Results 1 to 10 of 13

Thread: Multipath Doubts

Threaded View

  1. #1

    Multipath Doubts

    Hi all.

    I have a cluster with 5 nodes.
    Servers: Dell PowerEdge R710
    HBAs: 2xLpe12000 each
    Switch: Brocade 48000B
    Disk Array: EMC Clariion CX4
    Multipath-tools: v0.4.8 (08/02, 2007)
    (This solution is certified by EMC).

    1) Can you post your multipath.conf file?

    2) I think Novell recommends using descriptive names for the paths.
    Should I put the names in the multipath section of multipath.conf
    file as alias, or in /var/lib/multipath/bindings, or not use that and use the
    defaults mpath<letter>?

    3) Should I use "prio emc"?
    man says default is none.
    Red Hat documentation says is const.
    multipath -t says its const.

    4) Why "multipath -t" says I have polling_interval 0, when I didn't
    change the default multipath.conf.
    What does 0 means?

    5) path_grouping_policy
    man says the default is multibus.
    Red hat documentation says it's failover.
    I should have multibus correct?

    6) path_checker
    Should I use emc_clariion, or the default?

    7) Should I change dev_loss_tmo, fast_io_fail_tmo and async_timeout to
    smaller values?
    If I do that what values do you recommend?
    Should I still lower the timeout of the HBA, or by doing that the
    timeout in the HBA can continue in 30 seconds?
    What is the timeout that you have for yours HBAs?
    (I have Emulex lpe12002 so in /etc/modprobe.conf I have the following
    options lpfc lpfc_nodev_tmo=3 lpfc_devloss_tmo=3)

    8) What other recommendations/suggestions do you have for a correct configuration?

    Thanks in advance, TA.
    Last edited by ta; 02-Nov-2011 at 05:02 PM.

Tags for this Thread

Posting Permissions

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