From 49df29027005c7f7c7e3e4c8e3ab01a40bd03680 Mon Sep 17 00:00:00 2001
From: Harshavardhana <harsha@minio.io>
Date: Thu, 6 Feb 2020 13:52:11 +0530
Subject: [PATCH] Add metadata parsing to be inside mutex to slow down (#8952)

Adding mutex slows down the crawler to avoid large
spikes in CPU, also add millisecond interval jitter
in calculation of disk usage to slow down the spikes
further.
---
 cmd/data-usage.go | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)

diff --git a/cmd/data-usage.go b/cmd/data-usage.go
index 637437dbb..a77463c02 100644
--- a/cmd/data-usage.go
+++ b/cmd/data-usage.go
@@ -20,6 +20,7 @@ import (
 	"bytes"
 	"context"
 	"encoding/json"
+	"math/rand"
 	"os"
 	"path/filepath"
 	"sync"
@@ -159,13 +160,19 @@ func updateUsage(basePath string, doneCh <-chan struct{}, waitForLowActiveIO fun
 	}
 
 	numWorkers := 4
+	walkInterval := 1 * time.Millisecond
 
 	var mutex sync.Mutex // Mutex to update dataUsageInfo
 
+	r := rand.New(rand.NewSource(UTCNow().UnixNano()))
+
 	fastWalk(basePath, numWorkers, doneCh, func(path string, typ os.FileMode) error {
 		// Wait for I/O to go down.
 		waitForLowActiveIO()
 
+		// Randomize sleep intervals, to stagger the walk.
+		defer time.Sleep(time.Duration(r.Float64() * float64(walkInterval)))
+
 		bucket, entry := path2BucketObjectWithBasePath(basePath, path)
 		if bucket == "" {
 			return nil
@@ -183,6 +190,9 @@ func updateUsage(basePath string, doneCh <-chan struct{}, waitForLowActiveIO fun
 			return nil
 		}
 
+		mutex.Lock()
+		defer mutex.Unlock()
+
 		if typ&os.ModeDir != 0 {
 			return nil
 		}
@@ -192,12 +202,10 @@ func updateUsage(basePath string, doneCh <-chan struct{}, waitForLowActiveIO fun
 			return errSkipFile
 		}
 
-		mutex.Lock()
 		dataUsageInfo.ObjectsCount++
 		dataUsageInfo.ObjectsTotalSize += uint64(size)
 		dataUsageInfo.BucketsSizes[bucket] += uint64(size)
 		dataUsageInfo.ObjectsSizesHistogram[objSizeToHistoInterval(uint64(size))]++
-		mutex.Unlock()
 		return nil
 	})