VCS setting low nice values for managed resources - Veritas Cluster Server

This is a discussion on VCS setting low nice values for managed resources - Veritas Cluster Server ; Recently, one of our managed web hosting customers logged a trouble ticket complaining about poor system performance on their VCS-managed dynamic failover cluster. They are running an Oracle 8i (8.1.6) database on VCS 1.1.2. Upon examination of the system, it ...

+ Reply to Thread
Results 1 to 2 of 2

Thread: VCS setting low nice values for managed resources

  1. VCS setting low nice values for managed resources


    Recently, one of our managed web hosting customers logged a
    trouble ticket complaining about poor system performance on
    their VCS-managed dynamic failover cluster. They are running an
    Oracle 8i (8.1.6) database on VCS 1.1.2.

    Upon examination of the system, it was discovered that Oracle
    and other VCS-managed resources were running with a System V
    nice value of 4 (equivalent to -16 on BSD).

    After making this discovery, I examined other similar clusters
    and found the same issue repeatedly.

    Is this a known issue? Is there a way to configure VCS-managed
    resources to run with a normal nice value.

    Thanks.

    Chuck Wolfe
    Senior Oracle DBA

  2. Re: VCS setting low nice values for managed resources


    Chuck,

    This is an attribute that was changed in the 1.3 release. You now have the

    capability to set scheduling class and priorites. See the Users Guide
    for VCS 1.3.

    Chris


    "Chuck Wolfe" wrote:
    >
    >Recently, one of our managed web hosting customers logged a
    >trouble ticket complaining about poor system performance on
    >their VCS-managed dynamic failover cluster. They are running an
    >Oracle 8i (8.1.6) database on VCS 1.1.2.
    >
    >Upon examination of the system, it was discovered that Oracle
    >and other VCS-managed resources were running with a System V
    >nice value of 4 (equivalent to -16 on BSD).
    >
    >After making this discovery, I examined other similar clusters
    >and found the same issue repeatedly.
    >
    >Is this a known issue? Is there a way to configure VCS-managed
    >resources to run with a normal nice value.
    >
    >Thanks.
    >
    >Chuck Wolfe
    >Senior Oracle DBA



+ Reply to Thread