Slow transfer speeds

  • So I recently purchased a ASUS H87I-PLUS for my new NAS build (rest of the hardware here: http://ca.pcpartpicker.com/p/2kbJTW)


    I've been trying to transfer files to a share and can't hit anymore than ~40MB/s for speed


    I have tried this with both the regular and backport kernel and I see the same results


    I've looked in the support info and it states my speed is 1000Mb/s so it should be full gigabit (yes, I have a gigabit switch and nothing is wireless). I expect there to be overheard, but 40MB/s seems really low even when taking that into account


    When I was testing out OMV on my old computer I saw the same transfer speeds (40MB/s). However, when I finally turned that machine back into a windows box transfer speeds went up to ~100MB/s. So it doesn't appear to be a hardware issue (at least not with the old machine).


    Is there a way to speed up transfers in OMV?

    • Offizieller Beitrag

    The new system should easily be able to saturate gigabit networking. What size of files are you transferring? What OS is the client?

    omv 7.0.4-2 sandworm | 64 bit | 6.5 proxmox kernel

    plugins :: omvextrasorg 7.0 | kvm 7.0.10 | compose 7.1.2 | k8s 7.0-6 | cputemp 7.0 | mergerfs 7.0.3


    omv-extras.org plugins source code and issue tracker - github


    Please try ctrl-shift-R and read this before posting a question.

    Please put your OMV system details in your signature.
    Please don't PM for support... Too many PMs!

  • Same speed among the win8.1 clients? Did you check out the transfer speed with FTP?

    Homebox: Bitfenix Prodigy Case, ASUS E45M1-I DELUXE ITX, 8GB RAM, 5x 4TB HGST Raid-5 Data, 1x 320GB 2,5" WD Bootdrive via eSATA from the backside
    Companybox 1: Standard Midi-Tower, Intel S3420 MoBo, Xeon 3450 CPU, 16GB RAM, 5x 2TB Seagate Data, 1x 80GB Samsung Bootdrive - testing for iSCSI to ESXi-Hosts
    Companybox 2: 19" Rackservercase 4HE, Intel S975XBX2 MoBo, C2D@2200MHz, 8GB RAM, HP P212 Raidcontroller, 4x 1TB WD Raid-0 Data, 80GB Samsung Bootdrive, Intel 1000Pro DualPort (Bonded in a VLAN) - Temp-NFS-storage for ESXi-Hosts

  • FTP shares will be stored inside the media/<raid-id>/ folder, so copying files to the FTP server inside your OMV box will show how fast that will be.
    But you cannot reach /root or the boot disk with FTP, that needs an WinSCP connection like subzero said and with WinSCP you can check both targets.

    Homebox: Bitfenix Prodigy Case, ASUS E45M1-I DELUXE ITX, 8GB RAM, 5x 4TB HGST Raid-5 Data, 1x 320GB 2,5" WD Bootdrive via eSATA from the backside
    Companybox 1: Standard Midi-Tower, Intel S3420 MoBo, Xeon 3450 CPU, 16GB RAM, 5x 2TB Seagate Data, 1x 80GB Samsung Bootdrive - testing for iSCSI to ESXi-Hosts
    Companybox 2: 19" Rackservercase 4HE, Intel S975XBX2 MoBo, C2D@2200MHz, 8GB RAM, HP P212 Raidcontroller, 4x 1TB WD Raid-0 Data, 80GB Samsung Bootdrive, Intel 1000Pro DualPort (Bonded in a VLAN) - Temp-NFS-storage for ESXi-Hosts

  • so I tried with FTP: transfer rate settles at around 38MB/s (although at the beginning it was spiking up over 100MB/s but I don't know if thats just an error)


    also tried with winSCP to both the system drive (/opt) and the data drive (/media/..). Speed is the same to both drives, but is slower than with FTP or through windows at around 19MB/s

  • That's rather slow. SCP is always slower due to the encryption but FTP should usually transfer faster than SMB.


    Have you do internal read write tests for the disks?


    What about that?
    You can use pv to check speed when copying files (apt-get install pv) and copy a large file from the raid to the system disk using for exampe pv test.mp4 > /tmp/test.mp4
    I just tried that and it copied a file from my raid5 to the system disk with an average transfer speed of about 98mb/s (The system disk is rather slow). Copying inside the raid shows an average speed of about 170mb/s.

    Homebox: Bitfenix Prodigy Case, ASUS E45M1-I DELUXE ITX, 8GB RAM, 5x 4TB HGST Raid-5 Data, 1x 320GB 2,5" WD Bootdrive via eSATA from the backside
    Companybox 1: Standard Midi-Tower, Intel S3420 MoBo, Xeon 3450 CPU, 16GB RAM, 5x 2TB Seagate Data, 1x 80GB Samsung Bootdrive - testing for iSCSI to ESXi-Hosts
    Companybox 2: 19" Rackservercase 4HE, Intel S975XBX2 MoBo, C2D@2200MHz, 8GB RAM, HP P212 Raidcontroller, 4x 1TB WD Raid-0 Data, 80GB Samsung Bootdrive, Intel 1000Pro DualPort (Bonded in a VLAN) - Temp-NFS-storage for ESXi-Hosts

    • Offizieller Beitrag
    Code
    root@nazbox:~# hdparm -Tt /dev/md1
    
    
    /dev/md1:
     Timing cached reads:   5246 MB in  2.00 seconds = 2623.28 MB/sec
     Timing buffered disk reads: 374 MB in  3.01 seconds = 124.40 MB/sec


    This is read test in my machine, raid1. Almost the nic bandwidth.


    dd can give you write speeds like


    Code
    root@nazbox:~# dd if=/dev/zero of=/media/46e48dd3-8f8a-42f4-a503-0084368be19e/output.img bs=3k count=256k conv=fsync
    262144+0 records in
    262144+0 records out
    805306368 bytes (805 MB) copied, 11.5858 s, 69.5 MB/
  • BTW I have a realtek nic, I can rarely pass 60MB/s with the wire, but i don't care. Everything is connected by wifi.But I guess with your intel nic you should be reaching 80-90MB/s easily


    A realtek nic does not automatically means slower transfer speed. Which type?
    The Asus ITX mobo in my home box has a realtek nic, too - 8111 type. And with big files it always hits the 100mb/s barrier.

    Homebox: Bitfenix Prodigy Case, ASUS E45M1-I DELUXE ITX, 8GB RAM, 5x 4TB HGST Raid-5 Data, 1x 320GB 2,5" WD Bootdrive via eSATA from the backside
    Companybox 1: Standard Midi-Tower, Intel S3420 MoBo, Xeon 3450 CPU, 16GB RAM, 5x 2TB Seagate Data, 1x 80GB Samsung Bootdrive - testing for iSCSI to ESXi-Hosts
    Companybox 2: 19" Rackservercase 4HE, Intel S975XBX2 MoBo, C2D@2200MHz, 8GB RAM, HP P212 Raidcontroller, 4x 1TB WD Raid-0 Data, 80GB Samsung Bootdrive, Intel 1000Pro DualPort (Bonded in a VLAN) - Temp-NFS-storage for ESXi-Hosts

  • so it seems that different files will transfer at different speeds. some at 40MB/s, others at 100MB/s. Is it purely file size that will cause this inconsistency? Is that a normal amount of variation?

    • Offizieller Beitrag

    Small files will always transfer slower because the transfer has to start and stop so many times (for each file). So 40 MB/s for small files and 100 MB/s for large files seems normal to me.

    omv 7.0.4-2 sandworm | 64 bit | 6.5 proxmox kernel

    plugins :: omvextrasorg 7.0 | kvm 7.0.10 | compose 7.1.2 | k8s 7.0-6 | cputemp 7.0 | mergerfs 7.0.3


    omv-extras.org plugins source code and issue tracker - github


    Please try ctrl-shift-R and read this before posting a question.

    Please put your OMV system details in your signature.
    Please don't PM for support... Too many PMs!

Jetzt mitmachen!

Sie haben noch kein Benutzerkonto auf unserer Seite? Registrieren Sie sich kostenlos und nehmen Sie an unserer Community teil!