Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


New on LowEndTalk? Please Register and read our Community Rules.

All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

Test the disk I/O of your VPS

145791014

Comments

  • NickPNickP Member
    edited March 2012

    @vps6net said: @NickP @kbar Let us know if you're seeing speeds like that, even 60-80 MB/s is no good -- you should be getting 125 or above with those write tests.

    Looking good today ;)

    (VPS6.NET VPS)
    [root@server ~]# dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync
    16384+0 records in
    16384+0 records out
    1073741824 bytes (1.1 GB) copied, 8.43322 seconds, 127 MB/s

    Also sending you a Pm :)

  • JTRJTR Member

    @Francisco said: @JTR said: 1073741824 bytes (1.1 GB) copied, 10.026 s, 107 MB/s

    What's uname -a report?

    I'm thinking pony6-1?

    If not you should log a ticket for us to check that box, the network should be a lot more like 60 - 80MB/sec

    Francisco

    I had changed my host name, but the control panel reports "512 (node07)".

    I re-ran the test, here's another run through it — IO seems better, network is not too good.

    CPU model :  Intel(R) Xeon(R) CPU           L5520  @ 2.27GHz
    Number of cores : 2
    CPU frequency :  2266.746 MHz
    Total amount of ram : 1024 MB
    Total amount of swap : 0 MB
    System uptime :   35 days, 14:58,       
    Download speed from CacheFly: 29.6MB/s 
    Download speed from Linode, Atlanta GA: 1.99MB/s 
    Download speed from Linode, Dallas, TX: 5.69MB/s 
    Download speed from Linode, Tokyo, JP: 836KB/s 
    Download speed from Softlayer, Seattle, WA: 3.17MB/s 
    Download speed from Softlayer, San Jose, CA: 4.07MB/s 
    Download speed from Softlayer, Washington, DC: 1.23MB/s 
    I/O speed :  130 MB/s
    
    1073741824 bytes (1.1 GB) copied, 5.14333 s, 209 MB/s
    

    Should I open up a ticket?

  • flyfly Member

    @jtr can you post that test script? I can't find it with a quick google.

  • JTRJTR Member

    @kbar

    freevps.us/downloads/bench.sh

    Thanked by 1fly
  • aubsaubs Member
    edited March 2012

    vooservers.com
    Node : MAIVPS05
    (Maidtone, UK)

    dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync

    16384+0 records in
    16384+0 records out
    1073741824 bytes (1.1 GB) copied, 19.0031 s, 56.5 MB/s

    quickweb.co.nz
    Node : Spitfire (UK)
    (London, UK)

    dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync

    16384+0 records in
    16384+0 records out
    1073741824 bytes (1.1 GB) copied, 30.1121 s, 35.7 MB/s

  • FranciscoFrancisco Top Host, Host Rep, Veteran

    @JTR said: Should I open up a ticket?

    Sure, just link this post.

    It's possible when you did the IO test it was in the midst of crons or what have you. 200M/sec sounds good and it likely spikes higher.

    Francisco

  • BlueVM
    128 MB of Guaranteed Ram
    256 MB of Burstable RAM

    1073741824 bytes (1.1 GB) copied, 14.9847 s, 71.7 MB/s

  • First timer for KVM based VPS nabbed it on latest stock release last week :) Seems disk I/O differing from my OpenVZs all with buyvm ?

    Comparing the KVM vs OpenVZ both with CentOS 6.2 32bit EXT4:

    --------
    OpenVZ

    [root@openvz ~]# dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync; rm -rf test
    16384+0 records in
    16384+0 records out
    1073741824 bytes (1.1 GB) copied, 9.14105 s, 117 MB/s
    
    [root@openvz ~]# uname -r
    2.6.18-6-pve
    
    [root@openvz ~]# df -T
    Filesystem    Type   1K-blocks      Used Available Use% Mounted on
    /dev/simfs
              reiserfs    52428800   3707384  48721416   8% /
    none         tmpfs      524288         4    524284   1% /dev
    tmpfs        tmpfs      524288        16    524272   1% /tmp
    

    --------
    KVM

    [root@kvm ~]# dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync; rm -rf test
    16384+0 records in
    16384+0 records out
    1073741824 bytes (1.1 GB) copied, 17.8538 s, 60.1 MB/s
    
    [root@kvm ~]# uname -r
    2.6.32-220.7.1.el6.i686
    
    [root@kvm ~]# df -T
    Filesystem    Type   1K-blocks      Used Available Use% Mounted on
    /dev/mapper/VolGroup-lv_root
                  ext4    50065640   2216116  45306320   5% /
    tmpfs        tmpfs      255576         0    255576   0% /dev/shm
    /dev/vda1     ext4      495844     49713    420531  11% /boot
    tmpfs        tmpfs      255576         0    255576   0% /tmp
    

    **KVM disk and NIC = virtio **

    lsmod | grep virtio
    virtio_balloon          3519  0 
    virtio_net             12380  0 
    virtio_blk              5496  3 
    virtio_pci              5294  0 
    virtio_ring             6576  4 virtio_balloon,virtio_net,virtio_blk,virtio_pci
    virtio                  3669  4 virtio_balloon,virtio_net,virtio_blk,virtio_pci
    
    lspci | grep Virt
    00:03.0 Ethernet controller: Red Hat, Inc Virtio network device
    00:04.0 Ethernet controller: Red Hat, Inc Virtio network device
    00:05.0 SCSI storage controller: Red Hat, Inc Virtio block device
    00:06.0 RAM memory: Red Hat, Inc Virtio memory balloon
    

    However, network speed is excellent on KVM VPS

    single threaded wget (speedtest)
    vs
    multi-threaded (2 threads) axel download accelerator (speedtestaxel)

    [root@kvm]# ./speedtest2.sh speedtest
    -------------------------------------------------------------
    CPU model :  QEMU Virtual CPU version 1.0
    Number of cores : 2
    CPU frequency :  3192.748 MHz
    Total amount of ram : 499 MB
    Total amount of swap : 1023 MB
    System uptime :   3:47,       
    -------------------------------------------------------------
    Download speed from CacheFly: 68.8MB/s 
    Download speed from Linode, Atlanta GA: 2.23MB/s 
    Download speed from Linode, Dallas, TX: 8.76MB/s 
    Download speed from Linode, Tokyo, JP: 8.44MB/s 
    Download speed from Linode, London, UK: 4.37MB/s 
    Download speed from Leaseweb, Haarlem, NL: 3.92MB/s 
    Download speed from Softlayer, Singapore: 3.15MB/s 
    Download speed from Softlayer, Seattle, WA: 13.5MB/s 
    Download speed from Softlayer, San Jose, CA: 41.2MB/s 
    Download speed from Softlayer, Washington, DC: 6.79MB/s 
    
    [root@kvm]# ./speedtest2.sh speedtestaxel
    -------------------------------------------------------------
    CPU model :  QEMU Virtual CPU version 1.0
    Number of cores : 2
    CPU frequency :  3192.748 MHz
    Total amount of ram : 499 MB
    Total amount of swap : 1023 MB
    System uptime :   3:51,       
    -------------------------------------------------------------
    Download speed from CacheFly: 67.59 MB/s
    Download speed from Linode, Atlanta GA: 10.66 MB/s
    Download speed from Linode, Dallas, TX: 17.58 MB/s
    Download speed from Linode, Tokyo, JP: 15.35 MB/s
    Download speed from Linode, London, UK: 8.04 MB/s
    Download speed from Leaseweb, Haarlem, NL: 9.98 MB/s
    Download speed from Softlayer, Singapore: 4.30 MB/s
    Download speed from Softlayer, Seattle, WA: 31.07 MB/s
    Download speed from Softlayer, San Jose, CA: 67.28 MB/s
    Download speed from Softlayer, Washington, DC: 4.33 MB/s
    
  • FranciscoFrancisco Top Host, Host Rep, Veteran

    @eva2000 said: First timer for KVM based VPS nabbed it on latest stock release last week :) Seems disk I/O differing from my OpenVZs all with buyvm ?

    It will some since the OVZ's have a much bigger array. You can log a ticket and we can at least check our side to make sure all is OK :)

    The OVZ looks a bit low so you may want to ticket that as well. If you're on node57/58, then the IO looks right for now.

    Francisco

  • eva2000eva2000 Veteran
    edited March 2012

    Ah that would explain OpenVZ's better speed :)

    2nd run seems better/cached

    [root@openvz ~]# dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync; rm -rf test
    16384+0 records in
    16384+0 records out
    1073741824 bytes (1.1 GB) copied, 7.0963 s, 151 MB/s
    

    got 2x more openvz VPS with buyvm one on same node as first openvz above so disk I/O close

    2nd openvz same node VPS

    1st run

    [root@openvz2 ~]# dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync; rm -rf test
    16384+0 records in
    16384+0 records out
    1073741824 bytes (1.1 GB) copied, 6.12499 s, 175 MB/s
    

    2nd run

     [root@openvz2 ~]# dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync; rm -rf test
     16384+0 records in
     16384+0 records out
     1073741824 bytes (1.1 GB) copied, 5.97544 s, 180 MB/s
    

    3rd openvz different node VPS

    1st run

     [root@openvz3 ~]# dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync; rm -rf test
     16384+0 records in
     16384+0 records out
     1073741824 bytes (1.1 GB) copied, 12.2858 s, 87.4 MB/s
    

    2nd run

     [root@openvz3 ~]# dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync; rm -rf test
     16384+0 records in
     16384+0 records out
     1073741824 bytes (1.1 GB) copied, 4.50728 s, 238 MB/s
    
  • NateN34NateN34 Member
    edited March 2012

    BuyVM 1024 OpenVZ:

    root@custard:~# dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync

    16384+0 records in
    16384+0 records out
    1073741824 bytes (1.1 GB) copied, 3.16738 s, 339 MB/s

    Hostigation 1024 KVM:

    root@danish:~# dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync

    16384+0 records in
    16384+0 records out
    1073741824 bytes (1.1 GB) copied, 6.25823 s, 172 MB/s

    ChicagoVPS 512 XEN:

    root@creamfilled:~# dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync

    16384+0 records in
    16384+0 records out
    1073741824 bytes (1.1 GB) copied, 27.4499 s, 39.1 MB/s

    KiloServe 384 MB XEN:

    root@website:~# dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync

    16384+0 records in
    16384+0 records out
    1073741824 bytes (1.1 GB) copied, 9.47876 s, 113 MB/s

  • FranciscoFrancisco Top Host, Host Rep, Veteran

    @eva2000 - likely just a busy node then :) If you feel a bite in lag just let me know and we'll see what we can do. We have a 2.6.32 node still in action that we push troubled customers to, especially ones that love to eat harddrives.

    Nate's speeds are a lot more common around here for OVZ.

    Francisco

  • Here are from my latest VPSs:

    BuyVM 512 MB OpenVZ

    dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync && rm -rf test

    16384+0 records in
    16384+0 records out
    1073741824 bytes (1.1 GB) copied, 5.75904 s, 186 MB/s

    Hostigation OVZ-1024

    dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync && rm -rf test

    16384+0 records in
    16384+0 records out
    1073741824 bytes (1.1 GB) copied, 15.5739 s, 68.9 MB/s

  • @NickP Wow that's wrongly slow, I would contact support about that.

  • BuyVM 128 MB OpenVZ

    dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync
    16384+0 records in
    16384+0 records out
    1073741824 bytes (1.1 GB) copied, 4.93622 s, 218 MB/s

    BuyVM 256 MB OpenVZ

    dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync
    16384+0 records in
    16384+0 records out
    1073741824 bytes (1.1 GB) copied, 7.29418 s, 147 MB/s

    BuyVM 512 MB KVM Storage

    dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync
    16384+0 records in
    16384+0 records out
    1073741824 bytes (1.1 GB) copied, 1.63124 s, 658 MB/s

    Hostigation 256 MB OpenVZ

    dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync
    16384+0 records in
    16384+0 records out
    1073741824 bytes (1.1 GB) copied, 6.55324 s, 164 MB/s

    SecureDragon 96 MB OpenVZ

    dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync
    16384+0 records in
    16384+0 records out
    1073741824 bytes (1.1 GB) copied, 9.81746 s, 109 MB/s

  • JacobJacob Member

    @Francisco What the hell do you put in your servers! > . >

  • Pony sauce :3

  • AndriAndri Member

    @CBebop said: 1073741824 bytes (1.1 GB) copied, 1.63124 s, 658 MB/s

    O.O

  • JacobJacob Member

    @Aldryic You guys must spend a fortune on Drives and RAID Cards.

  • My reaction exactly when I first ran one on the new storage node.

  • I would guess a RAID card with crazy amount of cache (1GB or more).

  • FranciscoFrancisco Top Host, Host Rep, Veteran

    @rds100 said: I would guess a RAID card with crazy amount of cache (1GB or more).

    512M.

    There's a lot of optimizing that goes into each rig. We may take a long time to get stock runs out, but that's because we spend so much time making sure everything is smooth.

    New storage is a fucking beast.

    Francisco

  • RophRoph Member

    Thought I'd update this since I have multiple LEBs now :)

    Quickweb 384/512 OVZ in Phoenix, $40/yr:
    1073741824 bytes (1.1 GB) copied, 1.19009 s, 902 MB/s
    1073741824 bytes (1.1 GB) copied, 1.04502 s, 1.0 GB/s
    1073741824 bytes (1.1 GB) copied, 1.09241 s, 983 MB/s

    Yeah.... wow =o

    Minivps 128M OVZ, £1.50/mo:
    1073741824 bytes (1.1 GB) copied, 21.4787 s, 50.0 MB/s
    1073741824 bytes (1.1 GB) copied, 31.0556 s, 34.6 MB/s
    1073741824 bytes (1.1 GB) copied, 16.9466 s, 63.4 MB/s

    BuyVM 128/256M OVZ, $15/yr:
    1073741824 bytes (1.1 GB) copied, 5.25055 s, 205 MB/s
    1073741824 bytes (1.1 GB) copied, 4.80154 s, 224 MB/s
    1073741824 bytes (1.1 GB) copied, 5.39224 s, 199 MB/s

    And not really an LEB but for kicks, my ServInt 1024/2048 Virtuozzo VPS, $59/mo:
    1073741824 bytes (1.1 GB) copied, 14.8901 s, 72.1 MB/s
    1073741824 bytes (1.1 GB) copied, 29.4249 s, 36.5 MB/s
    1073741824 bytes (1.1 GB) copied, 10.2053 s, 105 MB/s
    1073741824 bytes (1.1 GB) copied, 12.2147 s, 87.9 MB/s

  • 128 OVZ from Secure Dragon

    stephen@nsa:~$ dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync
    16384+0 records in
    16384+0 records out
    1073741824 bytes (1.1 GB) copied, 5.39966 s, 199 MB/s

    256 OVZ with BuyVM

    stephen@kgb:~$ dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync
    16384+0 records in
    16384+0 records out
    1073741824 bytes (1.1 GB) copied, 3.98144 s, 270 MB/s
  • twaintwain Member

    Old thread revival of the day.

    $12/yr Semoweb:

    tp-sw:~ # dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync; rm test
    16384+0 records in
    16384+0 records out
    1073741824 bytes (1.1 GB) copied, 1.12451 s, 955 MB/s

  • 6.50/M 1GB ram from LFCVPS/LFCHosting
    16384+0 records in
    16384+0 records out
    1073741824 bytes (1.1 GB) copied, 7.58216 s, 142 MB/s

  • AnthonySmithAnthonySmith Member, Host Rep

    happy 1 year necro

    How is your read speed @twain ?

  • twaintwain Member

    @AnthonySmith - let me know if a different command is better..

    tp-sw:~ # dd if=test of=/dev/null bs=64k count=24000
    16384+0 records in
    16384+0 records out
    1073741824 bytes (1.1 GB) copied, 0.120681 s, 8.9 GB/s

    tp-sw:~ # dd if=test of=/dev/null bs=64k count=24000
    16384+0 records in
    16384+0 records out
    1073741824 bytes (1.1 GB) copied, 0.339974 s, 3.2 GB/s

    tp-sw:~ # dd if=test of=/dev/null bs=64k count=24000
    16384+0 records in
    16384+0 records out
    1073741824 bytes (1.1 GB) copied, 0.171461 s, 6.3 GB/s

  • AnthonySmithAnthonySmith Member, Host Rep

    hdparm -tT /dev/sda

  • twaintwain Member
    edited April 2013

    @AnthonySmith said: hdparm -tT /dev/sda

    hdparm not going to work on openvz container unless I am mistaken?

This discussion has been closed.