Page 4 of 10 FirstFirst 1234567 ... LastLast
Results 46 to 60 of 148
  1. #46
    Join Date
    May 2009
    Location
    US
    Posts
    2,502
    Quote Originally Posted by lilrichieh View Post
    hmm - VPSLatch

    root@inkdrop [~]# dd if=/dev/zero of=test bs=16k count=16k conv=fdatasync
    16384+0 records in
    16384+0 records out
    268435456 bytes (268 MB) copied, 11.9935 seconds, 22.4 MB/s
    Our nodes have been known to report such issues due with some BIOS/RAID card setting, I believe with write cache, we are already in the works to rectify that. However your sites should not be loading slow at all. Even if we run it on a new node with 0 VMs it would report similar speeds.

  2. #47
    Join Date
    Apr 2009
    Location
    United Kingdom
    Posts
    136
    Adam i was neither suggesting the node was poor or that my site was loading slow i was merely posting MY results. I am happy with my VPS and it's loading times...

    Happy to hear you are working on a fix though.

  3. #48
    Hi,

    Shouldn't random read or write, like in fio, be used to test disk IO speed instead of dd?

  4. #49
    Join Date
    Oct 2007
    Location
    United States
    Posts
    1,173
    Quote Originally Posted by VL-Adam View Post
    Our nodes have been known to report such issues due with some BIOS/RAID card setting, I believe with write cache, we are already in the works to rectify that. However your sites should not be loading slow at all. Even if we run it on a new node with 0 VMs it would report similar speeds.
    Most RAID controllers are around 150Mbps to 300Mbps sustained read/write like that, so the ~20-30MB/s is just fine and you probably wont get it better than that. The real test will be when you have a full node, if he can still push 25MB/s then there is nothing wrong and you wont be able to fix that unless you remove the RAID card in general. RAID cards that can push more generally cost a lot more ($400+), which most datacenters don't bother to carry unfortunately.

  5. #50

    Thumbs up

    clubuptime.com just got 1


    root@server1:/# dd if=/dev/zero of=test bs=16k count=16k conv=fdatasync
    16384+0 records in
    16384+0 records out
    268435456 bytes (268 MB) copied, 3.14158 s, 85.4 MB/s

  6. #51
    Join Date
    Mar 2009
    Location
    NL
    Posts
    553
    This is in our cloud, so it isn't very fair (because storage is on a SAN):

    [root@testserver ~]# dd if=/dev/zero of=test bs=16k count=16k conv=fdatasync
    16384+0 records in
    16384+0 records out
    268435456 bytes (268 MB) copied, 2.36134 seconds, 114 MB/s (limited to 1gbit line because it is a single thread)

    Still, on an empty node you should easily get > 20mb/s.

  7. #52
    Join Date
    Apr 2009
    Posts
    1,320
    What does that command do?
    dd if=/dev/zero of=test bs=16k count=16k conv=fdatasync

    Can it be run on live server? Will it wipe out any data?

  8. #53
    Join Date
    Mar 2009
    Location
    NL
    Posts
    553
    Quote Originally Posted by chasebug View Post
    What does that command do?
    dd if=/dev/zero of=test bs=16k count=16k conv=fdatasync

    Can it be run on live server? Will it wipe out any data?
    It only writes a file test, you can use it on a production server

  9. #54
    Join Date
    Sep 2006
    Location
    Toronto
    Posts
    158
    vpslatch (managed24 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, 26.8052 seconds, 40.1 MB/s

    ===================

    Directspace $2 server:

    [root@harber ~]# 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, 20.5059 seconds, 52.4 MB/s

  10. #55
    Join Date
    Jul 2010
    Location
    ~/
    Posts
    1,261

    My Results

    For the sake of comparison:

    PHOTON VPS:

    Code:
    [root@master ~]# dd if=/dev/zero of=test bs=64k count=4k conv=fdatasync
    4096+0 records in
    4096+0 records out
    268435456 bytes (268 MB) copied, 13.1446 seconds, 20.4 MB/s
    Racksrv.com VPS

    Code:
    root@server [~]# dd if=/dev/zero of=test bs=64k count=4k conv=fdatasync
    4096+0 records in
    4096+0 records out
    268435456 bytes (268 MB) copied, 0.774849 seconds, 346 MB/s
    2host.com VPS (actualy supprised it even ran lol)

    Code:
    [root@shout2 ~]#dd if=/dev/zero of=test bs=64k count=4k conv=fdatasync
    4096+0 records in
    4096+0 records out
    268435456 bytes (268 MB) copied, 1.82974 seconds, 147 MB/s
    My own VPS (XEN) on a crappy desktop system that is totaly overloaded without any raid and I am shocked it even still runs

    Code:
    [root@backup ~]#dd if=/dev/zero of=test bs=64k count=4k conv=fdatasync
    4096+0 records in
    4096+0 records out
    268435456 bytes (268 MB) copied, 4.51338 seconds, 59.5 MB/s
    So rackserv win hands down, I ran each test 5 times and took a result that sat in the middle on all of them

  11. #56
    Join Date
    Jun 2003
    Location
    Los Angeles, CA
    Posts
    1,301
    We have a few nodes undergoing RAID rebuilds, if you can PM your IP I can verify with what's going on with your VPS.

  12. #57
    Join Date
    Jul 2010
    Location
    ~/
    Posts
    1,261
    PM sent, let me know when the rebuilds are done (If my node is one of them) and I will run it again.

  13. #58
    Join Date
    Jan 2006
    Location
    Charlotte,NC
    Posts
    138
    One thing I would caution people on are using raw shorterm numbers like these to judge overall performance. Would you have 20MB/s of extremely stable storage on an HA array off the box, or 400MB/s on a Raid0 array? I always recommend people I consult with to take a overall view of performance and availability, and try not to get caught up in benchmarks.
    And to be perfectly honest, real world disk usage doesn't come close to saturating interface like dd is capably of. So I would also recommend understanding what TYPE of storage a host is using, versus what raw scores you can come up with.

    Case in point, it would take 2 and maybe 3 1G fiber drops to hit the half the 346MB/s, but you could fairly easily hit that number with 4-6 consumer grade sata disks in a single server config. But would you rather have a HA fiber storage cluster that can do 80MB/s and support multiple full server failures, and more bandwidth versus pure speed, or 400MB/s that goes completely down when a single server fails. It's a balancing act, and I think it's our job as providers to better educate consumers, if we want to see the industry grow in the right direction.

  14. #59
    Join Date
    Jul 2010
    Location
    ~/
    Posts
    1,261
    Its a fair point you make Linology, I dont know what they all use hardware wise, although from the 2 host website my VPS that got 147MB/s is using SATA II drives in raid 10.

    Is there a query I can run from the VPS that would give me a hint or would that need to be ran on the node itself?

  15. #60
    Join Date
    Jul 2010
    Location
    ~/
    Posts
    1,261
    Quote Originally Posted by PhotonVPS-Jim View Post
    We have a few nodes undergoing RAID rebuilds, if you can PM your IP I can verify with what's going on with your VPS.
    Have you finished the raid rebuild yet?

    Code:
    [root@master ~]# dd if=/dev/zero of=test bs=64k count=4k conv=fdatasync && rm test
    4096+0 records in
    4096+0 records out
    268435456 bytes (268 MB) copied, 17.9879 seconds, 14.9 MB/s

Page 4 of 10 FirstFirst 1234567 ... LastLast

Similar Threads

  1. Tool for stress test , check disk i/o and health of drives in array
    By turbovps in forum Colocation and Data Centers
    Replies: 2
    Last Post: 11-17-2010, 03:47 PM
  2. Replies: 0
    Last Post: 07-22-2010, 12:06 PM
  3. Replies: 15
    Last Post: 02-12-2010, 02:23 PM
  4. 1 Test VPS Only! Windows 2003 VPS - 512MB RAM / 50GB / 100GB Bandwidth
    By A Grateful Dad in forum VPS Hosting Offers
    Replies: 9
    Last Post: 08-22-2009, 09:48 PM
  5. Hard disk test 'surprises' Google.
    By encoderX in forum Computers and Peripherals
    Replies: 1
    Last Post: 02-21-2007, 03:06 AM

Related Posts from theWHIR.com

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •