Howdy, Stranger!

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


IO speed is important for server performance? Hetzner server
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.

IO speed is important for server performance? Hetzner server

rrrrrr Member
edited November 2014 in General

We have got this server today

http://www.hetzner.de/en/hosting/produkte_rootserver/px120ssd

WE got following benchmark I am not satisfied with IO speed what can be issue?

[root@server ~]# sh bench.sh
CPU model : Intel(R) Xeon(R) CPU E5-1650 v2 @ 3.50GHz
Number of cores : 12
CPU frequency : 3500.000 MHz
Total amount of ram : 129166 MB
Total amount of swap : 4095 MB
System uptime : 3 min,
Download speed from CacheFly: 59.4MB/s
Download speed from Coloat, Atlanta GA: 13.8MB/s
Download speed from Softlayer, Dallas, TX: 13.6MB/s
Download speed from Linode, Tokyo, JP: 5.88MB/s
Download speed from i3d.net, Rotterdam, NL: 61.7MB/s
Download speed from Leaseweb, Haarlem, NL: 60.8MB/s
Download speed from Softlayer, Singapore: 3.21MB/s
Download speed from Softlayer, Seattle, WA: 13.6MB/s
Download speed from Softlayer, San Jose, CA: 11.1MB/s
Download speed from Softlayer, Washington, DC: 21.1MB/s
I/O speed : 249 MB/s

[root@server ~]# dd if=/dev/zero of=sb-io-test bs=1M count=1k conv=fdatasync
1024+0 records in
1024+0 records out
1073741824 bytes (1.1 GB) copied, 4.26972 s, 251 MB/s

Instead smaller server is giving better performance...
http://www.hetzner.de/en/hosting/produkte_rootserver/ex40ssd

This server gave us 440MB/s IO..

Let us know which server we should keep.

Comments

  • NeoonNeoon Community Contributor, Veteran

    is the sync still running?

  • What SSD models are these? hdparm -i /dev/sda

  • @rds100

    /dev/sda:

    Model=SAMSUNG MZ7GE240HMGR-00003, FwRev=EXT0303Q, SerialNo=S1M9NYAF800312
    Config={ Fixed }
    RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=0
    BuffType=unknown, BuffSize=unknown, MaxMultSect=16, MultSect=off
    CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=468862128
    IORDY=on/off, tPIO={min:120,w/IORDY:120}, tDMA={min:120,rec:120}
    PIO modes: pio0 pio1 pio2 pio3 pio4
    DMA modes: mdma0 mdma1 mdma2
    UDMA modes: udma0 udma1 udma2 udma3 udma4 udma5 *udma6
    AdvancedPM=no WriteCache=enabled
    Drive conforms to: unknown: ATA/ATAPI-2,3,4,5,6,7

    • signifies the current active mode
  • @Infinity580 said:
    is the sync still running?

    Not sure what you mean

  • @rrr and on the other server? In the Hetzner's links you posted the two servers are advertised with different SSDs - one with "datacenter SSDs" and the other one doesn't mention datacenter, just SSDs.

  • @rds100

    Its intel i think

    /dev/sda:

    Model=INTEL SSDSC2BW240A4, FwRev=DC12, SerialNo=BTDA330401V22403GN
    Config={ Fixed }
    RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=0
    BuffType=unknown, BuffSize=unknown, MaxMultSect=16, MultSect=16
    CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=468862128
    IORDY=on/off, tPIO={min:120,w/IORDY:120}, tDMA={min:120,rec:120}
    PIO modes: pio0 pio3 pio4
    DMA modes: mdma0 mdma1 mdma2
    UDMA modes: udma0 udma1 udma2 udma3 udma4 udma5 *udma6
    AdvancedPM=yes: unknown setting WriteCache=enabled
    Drive conforms to: unknown: ATA/ATAPI-2,3,4,5,6,7

    • signifies the current active mode
  • The intel one is faster in your tests, right? It has a controller which compresses the data before writing to the flash, so zero writes there are faster.

  • @rds100 ok so which server we should keep? First one is bigger server with 128gb ram and 480SSD second one is I7 and 32GB ram but has good IO..

    We need faster server thats our main goal.

  • @rrr run a more meaningful benchmark that is more relevant to your workload. For instance if you are going to run it for a database server, run a database benchmark. dd test is mostly fake, in the real world nobody is writing many zeroes to the SSDs.

  • @rds100 We will host mainly WordPress websites. so what sort of test I should run? How can I run database benchmark?

  • This one should be better for wordress sites - http://www.hetzner.de/en/hosting/produkte_rootserver/px120ssd

    However it is twice the cost of the other one, so you have to decide in the end :)

  • raindog308raindog308 Administrator, Veteran

    rrr said: We will host mainly WordPress websites. so what sort of test I should run? How can I run database benchmark?

    http://lmgtfy.com/?q=database+benchmark+script+mysql

    If you're not willing to put forth minimal effort, you shouldn't be trying to run a server.

  • @raindog308 Thank you very much for your so kind words :) Guess what that what I am already doing it! I am newbie for sure but I am not dumb :) When I work I do all my best. I give my 200% Its just that I dont know much about CPU etc and how to measure performance. I always search Google first then post in forums :) I really like your efforts in provide me your solutions :) thanks. No offense :) Peace

    http://goo.gl/aty8sV This is what I have been doing :)

  • @rds100 yes cost is no issue but performance is. I will take your words and we will keep that PX120 SSD. Yes I do like that it has powerful CPU and more RAM. Thanks for your help. I really appreciate it!

  • If it's SSD don't worry about the disk speed, as long as you have healthy disks. If you have them in RAID the disks might still be syncing data to each other, so you will get slower benchmarks.

  • @rrr those DD numbers really mean nothing in real world scenario. Unfortunately this forum has loadsof DD benchmark sluts(including some providers) who will try to misguide you as much as possible so that they can make you believe that your server is good for nothing & you should get a server from insert any hosts name who sells DD porn but at the end it depends on your requirement if you dont need that much ram, just get the one which is cheaper as it was really not going to make much difference if expensive server was better in benchmark results.

Sign In or Register to comment.