Visit Open-E website
Results 1 to 2 of 2

Thread: NFS performance problems

  1. #1

    Default NFS performance problems

    Hi all,

    we are experiencing severe NFS performance problems. Small files seem to work OK, but the server becomes irresponssive when we write large files from the local disks to the NFS mount on the storage server, up to a point where we cannot even do a nomal ls.

    the NFS storage is mounted in fstab with:
    192.168.10.220:/data /mnt/data nfs rsize=9700,wsize=9700,hard,intr,async,proto=tcp,po rt=2049,timeo=5,retrans=3 0 0

    network connection consists of two 10 Gbit cards in bond mode:

    auto bond0
    iface bond0 inet static
    address 192.168.10.110
    netmask 255.255.255.0
    bond-mode 802.3ad
    bond-miimon 1000
    bond-lacp-rate 1
    bond-slaves eth2 eth3
    mtu 9700

    server box: 32 core xeon / 512 GB memory and 1.7 GB internal SSDs
    open-e box: 8 core xeon / 32 GB memory and 12x4TB sata disks on LSI controler

    any ideas how to tackle this?

  2. #2

    Default

    Make sure your on the latest DSS V7 w/ build 10529.

    Test with the mount command with fallowing options :
    -o rsize=524288,wsize=524288

    Also make sure this is check in the GUI in the NFS section:

    no root squash
    insecure
    insecure locks


    If there are many NFS connection consider increasing the number of nfsd threads. This is located under Hardware console tools (ALT+CTRL+W) -> Tuning options -> NFS Daemon tuning (default is 64) try higher.

    The MTU value for the 10GbE should 9000 or what the vendor recommends and make sure that all are on the same MTU values (DSS V7 and switches).
    All the best,

    Todd Maxwell


    Follow the red "E"
    Facebook | Twitter | YouTube

Posting Permissions

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