Set{}
	for i := uint64(0); i < mLen; i++ {
		metricID := encoding.UnmarshalUint64(src)
		src = src[8:]
		m.Add(metricID)
	}
	return m, src, nil
}

func marshalUint64Set(dst []byte, m *uint64set.Set) []byte {
	dst = encoding.MarshalUint64(dst, uint64(m.Len()))
	m.ForEach(func(part []uint64) bool {
		for _, metricID := range part {
			dst = encoding.MarshalUint64(dst, metricID)
		}
		return true
	})
	return dst
}

func (s *Storage) mustLoadCache(info, name string, sizeBytes int) *workingsetcache.Cache {
	path := s.cachePath + "/" + name
	logger.Infof("loading %s cache from %q...", info, path)
	startTime := time.Now()
	c := workingsetcache.Load(path, sizeBytes, time.Hour)
	var cs fastcache.Stats
	c.UpdateStats(&cs)
	logger.Infof("loaded %s cache from %q in %.3f seconds; entriesCount: %d; sizeBytes: %d",
		info, path, time.Since(startTime).Seconds(), cs.EntriesCount, cs.BytesSize)
	return c
}

func (s *Storage) mustSaveAndStopCache(c *workingsetcache.Cache, info, name string) {
	path := s.cachePath + "/" + name
	logger.Infof("saving %s cache to %q...", info, path)
	startTime := time.Now()
	if err := c.Save(path); err != nil {
		logger.Panicf("FATAL: cannot save %s cache to %q: %s", info, path, err)
	}
	var cs fastcache.Stats
	c.UpdateStats(&cs)
	c.Stop()
	logger.Infof("saved %s cache to %q in %.3f seconds; entriesCount: %d; sizeBytes: %d",
		info, path, time.Since(startTime).Seconds(), cs.EntriesCount, cs.BytesSize)
}

func nextRetentionDuration(retentionMonths int) time.Duration {
	t := time.Now().UTC()
	n := t.Year()*12 + int(t.Month()) - 1 + retentionMonths
	n -= n % retentionMonths
	y := n / 12
	m := time.Month((n % 12) + 1)
	// Schedule the deadline to +4 hours from the next retention period start.
	// This should prevent from possible double deletion of indexdb
	// due to time drift - see https://github.com/VictoriaMetrics/VictoriaMetrics/issues/248 .
	deadline := time.Date(y, m, 1, 4, 0, 0, 0, time.UTC)
	return deadline.Sub(t)
}

// SearchMetricNames returns metric names matching the given tfss on the given tr.
func (s *Storage) SearchMetricNames(tfss []*TagFilters, tr TimeRange, maxMetrics int, deadline uint64) ([]MetricName, error) {
	tsids, err := s.searchTSIDs(tfss, tr, maxMetrics, deadline)
	if err != nil {
		return nil, err
	}
	if err = s.prefetchMetricNames(tsids, deadline); err != nil {
		return nil, err
	}
	idb := s.idb()
	is := idb.getIndexSearch(deadline)
	defer idb.putIndexSearch(is)
	mns := make([]MetricName, 0, len(tsids))
	var metricName []byte
	for i := range tsids {
		metricID := tsids[i].MetricID
		var err error
		metricName, err = is.searchMetricName(metricName[:0], metricID)
		if err != nil {
			if err == io.EOF {
				// Skip missing metricName for metricID.
				// It should be automatically fixed. See indexDB.searchMetricName for details.
				continue
			}
			return nil, fmt.Errorf("error when searching metricName for metricID=%d: %w", metricID, err)
		}
		mns = mns[:len(mns)+1]
		mn := &mns[len(mns)-1]
		if err = mn.Unmarshal(metricName); err != nil {
			return nil, fmt.Errorf("cannot unmarshal metricName=%q: %w", metricName, err)
		}
	}
	return mns, nil
}

// searchTSIDs returns sorted TSIDs for the given tfss and the given tr.
func (s *Storage) searchTSIDs(tfss []*TagFilters, tr TimeRange, maxMetrics int, deadline uint64) ([]TSID, error) {
	// Do not cache tfss -> tsids here, since the caching is performed
	// on idb level.

	// Limit the number of concurrent goroutines that may search TSIDS in the storage.
	// This should prevent from out of memory errors and CPU trashing when too many
	// goroutines call searchTSIDs.
	select {
	case searchTSIDsConcurrencyCh <- struct{}{}:
	default:
		// Sleep for a while until giving up
		atomic.AddUint64(&s.searchTSIDsConcurrencyLimitReached, 1)
		currentTime := fasttime.UnixTimestamp()
		timeoutSecs := uint64(0)
		if currentTime < deadline {
			timeoutSecs = deadline - currentTime
		}
		timeout := time.Second * time.Duration(timeoutSecs)
		t := timerpool.Get(timeout)
		select {
		case searchTSIDsConcurrencyCh <- struct{}{}:
			timerpool.Put(t)
		case <-t.C:
			timerpool.Put(t)
			atomic.AddUint64(&s.searchTSIDsConcurrencyLimitTimeout, 1)
			return nil, fmt.Errorf("cannot search for tsids, since more than %d concurrent searches are performed during %.3f secs; add more CPUs or reduce query load",
				cap(searchTSIDsConcurrencyCh), timeout.Seconds())
		}
	}
	tsids, err := s.idb().searchTSIDs(tfss, tr, maxMetrics, deadline)
	<-searchTSIDsConcurrencyCh
	if err != nil {
		return nil, fmt.Errorf("error when searching tsids: %w", err)
	}
	return tsids, nil
}

var (
	// Limit the concurrency for TSID searches to GOMAXPROCS*2, since this operation
	// is CPU bound and sometimes disk IO bound, so there is no sense in running more
	// than GOMAXPROCS*2 concurrent goroutines for TSID searches.
	searchTSIDsConcurrencyCh = make(chan struct{}, runtime.GOMAXPROCS(-1)*2)
)

// prefetchMetricNames pre-fetches metric names for the given tsids into metricID->metricName cache.
//
// This should speed-up further searchMetricName calls for metricIDs from tsids.
func (s *Storage) prefetchMetricNames(tsids []TSID, deadline uint64) error {
	if len(tsids) == 0 {
		return nil
	}
	var metricIDs uint64Sorter
	prefetchedMetricIDs := s.prefetchedMetricIDs.Load().(*uint64set.Set)
	for i := range tsids {
		tsid := &tsids[i]
		metricID := tsid.MetricID
		if prefetchedMetricIDs.Has(metricID) {
			continue
		}
		metricIDs = append(metricIDs, metricID)
	}
	if len(metricIDs) < 500 {
		// It is cheaper to skip pre-fetching and obtain metricNames inline.
		return nil
	}
	atomic.AddUint64(&s.slowMetricNameLoads, uint64(len(metricIDs)))

	// Pre-fetch metricIDs.
	sort.Sort(metricIDs)
	var metricName []byte
	var err error
	idb := s.idb()
	is := idb.getIndexSearch(deadline)
	defer idb.putIndexSearch(is)
	for loops, metricID := range metricIDs {
		if loops&paceLimiterSlowIterationsMask == 0 {
			if err := checkSearchDeadlineAndPace(is.deadline); err != nil {
				return err
			}
		}
		metricName, err = is.searchMetricName(metricName[:0], metricID)
		if err != nil && err != io.EOF {
			return fmt.Errorf("error in pre-fetching metricName for metricID=%d: %w", metricID, err)
		}
	}

	// Store the pre-fetched metricIDs, so they aren't pre-fetched next time.

	prefetchedMetricIDsNew := prefetchedMetricIDs.Clone()
	prefetchedMetricIDsNew.AddMulti(metricIDs)
	if prefetchedMetricIDsNew.SizeBytes() > uint64(memory.Allowed())/32 {
		// Reset prefetchedMetricIDsNew if it occupies too much space.
		prefetchedMetricIDsNew = &uint64set.Set{}
	}
	s.prefetchedMetricIDs.Store(prefetchedMetricIDsNew)
	return nil
}

// ErrDeadlineExceeded is returned when the request times out.
var ErrDeadlineExceeded = fmt.Errorf("deadline exceeded")

// DeleteMetrics deletes all the metrics matching the given tfss.
//
// Returns the number of metrics deleted.
func (s *Storage) DeleteMetrics(tfss []*TagFilters) (int, error) {
	deletedCount, err := s.idb().DeleteTSIDs(tfss)
	if err != nil {
		return deletedCount, fmt.Errorf("cannot delete tsids: %w", err)
	}
	// Do not reset MetricName->TSID cache in order to prevent from adding new data points
	// to deleted time series in Storage.add, since it is already reset inside DeleteTSIDs.

	// Do not reset MetricID->MetricName cache, since it must be used only
	// after filtering out deleted metricIDs.

	return deletedCount, nil
}

// searchMetricName appends metric name for the given metricID to dst
// and returns the result.
func (s *Storage) searchMetricName(dst []byte, metricID uint64) ([]byte, error) {
	return s.idb().searchMetricName(dst, metricID)
}

// SearchTagKeysOnTimeRange searches for tag keys on tr.
func (s *Storage) SearchTagKeysOnTimeRange(tr TimeRange, maxTagKeys int, deadline uint64) ([]string, error) {
	return s.idb().SearchTagKeysOnTimeRange(tr, maxTagKeys, deadline)
}

// SearchTagKeys searches for tag keys
func (s *Storage) SearchTagKeys(maxTagKeys int, deadline uint64) ([]string, error) {
	return s.idb().SearchTagKeys(maxTagKeys, deadline)
}

// SearchTagValuesOnTimeRange searches for tag values for the given tagKey on tr.
func (s *Storage) SearchTagValuesOnTimeRange(tagKey []byte, tr TimeRange, maxTagValues int, deadline uint64) ([]string, error) {
	return s.idb().SearchTagValuesOnTimeRange(tagKey, tr, maxTagValues, deadline)
}

// SearchTagValues searches for tag values for the given tagKey
func (s *Storage) SearchTagValues(tagKey []byte, maxTagValues int, deadline uint64) ([]string, error) {
	return s.idb().SearchTagValues(tagKey, maxTagValues, deadline)
}

// SearchTagValueSuffixes returns all the tag value suffixes for the given tagKey and tagValuePrefix on the given tr.
//
// This allows implementing https://graphite-api.readthedocs.io/en/latest/api.html#metrics-find or similar APIs.
func (s *Storage) SearchTagValueSuffixes(tr TimeRange, tagKey, tagValuePrefix []byte, delimiter byte, maxTagValueSuffixes int, deadline uint64) ([]string, error) {
	return s.idb().SearchTagValueSuffixes(tr, tagKey, tagValuePrefix, delimiter, maxTagValueSuffixes, deadline)
}

// SearchTagEntries returns a list of (tagName -> tagValues)
func (s *Storage) SearchTagEntries(maxTagKeys, maxTagValues int, deadline uint64) ([]TagEntry, error) {
	idb := s.idb()
	keys, err := idb.SearchTagKeys(maxTagKeys, deadline)
	if err != nil {
		return nil, fmt.Errorf("cannot search tag keys: %w", err)
	}

	// Sort keys for faster seeks below
	sort.Strings(keys)

	tes := make([]TagEntry, len(keys))
	for i, key := range keys {
		values, err := idb.SearchTagValues([]byte(key), maxTagValues, deadline)
		if err != nil {
			return nil, fmt.Errorf("cannot search values for tag %q: %w", key, err)
		}
		te := &tes[i]
		te.Key = key
		te.Values = values
	}
	return tes, nil
}

// TagEntry contains (tagName -> tagValues) mapping
type TagEntry struct {
	// Key is tagName
	Key string

	// Values contains all the values for Key.
	Values []string
}

// GetSeriesCount returns the approximate number of unique time series.
//
// It includes the deleted series too and may count the same series
// up to two times - in db and extDB.
func (s *Storage) GetSeriesCount(deadline uint64) (uint64, error) {
	return s.idb().GetSeriesCount(deadline)
}

// GetTSDBStatusForDate returns TSDB status data for /api/v1/status/tsdb.
//
// See https://prometheus.io/docs/prometheus/latest/querying/api/#tsdb-stats
func (s *Storage) GetTSDBStatusForDate(date uint64, topN int, deadline uint64) (*TSDBStatus, error) {
	return s.idb().GetTSDBStatusForDate(date, topN, deadline)
}

// MetricRow is a metric to insert into storage.
type MetricRow struct {
	// MetricNameRaw contains raw metric name, which must be decoded
	// with MetricName.unmarshalRaw.
	MetricNameRaw []byte

	Timestamp int64
	Value     float64
}

// CopyFrom copies src to mr.
func (mr *MetricRow) CopyFrom(src *MetricRow) {
	mr.MetricNameRaw = append(mr.MetricNameRaw[:0], src.MetricNameRaw...)
	mr.Timestamp = src.Timestamp
	mr.Value = src.Value
}

// String returns string representation of the mr.
func (mr *MetricRow) String() string {
	metricName := string(mr.MetricNameRaw)
	var mn MetricName
	if err := mn.unmarshalRaw(mr.MetricNameRaw); err == nil {
		metricName = mn.String()
	}
	return fmt.Sprintf("MetricName=%s, Timestamp=%d, Value=%f\n", metricName, mr.Timestamp, mr.Value)
}

// Marshal appends marshaled mr to dst and returns the result.
func (mr *MetricRow) Marshal(dst []byte) []byte {
	dst = encoding.MarshalBytes(dst, mr.MetricNameRaw)
	dst = encoding.MarshalUint64(dst, uint64(mr.Timestamp))
	dst = encoding.MarshalUint64(dst, math.Float64bits(mr.Value))
	return dst
}

// Unmarshal unmarshals mr from src and returns the remaining tail from src.
func (mr *MetricRow) Unmarshal(src []byte) ([]byte, error) {
	tail, metricNameRaw, err := encoding.UnmarshalBytes(src)
	if err != nil {
		return tail, fmt.Errorf("cannot unmarshal MetricName: %w", err)
	}
	mr.MetricNameRaw = append(mr.MetricNameRaw[:0], metricNameRaw...)

	if len(tail) < 8 {
		return tail, fmt.Errorf("cannot unmarshal Timestamp: want %d bytes; have %d bytes", 8, len(tail))
	}
	timestamp := encoding.UnmarshalUint64(tail)
	mr.Timestamp = int64(timestamp)
	tail = tail[8:]

	if len(tail) < 8 {
		return tail, fmt.Errorf("cannot unmarshal Value: want %d bytes; have %d bytes", 8, len(tail))
	}
	value := encoding.UnmarshalUint64(tail)
	mr.Value = math.Float64frombits(value)
	tail = tail[8:]

	return tail, nil
}

// ForceMergePartitions force-merges partitions in s with names starting from the given partitionNamePrefix.
//
// Partitions are merged sequentially in order to reduce load on the system.
func (s *Storage) ForceMergePartitions(partitionNamePrefix string) error {
	return s.tb.ForceMergePartitions(partitionNamePrefix)
}

var rowsAddedTotal uint64

// AddRows adds the given mrs to s.
func (s *Storage) AddRows(mrs []MetricRow, precisionBits uint8) error {
	if len(mrs) == 0 {
		return nil
	}

	// Limit the number of concurrent goroutines that may add rows to the storage.
	// This should prevent from out of memory errors and CPU trashing when too many
	// goroutines call AddRows.
	select {
	case addRowsConcurrencyCh <- struct{}{}:
	default:
		// Sleep for a while until giving up
		atomic.AddUint64(&s.addRowsConcurrencyLimitReached, 1)
		t := timerpool.Get(addRowsTimeout)

		// Prioritize data ingestion over concurrent searches.
		storagepacelimiter.Search.Inc()

		select {
		case addRowsConcurrencyCh <- struct{}{}:
			timerpool.Put(t)
			storagepacelimiter.Search.Dec()
		case <-t.C:
			timerpool.Put(t)
			storagepacelimiter.Search.Dec()
			atomic.AddUint64(&s.addRowsConcurrencyLimitTimeout, 1)
			atomic.AddUint64(&s.addRowsConcurrencyDroppedRows, uint64(len(mrs)))
			return fmt.Errorf("cannot add %d rows to storage in %s, since it is overloaded with %d concurrent writers; add more CPUs or reduce load",
				len(mrs), addRowsTimeout, cap(addRowsConcurrencyCh))
		}
	}

	// Add rows to the storage.
	var err error
	rr := getRawRowsWithSize(len(mrs))
	rr.rows, err = s.add(rr.rows, mrs, precisionBits)
	putRawRows(rr)

	<-addRowsConcurrencyCh

	atomic.AddUint64(&rowsAddedTotal, uint64(len(mrs)))
	return err
}

var (
	// Limit the concurrency for data ingestion to GOMAXPROCS, since this operation
	// is CPU bound, so there is no sense in running more than GOMAXPROCS concurrent
	// goroutines on data ingestion path.
	addRowsConcurrencyCh = make(chan struct{}, runtime.GOMAXPROCS(-1))
	addRowsTimeout       = 30 * time.Second
)

// RegisterMetricNames registers all the metric names from mns in the indexdb, so they can be queried later.
//
// The the MetricRow.Timestamp is used for registering the metric name starting from the given timestamp.
// Th MetricRow.Value field is ignored.
func (s *Storage) RegisterMetricNames(mrs []MetricRow) error {
	var (
		tsid       TSID
		mn         MetricName
		metricName []byte
	)
	idb := s.idb()
	is := idb.getIndexSearch(noDeadline)
	defer idb.putIndexSearch(is)
	for i := range mrs {
		mr := &mrs[i]
		if s.getTSIDFromCache(&tsid, mr.MetricNameRaw) {
			// Fast path - mr.MetricNameRaw has been already registered.
			continue
		}

		// Slow path - register mr.MetricNameRaw.
		if err := mn.unmarshalRaw(mr.MetricNameRaw); err != nil {
			return fmt.Errorf("cannot register the metric because cannot unmarshal MetricNameRaw %q: %w", mr.MetricNameRaw, err)
		}
		mn.sortTags()
		metricName = mn.Marshal(metricName[:0])
		if err := is.GetOrCreateTSIDByName(&tsid, metricName); err != nil {
			return fmt.Errorf("cannot register the metric because cannot create TSID for metricName %q: %w", metricName, err)
		}
		s.putTSIDToCache(&tsid, mr.MetricNameRaw)

		// Register the metric in per-day inverted index.
		date := uint64(mr.Timestamp) / msecPerDay
		metricID := tsid.MetricID
		if s.dateMetricIDCache.Has(date, metricID) {
			// Fast path: the metric has been already registered in per-day inverted index
			continue
		}

		// Slow path: acutally register the metric in per-day inverted index.
		ok, err := is.hasDateMetricID(date, metricID)
		if err != nil {
			return fmt.Errorf("cannot register the metric in per-date inverted index because of error when locating (date=%d, metricID=%d) in database: %w",
				date, metricID, err)
		}
		if !ok {
			// The (date, metricID) entry is missing in the indexDB. Add it there.
			if err := is.storeDateMetricID(date, metricID); err != nil {
				return fmt.Errorf("cannot register the metric in per-date inverted index because of error when storing (date=%d, metricID=%d) in database: %w",
					date, metricID, err)
			}
		}
		// The metric must be added to cache only after it has been successfully added to indexDB.
		s.dateMetricIDCache.Set(date, metricID)
	}
	return nil
}

func (s *Storage) add(rows []rawRow, mrs []MetricRow, precisionBits uint8) ([]rawRow, error) {
	idb := s.idb()
	rowsLen := len(rows)
	if n := rowsLen + len(mrs) - cap(rows); n > 0 {
		rows = append(rows[:cap(rows)], make([]rawRow, n)...)
	}
	rows = rows[:rowsLen+len(mrs)]
	j := 0
	var (
		// These vars are used for speeding up bulk imports of multiple adjancent rows for the same metricName.
		prevTSID          TSID
		prevMetricNameRaw []byte
	)
	var pmrs *pendingMetricRows
	minTimestamp, maxTimestamp := s.tb.getMinMaxTimestamps()
	// Return only the first error, since it has no sense in returning all errors.
	var firstWarn error
	for i := range mrs {
		mr := &mrs[i]
		if math.IsNaN(mr.Value) {
			// Just skip NaNs, since the underlying encoding
			// doesn't know how to work with them.
			continue
		}
		if mr.Timestamp < minTimestamp {
			// Skip rows with too small timestamps outside the retention.
			if firstWarn == nil {
				firstWarn = fmt.Errorf("cannot insert row with too small timestamp %d outside the retention; minimum allowed timestamp is %d; "+
					"probably you need updating -retentionPeriod command-line flag",
					mr.Timestamp, minTimestamp)
			}
			atomic.AddUint64(&s.tooSmallTimestampRows, 1)
			continue
		}
		if mr.Timestamp > maxTimestamp {
			// Skip rows with too big timestamps significantly exceeding the current time.
			if firstWarn == nil {
				firstWarn = fmt.Errorf("cannot insert row with too big timestamp %d exceeding the current time; maximum allowd timestamp is %d; "+
					"propbably you need updating -retentionPeriod command-line flag",
					mr.Timestamp, maxTimestamp)
			}
			atomic.AddUint64(&s.tooBigTimestampRows, 1)
			continue
		}
		r := &rows[rowsLen+j]
		j++
		r.Timestamp = mr.Timestamp
		r.Value = mr.Value
		r.PrecisionBits = precisionBits
		if string(mr.MetricNameRaw) == string(prevMetricNameRaw) {
			// Fast path - the current mr contains the same metric name as the previous mr, so it contains the same TSID.
			// This path should trigger on bulk imports when many rows contain the same MetricNameRaw.
			r.TSID = prevTSID
			continue
		}
		if s.getTSIDFromCache(&r.TSID, mr.MetricNameRaw) {
			// Fast path - the TSID for the given MetricName has been found in cache and isn't deleted.
			// There is no need in checking whether r.TSID.MetricID is deleted, since tsidCache doesn't
			// contain MetricName->TSID entries for deleted time series.
			// See Storage.DeleteMetrics code for details.
			prevTSID = r.TSID
			prevMetricNameRaw = mr.MetricNameRaw
			continue
		}

		// Slow path - the TSID is missing in the cache.
		// Postpone its search in the loop below.
		j--
		if pmrs == nil {
			pmrs = getPendingMetricRows()
		}
		if err := pmrs.addRow(mr); err != nil {
			// Do not stop adding rows on error - just skip invalid row.
			// This guarantees that invalid rows don't prevent
			// from adding valid rows into the storage.
			if firstWarn == nil {
				firstWarn = err
			}
			continue
		}
	}
	if pmrs != nil {
		// Sort pendingMetricRows by canonical metric name in order to speed up search via `is` in the loop below.
		pendingMetricRows := pmrs.pmrs
		sort.Slice(pendingMetricRows, func(i, j int) bool {
			return string(pendingMetricRows[i].MetricName) < string(pendingMetricRows[j].MetricName)
		})
		is := idb.getIndexSearch(noDeadline)
		prevMetricNameRaw = nil
		var slowInsertsCount uint64
		for i := range pendingMetricRows {
			pmr := &pendingMetricRows[i]
			mr := &pmr.mr
			r := &rows[rowsLen+j]
			j++
			r.Timestamp = mr.Timestamp
			r.Value = mr.Value
			r.PrecisionBits = precisionBits
			if string(mr.MetricNameRaw) == string(prevMetricNameRaw) {
				// Fast path - the current mr contains the same metric name as the previous mr, so it contains the same TSID.
				// This path should trigger on bulk imports when many rows contain the same MetricNameRaw.
				r.TSID = prevTSID
				continue
			}
			if s.getTSIDFromCache(&r.TSID, mr.MetricNameRaw) {
				// Fast path - the TSID for the given MetricName has been found in cache and isn't deleted.
				// There is no need in checking whether r.TSID.MetricID is deleted, since tsidCache doesn't
				// contain MetricName->TSID entries for deleted time series.
				// See Storage.DeleteMetrics code for details.
				prevTSID = r.TSID
				prevMetricNameRaw = mr.MetricNameRaw
				continue
			}
			slowInsertsCount++
			if err := is.GetOrCreateTSIDByName(&r.TSID, pmr.MetricName); err != nil {
				// Do not stop adding rows on error - just skip invalid row.
				// This guarantees that invalid rows don't prevent
				// from adding valid rows into the storage.
				if firstWarn == nil {
					firstWarn = fmt.Errorf("cannot obtain or create TSID for MetricName %q: %w", pmr.MetricName, err)
				}
				j--
				continue
			}
			s.putTSIDToCache(&r.TSID, mr.MetricNameRaw)
		}
		idb.putIndexSearch(is)
		putPendingMetricRows(pmrs)
		atomic.AddUint64(&s.slowRowInserts, slowInsertsCount)
	}
	if firstWarn != nil {
		logger.Errorf("warn occurred during rows addition: %s", firstWarn)
	}
	rows = rows[:rowsLen+j]

	var firstError error
	if err := s.tb.AddRows(rows); err != nil {
		firstError = fmt.Errorf("cannot add rows to table: %w", err)
	}
	if err := s.updatePerDateData(rows); err != nil && firstError == nil {
		firstError = fmt.Errorf("cannot update per-date data: %w", err)
	}
	if firstError != nil {
		return rows, fmt.Errorf("error occurred during rows addition: %w", firstError)
	}
	return rows, nil
}

type pendingMetricRow struct {
	MetricName []byte
	mr         MetricRow
}

type pendingMetricRows struct {
	pmrs           []pendingMetricRow
	metricNamesBuf []byte

	lastMetricNameRaw []byte
	lastMetricName    []byte
	mn                MetricName
}

func (pmrs *pendingMetricRows) reset() {
	for _, pmr := range pmrs.pmrs {
		pmr.MetricName = nil
		pmr.mr.MetricNameRaw = nil
	}
	pmrs.pmrs = pmrs.pmrs[:0]
	pmrs.metricNamesBuf = pmrs.metricNamesBuf[:0]
	pmrs.lastMetricNameRaw = nil
	pmrs.lastMetricName = nil
	pmrs.mn.Reset()
}

func (pmrs *pendingMetricRows) addRow(mr *MetricRow) error {
	// Do not spend CPU time on re-calculating canonical metricName during bulk import
	// of many rows for the same metric.
	if string(mr.MetricNameRaw) != string(pmrs.lastMetricNameRaw) {
		if err := pmrs.mn.unmarshalRaw(mr.MetricNameRaw); err != nil {
			return fmt.Errorf("cannot unmarshal MetricNameRaw %q: %w", mr.MetricNameRaw, err)
		}
		pmrs.mn.sortTags()
		metricNamesBufLen := len(pmrs.metricNamesBuf)
		pmrs.metricNamesBuf = pmrs.mn.Marshal(pmrs.metricNamesBuf)
		pmrs.lastMetricName = pmrs.metricNamesBuf[metricNamesBufLen:]
		pmrs.lastMetricNameRaw = mr.MetricNameRaw
	}
	pmrs.pmrs = append(pmrs.pmrs, pendingMetricRow{
		MetricName: pmrs.lastMetricName,
		mr:         *mr,
	})
	return nil
}

func getPendingMetricRows() *pendingMetricRows {
	v := pendingMetricRowsPool.Get()
	if v == nil {
		v = &pendingMetricRows{}
	}
	return v.(*pendingMetricRows)
}

func putPendingMetricRows(pmrs *pendingMetricRows) {
	pmrs.reset()
	pendingMetricRowsPool.Put(pmrs)
}

var pendingMetricRowsPool sync.Pool

func (s *Storage) updatePerDateData(rows []rawRow) error {
	var date uint64
	var hour uint64
	var prevTimestamp int64
	var (
		// These vars are used for speeding up bulk imports when multiple adjancent rows
		// contain the same (metricID, date) pairs.
		prevDate     uint64
		prevMetricID uint64
	)
	hm := s.currHourMetricIDs.Load().(*hourMetricIDs)
	nextDayMetricIDs := &s.nextDayMetricIDs.Load().(*byDateMetricIDEntry).v
	todayShare16bit := uint64((float64(fasttime.UnixTimestamp()%(3600*24)) / (3600 * 24)) * (1 << 16))
	type pendingDateMetricID struct {
		date     uint64
		metricID uint64
	}
	var pendingDateMetricIDs []pendingDateMetricID
	for i := range rows {
		r := &rows[i]
		if r.Timestamp != prevTimestamp {
			date = uint64(r.Timestamp) / msecPerDay
			hour = uint64(r.Timestamp) / msecPerHour
			prevTimestamp = r.Timestamp
		}
		metricID := r.TSID.MetricID
		if hour == hm.hour {
			// The r belongs to the current hour. Check for the current hour cache.
			if hm.m.Has(metricID) {
				// Fast path: the metricID is in the current hour cache.
				// This means the metricID has been already added to per-day inverted index.

				// Gradually pre-populate per-day inverted index for the next day
				// during the current day.
				// This should reduce CPU usage spike and slowdown at the beginning of the next day
				// when entries for all the active time series must be added to the index.
				// This should address https://github.com/VictoriaMetrics/VictoriaMetrics/issues/430 .
				if todayShare16bit > (metricID&(1<<16-1)) && !nextDayMetricIDs.Has(metricID) {
					pendingDateMetricIDs = append(pendingDateMetricIDs, pendingDateMetricID{
						date:     date + 1,
						metricID: metricID,
					})
					s.pendingNextDayMetricIDsLock.Lock()
					s.pendingNextDayMetricIDs.Add(metricID)
					s.pendingNextDayMetricIDsLock.Unlock()
				}
				continue
			}
			s.pendingHourEntriesLock.Lock()
			s.pendingHourEntries.Add(metricID)
			s.pendingHourEntriesLock.Unlock()
		}

		// Slower path: check global cache for (date, metricID) entry.
		if metricID == prevMetricID && date == prevDate {
			// Fast path for bulk import of multiple rows with the same (date, metricID) pairs.
			continue
		}
		prevDate = date
		prevMetricID = metricID

		if !s.dateMetricIDCache.Has(date, metricID) {
			// Slow path: store the (date, metricID) entry in the indexDB.
			// It is OK if the (date, metricID) entry is added multiple times to db
			// by concurrent goroutines.
			pendingDateMetricIDs = append(pendingDateMetricIDs, pendingDateMetricID{
				date:     date,
				metricID: metricID,
			})
		}
	}
	if len(pendingDateMetricIDs) == 0 {
		// Fast path - there are no new (date, metricID) entires in rows.
		return nil
	}

	// Slow path - add new (date, metricID) entries to indexDB.

	atomic.AddUint64(&s.slowPerDayIndexInserts, uint64(len(pendingDateMetricIDs)))
	// Sort pendingDateMetricIDs by (date, metricID) in order to speed up `is` search in the loop below.
	sort.Slice(pendingDateMetricIDs, func(i, j int) bool {
		a := pendingDateMetricIDs[i]
		b := pendingDateMetricIDs[j]
		if a.date != b.date {
			return a.date < b.date
		}
		return a.metricID < b.metricID
	})
	idb := s.idb()
	is := idb.getIndexSearch(noDeadline)
	defer idb.putIndexSearch(is)
	var firstError error
	prevMetricID = 0
	prevDate = 0
	for _, dateMetricID := range pendingDateMetricIDs {
		date := dateMetricID.date
		metricID := dateMetricID.metricID
		if metricID == prevMetricID && date == prevDate {
			// Fast path for bulk import of multiple rows with the same (date, metricID) pairs.
			continue
		}
		prevDate = date
		prevMetricID = metricID

		if s.dateMetricIDCache.Has(date, metricID) {
			// The metricID has been already added to per-day inverted index.
			continue
		}
		ok, err := is.hasDateMetricID(date, metricID)
		if err != nil {
			if firstError == nil {
				firstError = fmt.Errorf("error when locating (date=%d, metricID=%d) in database: %w", date, metricID, err)
			}
			continue
		}
		if !ok {
			// The (date, metricID) entry is missing in the indexDB. Add it there.
			if err := is.storeDateMetricID(date, metricID); err != nil {
				if firstError == nil {
					firstError = fmt.Errorf("error when storing (date=%d, metricID=%d) in database: %w", date, metricID, err)
				}
				continue
			}
		}
		// The metric must be added to cache only after it has been successfully added to indexDB.
		s.dateMetricIDCache.Set(date, metricID)
	}
	return firstError
}

// dateMetricIDCache is fast cache for holding (date, metricID) entries.
//
// It should be faster than map[date]*uint64set.Set on multicore systems.
type dateMetricIDCache struct {
	// 64-bit counters must be at the top of the structure to be properly aligned on 32-bit arches.
	syncsCount  uint64
	resetsCount uint64

	// Contains immutable map
	byDate atomic.Value

	// Contains mutable map protected by mu
	byDateMutable *byDateMetricIDMap
	lastSyncTime  uint64
	mu            sync.Mutex
}

func newDateMetricIDCache() *dateMetricIDCache {
	var dmc dateMetricIDCache
	dmc.Reset()
	return &dmc
}

func (dmc *dateMetricIDCache) Reset() {
	dmc.mu.Lock()
	// Do not reset syncsCount and resetsCount
	dmc.byDate.Store(newByDateMetricIDMap())
	dmc.byDateMutable = newByDateMetricIDMap()
	dmc.lastSyncTime = fasttime.UnixTimestamp()
	dmc.mu.Unlock()

	atomic.AddUint64(&dmc.resetsCount, 1)
}

func (dmc *dateMetricIDCache) EntriesCount() int {
	byDate := dmc.byDate.Load().(*byDateMetricIDMap)
	n := 0
	for _, e := range byDate.m {
		n += e.v.Len()
	}
	return n
}

func (dmc *dateMetricIDCache) SizeBytes() uint64 {
	byDate := dmc.byDate.Load().(*byDateMetricIDMap)
	n := uint64(0)
	for _, e := range byDate.m {
		n += e.v.SizeBytes()
	}
	return n
}

func (dmc *dateMetricIDCache) Has(date, metricID uint64) bool {
	byDate := dmc.byDate.Load().(*byDateMetricIDMap)
	v := byDate.get(date)
	if v.Has(metricID) {
		// Fast path.
		// The majority of calls must go here.
		return true
	}

	// Slow path. Check mutable map.
	currentTime := fasttime.UnixTimestamp()
	dmc.mu.Lock()
	v = dmc.byDateMutable.get(date)
	ok := v.Has(metricID)
	mustSync := false
	if currentTime-dmc.lastSyncTime > 10 {
		mustSync = true
		dmc.lastSyncTime = currentTime
	}
	dmc.mu.Unlock()

	if mustSync {
		dmc.sync()
	}
	return ok
}

func (dmc *dateMetricIDCache) Set(date, metricID uint64) {
	dmc.mu.Lock()
	v := dmc.byDateMutable.getOrCreate(date)
	v.Add(metricID)
	dmc.mu.Unlock()
}

func (dmc *dateMetricIDCache) sync() {
	dmc.mu.Lock()
	byDate := dmc.byDate.Load().(*byDateMetricIDMap)
	for date, e := range dmc.byDateMutable.m {
		v := byDate.get(date)
		e.v.Union(v)
	}
	dmc.byDate.Store(dmc.byDateMutable)
	dmc.byDateMutable = newByDateMetricIDMap()
	dmc.mu.Unlock()

	atomic.AddUint64(&dmc.syncsCount, 1)

	if dmc.EntriesCount() > memory.Allowed()/128 {
		dmc.Reset()
	}
}

type byDateMetricIDMap struct {
	hotEntry atomic.Value
	m        map[uint64]*byDateMetricIDEntry
}

func newByDateMetricIDMap() *byDateMetricIDMap {
	dmm := &byDateMetricIDMap{
		m: make(map[uint64]*byDateMetricIDEntry),
	}
	dmm.hotEntry.Store(&byDateMetricIDEntry{})
	return dmm
}

func (dmm *byDateMetricIDMap) get(date uint64) *uint64set.Set {
	hotEntry := dmm.hotEntry.Load().(*byDateMetricIDEntry)
	if hotEntry.date == date {
		// Fast path
		return &hotEntry.v
	}
	// Slow path
	e := dmm.m[date]
	if e == nil {
		return nil
	}
	dmm.hotEntry.Store(e)
	return &e.v
}

func (dmm *byDateMetricIDMap) getOrCreate(date uint64) *uint64set.Set {
	v := dmm.get(date)
	if v != nil {
		return v
	}
	e := &byDateMetricIDEntry{
		date: date,
	}
	dmm.m[date] = e
	return &e.v
}

type byDateMetricIDEntry struct {
	date uint64
	v    uint64set.Set
}

func (s *Storage) updateNextDayMetricIDs() {
	date := fasttime.UnixDate()
	e := s.nextDayMetricIDs.Load().(*byDateMetricIDEntry)
	s.pendingNextDayMetricIDsLock.Lock()
	pendingMetricIDs := s.pendingNextDayMetricIDs
	s.pendingNextDayMetricIDs = &uint64set.Set{}
	s.pendingNextDayMetricIDsLock.Unlock()
	if pendingMetricIDs.Len() == 0 && e.date == date {
		// Fast path: nothing to update.
		return
	}

	// Slow path: union pendingMetricIDs with e.v
	if e.date == date {
		pendingMetricIDs.Union(&e.v)
	}
	eNew := &byDateMetricIDEntry{
		date: date,
		v:    *pendingMetricIDs,
	}
	s.nextDayMetricIDs.Store(eNew)
}

func (s *Storage) updateCurrHourMetricIDs() {
	hm := s.currHourMetricIDs.Load().(*hourMetricIDs)
	s.pendingHourEntriesLock.Lock()
	newMetricIDs := s.pendingHourEntries
	s.pendingHourEntries = &uint64set.Set{}
	s.pendingHourEntriesLock.Unlock()
	hour := fasttime.UnixHour()
	if newMetricIDs.Len() == 0 && hm.hour == hour {
		// Fast path: nothing to update.
		return
	}

	// Slow path: hm.m must be updated with non-empty s.pendingHourEntries.
	var m *uint64set.Set
	isFull := hm.isFull
	if hm.hour == hour {
		m = hm.m.Clone()
	} else {
		m = &uint64set.Set{}
		isFull = true
	}
	m.Union(newMetricIDs)
	hmNew := &hourMetricIDs{
		m:      m,
		hour:   hour,
		isFull: isFull,
	}
	s.currHourMetricIDs.Store(hmNew)
	if hm.hour != hour {
		s.prevHourMetricIDs.Store(hm)
	}
}

type hourMetricIDs struct {
	m      *uint64set.Set
	hour   uint64
	isFull bool
}

func (s *Storage) getTSIDFromCache(dst *TSID, metricName []byte) bool {
	buf := (*[unsafe.Sizeof(*dst)]byte)(unsafe.Pointer(dst))[:]
	buf = s.tsidCache.Get(buf[:0], metricName)
	return uintptr(len(buf)) == unsafe.Sizeof(*dst)
}

func (s *Storage) putTSIDToCache(tsid *TSID, metricName []byte) {
	buf := (*[unsafe.Sizeof(*tsid)]byte)(unsafe.Pointer(tsid))[:]
	s.tsidCache.Set(metricName, buf)
}

func openIndexDBTables(path string, metricIDCache, metricNameCache, tsidCache *workingsetcache.Cache) (curr, prev *indexDB, err error) {
	if err := fs.MkdirAllIfNotExist(path); err != nil {
		return nil, nil, fmt.Errorf("cannot create directory %q: %w", path, err)
	}

	d, err := os.Open(path)
	if err != nil {
		return nil, nil, fmt.Errorf("cannot open directory: %w", err)
	}
	defer fs.MustClose(d)

	// Search for the two most recent tables - the last one is active,
	// the previous one contains backup data.
	fis, err := d.Readdir(-1)
	if err != nil {
		return nil, nil, fmt.Errorf("cannot read directory: %w", err)
	}
	var tableNames []string
	for _, fi := range fis {
		if !fs.IsDirOrSymlink(fi) {
			// Skip non-directories.
			continue
		}
		tableName := fi.Name()
		if !indexDBTableNameRegexp.MatchString(tableName) {
			// Skip invalid directories.
			continue
		}
		tableNames = append(tableNames, tableName)
	}
	sort.Slice(tableNames, func(i, j int) bool {
		return tableNames[i] < tableNames[j]
	})
	if len(tableNames) < 2 {
		// Create missing tables
		if len(tableNames) == 0 {
			prevName := nextIndexDBTableName()
			tableNames = append(tableNames, prevName)
		}
		currName := nextIndexDBTableName()
		tableNames = append(tableNames, currName)
	}

	// Invariant: len(tableNames) >= 2

	// Remove all the tables except two last tables.
	for _, tn := range tableNames[:len(tableNames)-2] {
		pathToRemove := path + "/" + tn
		logger.Infof("removing obsolete indexdb dir %q...", pathToRemove)
		fs.MustRemoveAll(pathToRemove)
		logger.Infof("removed obsolete indexdb dir %q", pathToRemove)
	}

	// Persist changes on the file system.
	fs.MustSyncPath(path)

	// Open the last two tables.
	currPath := path + "/" + tableNames[len(tableNames)-1]

	curr, err = openIndexDB(currPath, metricIDCache, metricNameCache, tsidCache)
	if err != nil {
		return nil, nil, fmt.Errorf("cannot open curr indexdb table at %q: %w", currPath, err)
	}
	prevPath := path + "/" + tableNames[len(tableNames)-2]
	prev, err = openIndexDB(prevPath, metricIDCache, metricNameCache, tsidCache)
	if err != nil {
		curr.MustClose()
		return nil, nil, fmt.Errorf("cannot open prev indexdb table at %q: %w", prevPath, err)
	}

	return curr, prev, nil
}

var indexDBTableNameRegexp = regexp.MustCompile("^[0-9A-F]{16}$")

func nextIndexDBTableName() string {
	n := atomic.AddUint64(&indexDBTableIdx, 1)
	return fmt.Sprintf("%016X", n)
}

var indexDBTableIdx = uint64(time.Now().UnixNano())