VCS Failover fail in SAN environment when all FC cables detached - Veritas Cluster Server

This is a discussion on VCS Failover fail in SAN environment when all FC cables detached - Veritas Cluster Server ; hi.. Any one have the following similar situation, please help? I configured a VCS 1.3 4-node cluster at customer site. All machines are Solaris8 with Recommend8 patch, VxFS 3.4 with latest patch, VxVM3.1 with latest patches(VxVM3.1.1 is also tested in ...

+ Reply to Thread
Results 1 to 3 of 3

Thread: VCS Failover fail in SAN environment when all FC cables detached

  1. VCS Failover fail in SAN environment when all FC cables detached

    hi..

    Any one have the following similar situation, please help?

    I configured a VCS 1.3 4-node cluster at customer site. All machines are
    Solaris8 with Recommend8 patch, VxFS 3.4 with latest patch, VxVM3.1 with
    latest patches(VxVM3.1.1 is also tested in this case), JNI FCI-1063(2 in
    each servers for DMP purpose), upon testing with detach all FC cables in one
    node, the failover was hang. The process was hang in umount the filesystem
    and also deploy the diskgroup, the fault node can't finish the offlining and
    therefore failover fail.
    Any one have same situation? Any one have fix to this problem? Any one have
    ideas to this problem?

    Thanks.

    Mike




  2. Re: VCS Failover fail in SAN environment when all FC cables detached

    I had and *have* the some problem.
    Veritas support said that this kind of situation isn't a failover condition.
    In "Veritas thinking" the connections and the RAS of the Storage Area
    Network
    aren't under the control of VCS.
    The only way to obtain failover is writing script code in the resfault
    trigger.
    I think that in a situation like the one you described, is a VCS duty
    performing a failover
    because perhaps other nodes in the cluster could
    effectively be in different conditions about the disks accessibility,
    and then the Service Group could online again.

    Cheers,
    napo

    "Mike" wrote in message news:3c597b88@hronntp01....
    > hi..
    >
    > Any one have the following similar situation, please help?
    >
    > I configured a VCS 1.3 4-node cluster at customer site. All machines are
    > Solaris8 with Recommend8 patch, VxFS 3.4 with latest patch, VxVM3.1 with
    > latest patches(VxVM3.1.1 is also tested in this case), JNI FCI-1063(2 in
    > each servers for DMP purpose), upon testing with detach all FC cables in

    one
    > node, the failover was hang. The process was hang in umount the filesystem
    > and also deploy the diskgroup, the fault node can't finish the offlining

    and
    > therefore failover fail.
    > Any one have same situation? Any one have fix to this problem? Any one

    have
    > ideas to this problem?
    >
    > Thanks.
    >
    > Mike
    >
    >
    >








  3. Re: VCS Failover fail in SAN environment when all FC cables detached


    In another scenario, is it possible to failover the node to another "SAN"
    if the primary SAN is unavailable?

    "Napo" wrote:
    >I had and *have* the some problem.
    >Veritas support said that this kind of situation isn't a failover condition.
    >In "Veritas thinking" the connections and the RAS of the Storage Area
    >Network
    >aren't under the control of VCS.
    >The only way to obtain failover is writing script code in the resfault
    >trigger.
    >I think that in a situation like the one you described, is a VCS duty
    >performing a failover
    >because perhaps other nodes in the cluster could
    >effectively be in different conditions about the disks accessibility,
    >and then the Service Group could online again.
    >
    >Cheers,
    >napo
    >
    >"Mike" wrote in message news:3c597b88@hronntp01....
    >> hi..
    >>
    >> Any one have the following similar situation, please help?
    >>
    >> I configured a VCS 1.3 4-node cluster at customer site. All machines are
    >> Solaris8 with Recommend8 patch, VxFS 3.4 with latest patch, VxVM3.1 with
    >> latest patches(VxVM3.1.1 is also tested in this case), JNI FCI-1063(2

    in
    >> each servers for DMP purpose), upon testing with detach all FC cables

    in
    >one
    >> node, the failover was hang. The process was hang in umount the filesystem
    >> and also deploy the diskgroup, the fault node can't finish the offlining

    >and
    >> therefore failover fail.
    >> Any one have same situation? Any one have fix to this problem? Any one

    >have
    >> ideas to this problem?
    >>
    >> Thanks.
    >>
    >> Mike
    >>
    >>
    >>

    >
    >
    >
    >
    >
    >



+ Reply to Thread