app/vlstorage: add -retention.maxDiskSpaceUsageBytes command-line flag for limiting the retention at VictoriaLogs by disk space usage

app/vlstorage/main.go

@@ -20,10 +20,12 @@ import (
 var (
 	retentionPeriod = flagutil.NewDuration("retentionPeriod", "7d", "Log entries with timestamps older than now-retentionPeriod are automatically deleted; "+
 		"log entries with timestamps outside the retention are also rejected during data ingestion; the minimum supported retention is 1d (one day); "+
-		"see https://docs.victoriametrics.com/victorialogs/#retention")
+		"see https://docs.victoriametrics.com/victorialogs/#retention ; see also -retention.maxDiskSpaceUsageBytes")
+	maxDiskSpaceUsageBytes = flagutil.NewBytes("retention.maxDiskSpaceUsageBytes", 0, "The maximum disk space usage at -storageDataPath before older per-day "+
+		"partitions are automatically dropped; see https://docs.victoriametrics.com/victorialogs/#retention-by-disk-space-usage ; see also -retentionPeriod")
 	futureRetention = flagutil.NewDuration("futureRetention", "2d", "Log entries with timestamps bigger than now+futureRetention are rejected during data ingestion; "+
 		"see https://docs.victoriametrics.com/victorialogs/#retention")
-	storageDataPath = flag.String("storageDataPath", "victoria-logs-data", "Path to directory with the VictoriaLogs data; "+
+	storageDataPath = flag.String("storageDataPath", "victoria-logs-data", "Path to directory where to store VictoriaLogs data; "+
 		"see https://docs.victoriametrics.com/victorialogs/#storage")
 	inmemoryDataFlushInterval = flag.Duration("inmemoryDataFlushInterval", 5*time.Second, "The interval for guaranteed saving of in-memory data to disk. "+
 		"The saved data survives unclean shutdowns such as OOM crash, hardware reset, SIGKILL, etc. "+
@@ -49,12 +51,13 @@ func Init() {
 		logger.Fatalf("-retentionPeriod cannot be smaller than a day; got %s", retentionPeriod)
 	}
 	cfg := &logstorage.StorageConfig{
-		Retention:             retentionPeriod.Duration(),
-		FlushInterval:         *inmemoryDataFlushInterval,
-		FutureRetention:       futureRetention.Duration(),
-		LogNewStreams:         *logNewStreams,
-		LogIngestedRows:       *logIngestedRows,
-		MinFreeDiskSpaceBytes: minFreeDiskSpaceBytes.N,
+		Retention:              retentionPeriod.Duration(),
+		MaxDiskSpaceUsageBytes: maxDiskSpaceUsageBytes.N,
+		FlushInterval:          *inmemoryDataFlushInterval,
+		FutureRetention:        futureRetention.Duration(),
+		LogNewStreams:          *logNewStreams,
+		LogIngestedRows:        *logIngestedRows,
+		MinFreeDiskSpaceBytes:  minFreeDiskSpaceBytes.N,
 	}
 	logger.Infof("opening storage at -storageDataPath=%s", *storageDataPath)
 	startTime := time.Now()

docs/VictoriaLogs/CHANGELOG.md

@@ -19,6 +19,8 @@ according to [these docs](https://docs.victoriametrics.com/victorialogs/quicksta
 
 ## tip
 
+* FEATURE: add `-retention.maxDiskSpaceUsageBytes` command-line flag, which allows limiting disk space usage for [VictoriaLogs data](https://docs.victoriametrics.com/victorialogs/#storage) by automatic dropping the oldest per-day partitions if the storage disk space usage becomes bigger than the `-retention.maxDiskSpaceUsageBytes`. See [these docs](https://docs.victoriametrics.com/victorialogs/#retention-by-disk-space-usage).
+
 ## [v0.23.0](https://github.com/VictoriaMetrics/VictoriaMetrics/releases/tag/v0.23.0-victorialogs)
 
 Released at 2024-06-25

docs/VictoriaLogs/README.md

@@ -77,6 +77,8 @@ For example, the following command starts VictoriaLogs with the retention of 8 w
 /path/to/victoria-logs -retentionPeriod=8w
 ```
 
+See also [retention by disk space usage](#retention-by-disk-space-usage).
+
 VictoriaLogs stores the [ingested](https://docs.victoriametrics.com/victorialogs/data-ingestion/) logs in per-day partition directories.
 It automatically drops partition directories outside the configured retention.
 
@@ -101,6 +103,23 @@ For example, the following command starts VictoriaLogs, which accepts logs with
 /path/to/victoria-logs -futureRetention=1y
 ```
 
+## Retention by disk space usage
+
+VictoriaLogs can be configured to automatically drop older per-day partitions if the total size of partitions at [`-storageDataPath` directory](#storage)
+becomes bigger than the given threshold at `-retention.maxDiskSpaceUsageBytes` command-line flag. For example, the following command starts VictoriaLogs,
+which drops old per-day partitions if the total [storage](#storage) size becomes bigger than `100GiB`:
+
+```sh
+/path/to/victoria-logs -retention.maxDiskSpaceUsageBytes=100GiB
+```
+
+VictoriaLogs keeps at least two last days of data in order to guarantee that the logs for the last day can be returned in queries.
+This means that the total disk space usage may exceed the `-retention.maxDiskSpaceUsageBytes` if the size of the last two days of data
+exceeds the `-retention.maxDiskSpaceUsageBytes`.
+
+See also [retention](#retention).
+
+
 ## Storage
 
 VictoriaLogs stores all its data in a single directory - `victoria-logs-data`. The path to the directory can be changed via `-storageDataPath` command-line flag.
@@ -263,8 +282,11 @@ Pass `-help` to VictoriaLogs in order to see the list of supported command-line
     	Optional URL to push metrics exposed at /metrics page. See https://docs.victoriametrics.com/#push-metrics . By default, metrics exposed at /metrics page aren't pushed to any remote storage
     	Supports an array of values separated by comma or specified via multiple flags.
     	Value can contain comma inside single-quoted or double-quoted string, {}, [] and () braces.
+  -retention.maxDiskSpaceUsageBytes size
+    	The maximum disk space usage at -storageDataPath before older per-day partitions are automatically dropped; see https://docs.victoriametrics.com/victorialogs/#retention-by-disk-space-usage ; see also -retentionPeriod
+    	Supports the following optional suffixes for size values: KB, MB, GB, TB, KiB, MiB, GiB, TiB (default 0)
   -retentionPeriod value
-    	Log entries with timestamps older than now-retentionPeriod are automatically deleted; log entries with timestamps outside the retention are also rejected during data ingestion; the minimum supported retention is 1d (one day); see https://docs.victoriametrics.com/victorialogs/#retention
+    	Log entries with timestamps older than now-retentionPeriod are automatically deleted; log entries with timestamps outside the retention are also rejected during data ingestion; the minimum supported retention is 1d (one day); see https://docs.victoriametrics.com/victorialogs/#retention ; see also -retention.maxDiskSpaceUsageBytes
     	The following optional suffixes are supported: s (second), m (minute), h (hour), d (day), w (week), y (year). If suffix isn't set, then the duration is counted in months (default 7d)
   -search.maxConcurrentRequests int
     	The maximum number of concurrent search requests. It shouldn't be high, since a single request can saturate all the CPU cores, while many concurrently executed requests may require high amounts of memory. See also -search.maxQueueDuration (default 16)
@@ -276,7 +298,7 @@ Pass `-help` to VictoriaLogs in order to see the list of supported command-line
     	The minimum free disk space at -storageDataPath after which the storage stops accepting new data
     	Supports the following optional suffixes for size values: KB, MB, GB, TB, KiB, MiB, GiB, TiB (default 10000000)
   -storageDataPath string
-    	Path to directory with the VictoriaLogs data; see https://docs.victoriametrics.com/victorialogs/#storage (default "victoria-logs-data")
+    	Path to directory where to store VictoriaLogs data; see https://docs.victoriametrics.com/victorialogs/#storage (default "victoria-logs-data")
   -syslog.compressMethod.tcp array
     	Compression method for syslog messages received at the corresponding -syslog.listenAddr.tcp. Supported values: none, gzip, deflate. See https://docs.victoriametrics.com/victorialogs/data-ingestion/syslog/
     	Supports an array of values separated by comma or specified via multiple flags.

lib/logstorage/storage.go

@@ -45,7 +45,12 @@ type StorageConfig struct {
 	// Older data is automatically deleted.
 	Retention time.Duration
 
-	// FlushInterval is the interval for flushing the in-memory data to disk at the Storage
+	// MaxDiskSpaceUsageBytes is an optional maximum disk space logs can use.
+	//
+	// The oldest per-day partitions are automatically dropped if the total disk space usage exceeds this limit.
+	MaxDiskSpaceUsageBytes int64
+
+	// FlushInterval is the interval for flushing the in-memory data to disk at the Storage.
 	FlushInterval time.Duration
 
 	// FutureRetention is the allowed retention from the current time to future for the ingested data.
@@ -53,7 +58,8 @@ type StorageConfig struct {
 	// Log entries with timestamps bigger than now+FutureRetention are ignored.
 	FutureRetention time.Duration
 
-	// MinFreeDiskSpaceBytes is the minimum free disk space at storage path after which the storage stops accepting new data.
+	// MinFreeDiskSpaceBytes is the minimum free disk space at storage path after which the storage stops accepting new data
+	// and enters read-only mode.
 	MinFreeDiskSpaceBytes int64
 
 	// LogNewStreams indicates whether to log newly created log streams.
@@ -81,6 +87,11 @@ type Storage struct {
 	// older data is automatically deleted
 	retention time.Duration
 
+	// maxDiskSpaceUsageBytes is an optional maximum disk space logs can use.
+	//
+	// The oldest per-day partitions are automatically dropped if the total disk space usage exceeds this limit.
+	maxDiskSpaceUsageBytes int64
+
 	// flushInterval is the interval for flushing in-memory data to disk
 	flushInterval time.Duration
 
@@ -247,15 +258,16 @@ func MustOpenStorage(path string, cfg *StorageConfig) *Storage {
 	filterStreamCache := workingsetcache.New(mem / 10)
 
 	s := &Storage{
-		path:                  path,
-		retention:             retention,
-		flushInterval:         flushInterval,
-		futureRetention:       futureRetention,
-		minFreeDiskSpaceBytes: minFreeDiskSpaceBytes,
-		logNewStreams:         cfg.LogNewStreams,
-		logIngestedRows:       cfg.LogIngestedRows,
-		flockF:                flockF,
-		stopCh:                make(chan struct{}),
+		path:                   path,
+		retention:              retention,
+		maxDiskSpaceUsageBytes: cfg.MaxDiskSpaceUsageBytes,
+		flushInterval:          flushInterval,
+		futureRetention:        futureRetention,
+		minFreeDiskSpaceBytes:  minFreeDiskSpaceBytes,
+		logNewStreams:          cfg.LogNewStreams,
+		logIngestedRows:        cfg.LogIngestedRows,
+		flockF:                 flockF,
+		stopCh:                 make(chan struct{}),
 
 		streamIDCache:     streamIDCache,
 		streamTagsCache:   streamTagsCache,
@@ -305,6 +317,7 @@ func MustOpenStorage(path string, cfg *StorageConfig) *Storage {
 
 	s.partitions = ptws
 	s.runRetentionWatcher()
+	s.runMaxDiskSpaceUsageWatcher()
 	return s
 }
 
@@ -318,6 +331,17 @@ func (s *Storage) runRetentionWatcher() {
 	}()
 }
 
+func (s *Storage) runMaxDiskSpaceUsageWatcher() {
+	if s.maxDiskSpaceUsageBytes <= 0 {
+		return
+	}
+	s.wg.Add(1)
+	go func() {
+		s.watchMaxDiskSpaceUsage()
+		s.wg.Done()
+	}()
+}
+
 func (s *Storage) watchRetention() {
 	d := timeutil.AddJitterToDuration(time.Hour)
 	ticker := time.NewTicker(d)
@@ -360,6 +384,62 @@ func (s *Storage) watchRetention() {
 	}
 }
 
+func (s *Storage) watchMaxDiskSpaceUsage() {
+	d := timeutil.AddJitterToDuration(10 * time.Second)
+	ticker := time.NewTicker(d)
+	defer ticker.Stop()
+	for {
+		s.partitionsLock.Lock()
+		var n uint64
+		ptws := s.partitions
+		var ptwsToDelete []*partitionWrapper
+		for i := len(ptws) - 1; i >= 0; i-- {
+			ptw := ptws[i]
+			var ps PartitionStats
+			ptw.pt.updateStats(&ps)
+			n += ps.IndexdbSizeBytes + ps.CompressedSmallPartSize + ps.CompressedBigPartSize
+			if n <= uint64(s.maxDiskSpaceUsageBytes) {
+				continue
+			}
+			if i >= len(ptws)-2 {
+				// Keep the last two per-day partitions, so logs could be queried for one day time range.
+				continue
+			}
+
+			// ptws are sorted by time, so just drop all the partitions until i, including i.
+			i++
+			ptwsToDelete = ptws[:i]
+			s.partitions = ptws[i:]
+
+			// Remove reference to deleted partitions from s.ptwHot
+			for _, ptw := range ptwsToDelete {
+				if ptw == s.ptwHot {
+					s.ptwHot = nil
+					break
+				}
+			}
+
+			break
+		}
+		s.partitionsLock.Unlock()
+
+		for i, ptw := range ptwsToDelete {
+			logger.Infof("the partition %s is scheduled to be deleted because the total size of partitions exceeds -retention.maxDiskSpaceUsageBytes=%d",
+				ptw.pt.path, s.maxDiskSpaceUsageBytes)
+			ptw.mustDrop.Store(true)
+			ptw.decRef()
+
+			ptwsToDelete[i] = nil
+		}
+
+		select {
+		case <-s.stopCh:
+			return
+		case <-ticker.C:
+		}
+	}
+}
+
 func (s *Storage) getMinAllowedDay() int64 {
 	return time.Now().UTC().Add(-s.retention).UnixNano() / nsecPerDay
 }