GCO bringing global service group online without force option after a site failure - Veritas Cluster Server

This is a discussion on GCO bringing global service group online without force option after a site failure - Veritas Cluster Server ; IHAC that has 2 1-node clusters and we’re doing some failover tests where we power off the cluster node that owns the service group. The global service group’s AutoFailoverPolicy is set to Manual. When we power off the node that ...

+ Reply to Thread
Results 1 to 2 of 2

Thread: GCO bringing global service group online without force option after a site failure

  1. GCO bringing global service group online without force option after a site failure


    IHAC that has 2 1-node clusters and we’re doing some failover tests where
    we power off the cluster node that owns the service group. The global service
    group’s AutoFailoverPolicy is set to Manual. When we power off the node
    that owns the global service group, no one owns the service group (that’s
    fine and expected). However, when we manually try to online the global service
    group online, it tells us that we have to use hagrp –online –force to get
    the service group to the failover cluster (this is fine too and the client
    fully understands why this is going on). When we use the –force option to
    online the global service group on the failover cluster, it comes up just
    fine (Of course it gets to this state because when the node that owns the
    global service group on the original site goes down via power off, it still
    going to think that it still owns the service group so the onlining of the
    global service group won’t go online without the force option because the
    cluster environment still thinks that the other node on the other cluster
    still owns the service group).

    My client understands why it would react this way. What he wanted to know
    if there’s a way to online the global service group on the node in cluster
    B after a power off on the node in cluster A without using the –force option?
    In particular he wants to know if there was a way to do this other than
    using triggers (because he wants to avoid copying the trigger scripts on
    50 other clusters later on) and keeping the global service group from automatically
    bringing it online to the other site.

    Any suggestions will be helpful.


  2. Re: GCO bringing global service group online without force option after a site failure


    Hy,

    you have the attribute ClusterFailOverPolicy for Service Group : Manual,
    Connected or Auto.
    Set it to Auto, and your service group will be brought online on cluster
    B when cluster A failed.

    Maybe this will help ?

    "Kenny Cho" wrote:
    >
    >IHAC that has 2 1-node clusters and we’re doing some failover tests where
    >we power off the cluster node that owns the service group. The global service
    >group’s AutoFailoverPolicy is set to Manual. When we power off the node
    >that owns the global service group, no one owns the service group (that’s
    >fine and expected). However, when we manually try to online the global

    service
    >group online, it tells us that we have to use hagrp –online –force to get
    >the service group to the failover cluster (this is fine too and the client
    >fully understands why this is going on). When we use the –force option

    to
    >online the global service group on the failover cluster, it comes up just
    >fine (Of course it gets to this state because when the node that owns the
    >global service group on the original site goes down via power off, it still
    >going to think that it still owns the service group so the onlining of the
    >global service group won’t go online without the force option because the
    >cluster environment still thinks that the other node on the other cluster
    >still owns the service group).
    >
    >My client understands why it would react this way. What he wanted to know
    >if there’s a way to online the global service group on the node in cluster
    >B after a power off on the node in cluster A without using the –force option?
    > In particular he wants to know if there was a way to do this other than
    >using triggers (because he wants to avoid copying the trigger scripts on
    >50 other clusters later on) and keeping the global service group from automatically
    >bringing it online to the other site.
    >
    >Any suggestions will be helpful.
    >



+ Reply to Thread