Dateisystem-Benchmarks unter Linux

Vergleich von EXT4, XFS und ZFS

Vergleich der Dateisysteme EXT4, XFS und ZFS

Unter Linux laufen viele Server mit zwei Festplatten im RAID-1 Verbund. Dabei werden die Daten auf den Festplatten gespiegelt um Datenverlust beim Ausfall einer Festplatte zu vermeiden. Wir haben uns angesehen welche Leistung verschiedene Dateisysteme in dieser Konfiguration bringen.


Der Test wurde mit folgenden Komponenten durchgeführt:

  • CentOS 8.1 / kernel 4.18.0-147.5.1.el8_1.x86_64
  • Pro Dateisystem je zwei Festplatten im RAID-1 Verbund:
    • Model Family: HGST Travelstar 5K1000
    • Device Model: HGST HTE541010A9E680
    • User Capacity: 1,000,204,886,016 bytes [1.00 TB]
    • Sector Sizes: 512 bytes logical, 4096 bytes physical
    • Rotation Rate: 5400 rpm
    • Form Factor: 2.5 inches

ZFS Dateisystem erstellen

[root@fstest ~]# zpool create -f -o ashift=12 -m /zfspool zfspool \
                 mirror ata-HGST_HTE541010A9E680_J540001MJGTEXC ata-HGST_HTE541010A9E680_J5400013JZ6NAC
[root@fstest ~]# zpool status -v
  pool: zfspool
 state: ONLINE
  scan: none requested
config:

    NAME                                         STATE     READ WRITE CKSUM
    zfspool                                      ONLINE       0     0     0
      mirror-0                                   ONLINE       0     0     0
        ata-HGST_HTE541010A9E680_J540001MJGTEXC  ONLINE       0     0     0
        ata-HGST_HTE541010A9E680_J5400013JZ6NAC  ONLINE       0     0     0


NAME            PROPERTY              VALUE                     SOURCE
zfspool         type                  filesystem                -
zfspool         available             899G                      -
zfspool         compressratio         1.00x                     -
zfspool         quota                 none                      default
zfspool         reservation           none                      default
zfspool         recordsize            128K                      default

EXT4 Dateisystem erstellen

[root@fstest ~]# mkfs.ext4 /dev/md0
mke2fs 1.44.3 (10-July-2018)
Creating filesystem with 244157360 4k blocks and 61046784 inodes
Filesystem UUID: d592518f-30d6-43f5-8b8a-3852c0c4fbb4
Superblock backups stored on blocks: 
    32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208, 
    4096000, 7962624, 11239424, 20480000, 23887872, 71663616, 78675968, 
    102400000, 214990848

Allocating group tables: done                            
Writing inode tables: done                            
Creating journal (262144 blocks): done
Writing superblocks and filesystem accounting information: done     

XFS Dateisystem erstellen

[root@fstest ~]# mkfs.xfs /dev/md1
meta-data=/dev/md1               isize=512    agcount=4, agsize=61039340 blks
         =                       sectsz=4096  attr=2, projid32bit=1
         =                       crc=1        finobt=1, sparse=1, rmapbt=0
         =                       reflink=1
data     =                       bsize=4096   blocks=244157360, imaxpct=25
         =                       sunit=0      swidth=0 blks
naming   =version 2              bsize=4096   ascii-ci=0, ftype=1
log      =internal log           bsize=4096   blocks=119217, version=2
         =                       sectsz=4096  sunit=1 blks, lazy-count=1
realtime =none                   extsz=4096   blocks=0, rtextents=0




Tests mit bonnie++

Mehr über das Programm bonnie++ finden sie auf Wikipedia (englisch): Bonnie++

Bonnie schreibt in diesem Test 128 GB Daten auf die Festplatten, da der Server über 64 GB Arbeitsspeicher verfügt und so Caching-Effekte minimiert werden sollen.


ZFS (Version) 0.8.3

Version 1.98Sequential OutputSequential InputRandom
Seeks
Sequential CreateRandom Create
SizePer CharBlockRewritePer CharBlockNum FilesCreateReadDeleteCreateReadDelete
M/sec% CPUM/sec% CPUM/sec% CPUM/sec% CPUM/sec% CPU/sec% CPU/sec% CPU/sec% CPU/sec% CPU/sec% CPU/sec% CPU/sec% CPU
#126G1239973163692559912213113.9516496692++++++++152415505992++++++++123713
Latency64300us13486ms21657ms65773us3948ms408msLatency3486us1152us3105ms3541us23us3762ms

EXT4 (mdraid / raid 1)

Version 1.98Sequential OutputSequential InputRandom
Seeks
Sequential CreateRandom Create
SizePer CharBlockRewritePer CharBlockNum FilesCreateReadDeleteCreateReadDelete
M/sec% CPUM/sec% CPUM/sec% CPUM/sec% CPUM/sec% CPU/sec% CPU/sec% CPU/sec% CPU/sec% CPU/sec% CPU/sec% CPU/sec% CPU
#126G424979012394958931005346.15161225536++++++++++++++++2353068+++++++++++++++
Latency18630us8545ms1140ms42759us289ms836msLatency326us496us1211us266us12us2516us

XFS (mdraid / raid 1)

Version 1.98Sequential OutputSequential InputRandom
Seeks
Sequential CreateRandom Create
SizePer CharBlockRewritePer CharBlockNum FilesCreateReadDeleteCreateReadDelete
M/sec% CPUM/sec% CPUM/sec% CPUM/sec% CPUM/sec% CPU/sec% CPU/sec% CPU/sec% CPU/sec% CPU/sec% CPU/sec% CPU/sec% CPU
#126G65185796384896941055232.62716787338++++++++1275531830440++++++++1283033
Latency10684us18498us4093ms23959us176ms112msLatency464us158us212us290us15us184us


Tests mit fio - 16KB Blockgröße

Die Dokumentation zu fio (flexible I/O tester) finden Sie hier.


Zufällige Lese-/Schreibanfragen, 70% lesend, 30% schreibend, 16 KB Blockgröße, Direct-IO
fio --name=randrw --rw=randrw --direct=1 --ioengine=libaio --bs=16k --numjobs=8 --rwmixread=70 --size=1G --runtime=600 --group_reporting


ZFS ⟩ 70% lesen ⟩ 30% schreiben ⟩ 16KB Blockgröße

 read: IOPS=706, BW=11.0MiB/s (11.6MB/s)(5731MiB/519310msec)
write: IOPS=303, BW=4852KiB/s (4969kB/s)(2461MiB/519310msec)

Run status group 0 (all jobs):
 READ: bw=11.0MiB/s (11.6MB/s), 11.0MiB/s-11.0MiB/s (11.6MB/s-11.6MB/s), io=5731MiB (6010MB), run=519310-519310msec
WRITE: bw=4852KiB/s (4969kB/s), 4852KiB/s-4852KiB/s (4969kB/s-4969kB/s), io=2461MiB (2580MB), run=519310-519310msec

EXT4 - ⟩ 70% lesen ⟩ 30% schreiben ⟩ 16KB Blockgröße

 read: IOPS=132, BW=2127KiB/s (2178kB/s)(1247MiB/600427msec)
write: IOPS=57,  BW= 913KiB/s ( 935kB/s)( 535MiB/600427msec)

Run status group 0 (all jobs):
 READ: bw=2127KiB/s (2178kB/s), 2127KiB/s-2127KiB/s (2178kB/s-2178kB/s), io=1247MiB (1307MB), run=600427-600427msec
WRITE: bw=913KiB/s (935kB/s), 913KiB/s-913KiB/s (935kB/s-935kB/s), io=535MiB (561MB), run=600427-600427msec

Disk stats (read/write):
md0: ios=79802/34639, merge=0/0, ticks=0/0, in_queue=0, util=0.00%, aggrios=39901/35743, aggrmerge=0/134,
                                         aggrticks=1519734/631509, aggrin_queue=2119860, aggrutil=15.01%
sde: ios=16184/35743, merge=0/134, ticks=205446/139020, in_queue=327748, util=8.74%
sdc: ios=63618/35743, merge=0/134, ticks=2834022/1123998, in_queue=3911973, util=15.01%

XFS - ⟩ 70% lesen ⟩ 30% schreiben ⟩ 16KB Blockgröße

 read: IOPS=104, BW=1680KiB/s (1720kB/s)(984MiB/600108msec)
write: IOPS=45,   BW=727KiB/s ( 745kB/s)(426MiB/600108msec)

Run status group 0 (all jobs):
 READ: bw=1680KiB/s (1720kB/s), 1680KiB/s-1680KiB/s (1720kB/s-1720kB/s), io=984MiB (1032MB), run=600108-600108msec
WRITE: bw=727KiB/s (745kB/s), 727KiB/s-727KiB/s (745kB/s-745kB/s), io=426MiB (447MB), run=600108-600108msec

Disk stats (read/write):
md1: ios=63000/27327, merge=0/0, ticks=0/0, in_queue=0, util=0.00%, aggrios=31500/28345, aggrmerge=0/3,
                                        aggrticks=813483/586447, aggrin_queue=1375129, aggrutil=12.18%
sdg: ios=13054/28345, merge=0/4, ticks=146353/101802, in_queue=235378, util=7.11%
sdf: ios=49946/28346, merge=0/3, ticks=1480614/1071093, in_queue=2514880, util=12.18%



Tests mit fio - 128KB Blockgröße

Zufällige Lese-/Schreibanfragen, 70% lesend, 30% schreibend, 128 KB Blockgröße, Direct-IO
fio --name=randrw --rw=randrw --direct=1 --ioengine=libaio --bs=128k --numjobs=8 --rwmixread=70 --size=1G --runtime=600 --group_reporting


ZFS ⟩ 70% lesen ⟩ 30% schreiben ⟩ 128KB Blockgröße ⟩ 128 GB Daten

[root@fstest fio]# fio --name=randrw --rw=randrw --direct=0 --ioengine=libaio --bs=128k --numjobs=8 \
                       --rwmixread=70 --size=16G --runtime=600 --group_reporting

randrw: (g=0): rw=randrw, bs=(R) 128KiB-128KiB, (W) 128KiB-128KiB, (T) 128KiB-128KiB, ioengine=libaio, iodepth=1
...
fio-3.7
Starting 8 processes
randrw: Laying out IO file (1 file / 16384MiB)
randrw: Laying out IO file (1 file / 16384MiB)
randrw: Laying out IO file (1 file / 16384MiB)
randrw: Laying out IO file (1 file / 16384MiB)
randrw: Laying out IO file (1 file / 16384MiB)
randrw: Laying out IO file (1 file / 16384MiB)
randrw: Laying out IO file (1 file / 16384MiB)
randrw: Laying out IO file (1 file / 16384MiB)
Jobs: 8 (f=8): [m(8)][100.0%][r=15.1MiB/s,w=4736KiB/s][r=121,w=37 IOPS][eta 00m:00s]
randrw: (groupid=0, jobs=8): err= 0: pid=25652: Mon Mar 23 00:31:15 2020
   read: IOPS=101, BW=12.6MiB/s (13.2MB/s)(7581MiB/600088msec)
    slat (usec): min=35, max=981460, avg=78730.98, stdev=66695.87
    clat (nsec): min=1428, max=37411, avg=3757.49, stdev=1925.39
     lat (usec): min=37, max=981466, avg=78736.05, stdev=66696.27
    clat percentiles (nsec):
     |  1.00th=[ 1768],  5.00th=[ 3152], 10.00th=[ 3216], 20.00th=[ 3248],
     | 30.00th=[ 3312], 40.00th=[ 3344], 50.00th=[ 3376], 60.00th=[ 3408],
     | 70.00th=[ 3472], 80.00th=[ 3728], 90.00th=[ 4256], 95.00th=[ 4896],
     | 99.00th=[16512], 99.50th=[17024], 99.90th=[18048], 99.95th=[20608],
     | 99.99th=[27520]
   bw (  KiB/s): min=  255, max= 7424, per=12.52%, avg=1618.84, stdev=922.75, samples=9588
   iops        : min=    1, max=   58, avg=12.61, stdev= 7.21, samples=9588
  write: IOPS=43, BW=5579KiB/s (5713kB/s)(3269MiB/600088msec)
    slat (usec): min=51, max=336266, avg=958.61, stdev=10476.92
    clat (nsec): min=1572, max=23886, avg=1981.22, stdev=629.80
     lat (usec): min=53, max=336271, avg=961.20, stdev=10477.13
    clat percentiles (nsec):
     |  1.00th=[ 1672],  5.00th=[ 1736], 10.00th=[ 1784], 20.00th=[ 1832],
     | 30.00th=[ 1864], 40.00th=[ 1896], 50.00th=[ 1928], 60.00th=[ 1960],
     | 70.00th=[ 1992], 80.00th=[ 2024], 90.00th=[ 2096], 95.00th=[ 2160],
     | 99.00th=[ 3376], 99.50th=[ 3664], 99.90th=[12736], 99.95th=[14144],
     | 99.99th=[18304]
   bw (  KiB/s): min=  255, max= 5120, per=15.22%, avg=848.86, stdev=612.50, samples=7887
   iops        : min=    1, max=   40, avg= 6.60, stdev= 4.79, samples=7887
  lat (usec)   : 2=24.28%, 4=65.70%, 10=8.45%, 20=1.53%, 50=0.04%
  cpu          : usr=0.02%, sys=0.15%, ctx=59928, majf=0, minf=142
  IO depths    : 1=100.0%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     issued rwts: total=60644,26155,0,0 short=0,0,0,0 dropped=0,0,0,0
     latency   : target=0, window=0, percentile=100.00%, depth=1

Run status group 0 (all jobs):
   READ: bw=12.6MiB/s (13.2MB/s), 12.6MiB/s-12.6MiB/s (13.2MB/s-13.2MB/s), io=7581MiB (7949MB), run=600088-600088msec
  WRITE: bw=5579KiB/s (5713kB/s), 5579KiB/s-5579KiB/s (5713kB/s-5713kB/s), io=3269MiB (3428MB), run=600088-600088msec

EXT4 ⟩ 70% lesen ⟩ 30% schreiben ⟩ 128KB Blockgröße ⟩ 128 GB Daten

[root@fstest fio]# fio --name=randrw --rw=randrw --direct=0 --ioengine=libaio --bs=128k --numjobs=8 \
                       --rwmixread=70 --size=16G --runtime=600 --group_reporting

randrw: (g=0): rw=randrw, bs=(R) 128KiB-128KiB, (W) 128KiB-128KiB, (T) 128KiB-128KiB, ioengine=libaio, iodepth=1
...
fio-3.7
Starting 8 processes
randrw: Laying out IO file (1 file / 16384MiB)
randrw: Laying out IO file (1 file / 16384MiB)
randrw: Laying out IO file (1 file / 16384MiB)
randrw: Laying out IO file (1 file / 16384MiB)
randrw: Laying out IO file (1 file / 16384MiB)
randrw: Laying out IO file (1 file / 16384MiB)
randrw: Laying out IO file (1 file / 16384MiB)
randrw: Laying out IO file (1 file / 16384MiB)
Jobs: 8 (f=8): [m(8)][100.0%][r=12.0MiB/s,w=5381KiB/s][r=96,w=42 IOPS][eta 00m:00s] 
randrw: (groupid=0, jobs=8): err= 0: pid=11169: Mon Mar 23 00:29:37 2020
   read: IOPS=115, BW=14.5MiB/s (15.2MB/s)(8671MiB/600062msec)
    slat (usec): min=554, max=610323, avg=68985.35, stdev=55557.33
    clat (nsec): min=1964, max=131469, avg=4047.76, stdev=1586.81
     lat (usec): min=558, max=610332, avg=68990.75, stdev=55557.73
    clat percentiles (nsec):
     |  1.00th=[ 2448],  5.00th=[ 3312], 10.00th=[ 3504], 20.00th=[ 3568],
     | 30.00th=[ 3600], 40.00th=[ 3632], 50.00th=[ 3696], 60.00th=[ 3728],
     | 70.00th=[ 3824], 80.00th=[ 4128], 90.00th=[ 4704], 95.00th=[ 5728],
     | 99.00th=[12224], 99.50th=[13632], 99.90th=[17792], 99.95th=[20096],
     | 99.99th=[28032]
   bw (  KiB/s): min=  255, max=12032, per=12.50%, avg=1849.42, stdev=1184.94, samples=9600
   iops        : min=    1, max=   94, avg=14.39, stdev= 9.26, samples=9600
  write: IOPS=49, BW=6391KiB/s (6544kB/s)(3745MiB/600062msec)
    slat (usec): min=77, max=160148, avg=408.05, stdev=4833.27
    clat (nsec): min=1317, max=30439, avg=1817.31, stdev=631.19
     lat (usec): min=78, max=160153, avg=410.31, stdev=4833.45
    clat percentiles (nsec):
     |  1.00th=[ 1448],  5.00th=[ 1512], 10.00th=[ 1560], 20.00th=[ 1640],
     | 30.00th=[ 1704], 40.00th=[ 1736], 50.00th=[ 1768], 60.00th=[ 1816],
     | 70.00th=[ 1848], 80.00th=[ 1880], 90.00th=[ 1960], 95.00th=[ 2064],
     | 99.00th=[ 2960], 99.50th=[ 3920], 99.90th=[12480], 99.95th=[13248],
     | 99.99th=[16320]
   bw (  KiB/s): min=  255, max= 6400, per=14.52%, avg=928.03, stdev=720.88, samples=8263
   iops        : min=    1, max=   50, avg= 7.19, stdev= 5.64, samples=8263
  lat (usec)   : 2=28.10%, 4=56.05%, 10=14.52%, 20=1.30%, 50=0.04%
  lat (usec)   : 250=0.01%
  cpu          : usr=0.02%, sys=0.22%, ctx=69859, majf=0, minf=140
  IO depths    : 1=100.0%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     issued rwts: total=69371,29959,0,0 short=0,0,0,0 dropped=0,0,0,0
     latency   : target=0, window=0, percentile=100.00%, depth=1

Run status group 0 (all jobs):
   READ: bw=14.5MiB/s (15.2MB/s), 14.5MiB/s-14.5MiB/s (15.2MB/s-15.2MB/s), io=8671MiB (9093MB), run=600062-600062msec
  WRITE: bw=6391KiB/s (6544kB/s), 6391KiB/s-6391KiB/s (6544kB/s-6544kB/s), io=3745MiB (3927MB), run=600062-600062msec

Disk stats (read/write):
    md0: ios=69368/30275, merge=0/0, ticks=0/0, in_queue=0, util=0.00%, aggrios=34685/30086,
           aggrmerge=0/231, aggrticks=2384338/888440, aggrin_queue=3240162, aggrutil=12.18%
  sde: ios=33255/30064, merge=0/250, ticks=2043693/809189, in_queue=2821071, util=11.65%
  sdc: ios=36116/30109, merge=0/212, ticks=2724983/967692, in_queue=3659254, util=12.18%

XFS ⟩ 70% lesen ⟩ 30% schreiben ⟩ 128KB Blockgröße ⟩ 128 GB Daten

[root@fstest fio]# fio --name=randrw --rw=randrw --direct=0 --ioengine=libaio --bs=128k --numjobs=8 \
                       --rwmixread=70 --size=16G --runtime=600 --group_reporting

randrw: (g=0): rw=randrw, bs=(R) 128KiB-128KiB, (W) 128KiB-128KiB, (T) 128KiB-128KiB, ioengine=libaio, iodepth=1
...
fio-3.7
Starting 8 processes
randrw: Laying out IO file (1 file / 16384MiB)
randrw: Laying out IO file (1 file / 16384MiB)
randrw: Laying out IO file (1 file / 16384MiB)
randrw: Laying out IO file (1 file / 16384MiB)
randrw: Laying out IO file (1 file / 16384MiB)
randrw: Laying out IO file (1 file / 16384MiB)
randrw: Laying out IO file (1 file / 16384MiB)
randrw: Laying out IO file (1 file / 16384MiB)
Jobs: 8 (f=8): [m(8)][100.0%][r=23.4MiB/s,w=8840KiB/s][r=187,w=69 IOPS][eta 00m:00s]
randrw: (groupid=0, jobs=8): err= 0: pid=10376: Mon Mar 23 00:42:49 2020
   read: IOPS=113, BW=14.2MiB/s (14.9MB/s)(8535MiB/600059msec)
    slat (usec): min=624, max=673486, avg=70227.61, stdev=61519.31
    clat (nsec): min=2109, max=24242, avg=3674.62, stdev=865.32
     lat (usec): min=628, max=673491, avg=70232.37, stdev=61519.33
    clat percentiles (nsec):
     |  1.00th=[ 3408],  5.00th=[ 3440], 10.00th=[ 3472], 20.00th=[ 3504],
     | 30.00th=[ 3536], 40.00th=[ 3568], 50.00th=[ 3600], 60.00th=[ 3600],
     | 70.00th=[ 3632], 80.00th=[ 3664], 90.00th=[ 3728], 95.00th=[ 3760],
     | 99.00th=[ 6112], 99.50th=[10816], 99.90th=[16768], 99.95th=[17792],
     | 99.99th=[21632]
   bw (  KiB/s): min=  256, max=12288, per=12.50%, avg=1820.82, stdev=1244.41, samples=9598
   iops        : min=    2, max=   96, avg=14.18, stdev= 9.72, samples=9598
  write: IOPS=49, BW=6280KiB/s (6431kB/s)(3680MiB/600059msec)
    slat (usec): min=63, max=409, avg=82.03, stdev=10.00
    clat (nsec): min=1297, max=26489, avg=1801.48, stdev=555.07
     lat (usec): min=65, max=443, avg=84.24, stdev=10.29
    clat percentiles (nsec):
     |  1.00th=[ 1432],  5.00th=[ 1480], 10.00th=[ 1512], 20.00th=[ 1576],
     | 30.00th=[ 1704], 40.00th=[ 1768], 50.00th=[ 1816], 60.00th=[ 1848],
     | 70.00th=[ 1880], 80.00th=[ 1928], 90.00th=[ 1976], 95.00th=[ 2024],
     | 99.00th=[ 2160], 99.50th=[ 2256], 99.90th=[12992], 99.95th=[14400],
     | 99.99th=[17536]
   bw (  KiB/s): min=  255, max= 7680, per=14.84%, avg=931.85, stdev=738.46, samples=8087
   iops        : min=    1, max=   60, avg= 7.23, stdev= 5.77, samples=8087
  lat (usec)   : 2=27.83%, 4=71.05%, 10=0.64%, 20=0.46%, 50=0.02%
  cpu          : usr=0.02%, sys=0.19%, ctx=68315, majf=0, minf=187
  IO depths    : 1=100.0%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     issued rwts: total=68283,29441,0,0 short=0,0,0,0 dropped=0,0,0,0
     latency   : target=0, window=0, percentile=100.00%, depth=1

Run status group 0 (all jobs):
   READ: bw=14.2MiB/s (14.9MB/s), 14.2MiB/s-14.2MiB/s (14.9MB/s-14.9MB/s), io=8535MiB (8950MB), run=600059-600059msec
  WRITE: bw=6280KiB/s (6431kB/s), 6280KiB/s-6280KiB/s (6431kB/s-6431kB/s), io=3680MiB (3859MB), run=600059-600059msec

Disk stats (read/write):
    md1: ios=68410/23675, merge=0/0, ticks=0/0, in_queue=0, util=0.00%, aggrios=34141/23760,
            aggrmerge=63/27, aggrticks=2395357/744590, aggrin_queue=3110750, aggrutil=11.10%
  sdg: ios=34611/23754, merge=64/34, ticks=2398039/710562, in_queue=3079169, util=11.10%
  sdf: ios=33672/23767, merge=63/20, ticks=2392676/778618, in_queue=3142332, util=10.98%


Tests mit Perl

Ein Perl-Skript das 100.000 Dateien mit 4KB bzw. 128KB Größe erzeugt:


# datasize: 4k
-------------------------
# ext4 file_create 100000
  write: 16592.003/sec,    write/6.027 secs,   write_sync/7.022 secs
rewrite:  9100.009/sec, rewrite/10.989 secs, rewrite_sync/0.524 secs
   read: 38387.716/sec,     read/2.605 secs,    read_sync/0.817 secs
 delete: 28161.081/sec,   delete/3.551 secs,  delete_sync/0.951 secs

# xfs file_create 100000
  write: 10093.873/sec,    write/9.907 secs,  write_sync/14.865 secs
rewrite:  8312.552/sec, rewrite/12.030 secs, rewrite_sync/8.461 secs
   read: 35880.875/sec,     read/2.787 secs,    read_sync/6.708 secs
 delete:  7859.781/sec,  delete/12.723 secs,  delete_sync/6.815 secs

# zfs file_create 100000
  write:  4301.075/sec,   write/23.250 secs,   write_sync/7.250 secs
rewrite:  1630.683/sec, rewrite/61.324 secs, rewrite_sync/1.212 secs
   read: 29770.765/sec,     read/3.359 secs,    read_sync/0.018 secs
 delete:  2045.073/sec,  delete/48.898 secs,  delete_sync/0.143 secs

# datasize: 128k
-------------------------
# ext4 file_create 100000
  write:  1185.579/sec,    write/84.347 secs, write_sync/181.638 secs
rewrite:   625.403/sec, rewrite/159.897 secs, rewrite_sync/0.263 secs
   read: 10447.137/sec,      read/9.572 secs,    read_sync/0.431 secs
 delete: 17721.070/sec,    delete/5.643 secs,  delete_sync/0.829 secs

# xfs file_create 100000
  write: 4438.132/sec,   write/22.532 secs, write_sync/179.358 secs
rewrite: 433.937/sec, rewrite/230.448 secs, rewrite_sync/6.740 secs
   read: 9834.776/sec,    read/10.168 secs,    read_sync/5.795 secs
 delete: 4660.918/sec,  delete/21.455 secs,  delete_sync/0.591 secs

# zfs file_create 100000
  write: 356.019/sec,   write/280.884 secs,   write_sync/103.739 secs
rewrite: 433.661/sec, rewrite/230.595 secs, rewrite_sync/103.261 secs
   read: 181.663/sec,    read/550.469 secs,      read_sync/0.005 secs
 delete: 1975.075/sec,  delete/50.631 secs,    delete_sync/5.093 secs
© 2020 GCore GmbH
Information zum Datenschutz | Helfen Sie uns diese Seite zu optimieren und erlauben Sie Cookies - daraus gewonnene Informationen werden nicht an Dritte weitergegeben.
Erlauben Ablehnen