Active/Active Active/Passive Win2003

Discussion in 'Clustering' started by DHamblin, Jul 30, 2007.

  1. DHamblin

    DHamblin Guest

    We currently have a 2 node Win2003 active/passive cluster that is just a file
    server. On it currently is 1 group (group0) that has the cluster name, IP,
    the two physical volume resources (E:, F:) and the various shares.

    E: and F: are seperate volumes on the SANS.

    Currently its active/passive with one or the others erver driving the whole
    cluster. We would like to have one server drive E: and the other drive F:.
    My understanding is that I would need to just create a new Group1, move the
    f: physical resource and file share resources for F: to it; then set the
    preferred owner to the other server.

    At that point I could failover the second group making the other server the
    owner of F: while the first server still contrls E: and its shares.

    That way the load is split between the two.

    The reason we are wanting to do this E: is 1.2 TB and F: is almost 1TB so by
    spillting them each server has less load.

    Is my logic on how this would work correct?
     
    DHamblin, Jul 30, 2007
    #1
    1. Advertisements

  2. Your Group1 should also have its own IP address and Network Name resources.

    Setting the preferred owner is not necessary unless you're using the
    "failback" feature (or if you had 3+ nodes)...though it doesn't hurt
    anything setting this.

    Regards,
    John

    Visit my blog: http://msmvps.com/blogs/jtoner
     
    John Toner [MVP], Jul 31, 2007
    #2
    1. Advertisements

  3. DHamblin

    DHamblin Guest

    Ok, this then confuses me a bit. If I setup a different name and IP address,
    then does this make a "new" virtual server?

    For example the current virtual server is CPLIBRA with IP 178.16.2.2 (not
    real name or IP). Under it users are mapped to shares that run on E: and F:.

    Today which ever of the two cluster servers that is controlling the virtual
    server CPLIBRA has both E: and F:. I was thinking I could split it and have
    one server do E: and one do F: and the users still see just CPLIBRA to map to.

    I admit to being new at clustering; wouldn't giving it (Group1) its own IP
    and name mean they need to map to another virtual server for the F: shares?

    If there is any documentation you can point me to that would be great as well.


    --
    Thanks,
    Dave Hamblin


     
    DHamblin, Jul 31, 2007
    #3
  4. DHamblin

    Chuck [MSFT] Guest

    To split the 'duties' in the cluster you will have to create another
    resource group with its own Virtual Server and then move the appropriate
    storage into that group. Users will then have to connect to the virtual
    server to access the file shares.

    --
    Chuck Timon, Jr.
    Microsoft Corporation
    Windows Server 2008 Readiness Team
    This posting is provided 'AS IS" with no warranties, and confers no rights.

     
    Chuck [MSFT], Jul 31, 2007
    #4
  5. DHamblin

    DHamblin Guest

    Well, that makes sense (I think :) ).

    I think whats confusing me a bit is the way ours was originally setup
    (named). When you open Cluster admin, the cluster name we connect to is
    CPLIBRA. Under that is the existing Group0 that has resources including IP
    and NAME for the virtual server CPLIBRA (the virtual server and cluster name
    are the same).

    Once I would complete this process; I would still open a connection to
    CPLIBRA cluster as before, only under the Group1 there would be a new server
    name and IP that users would then see that contains F: physical drive and
    share resources.

    My first thought was I would be creating a whole new cluster but I am not,
    correct?

    I found lots of info on setting up a two-node cluster like we have; but very
    little on having multiple groups and how to config them.
     
    DHamblin, Jul 31, 2007
    #5
  6. DHamblin

    Chuck [MSFT] Guest

    Unfortunately you have fallen into the same trap many others have....that
    being putting things in the Cluster Group that should not be there and using
    the Cluster Network Name for things other than Administration.

    Ideally, after the cluster is created you would see the Cluster Group listed
    and it would contain the Cluster Name, Cluster IP and the disk serving as
    the quorum disk. You would just leave that group alone.

    You would also notice other resource groups each containing a physical disk
    resource provided the cluster configuration detected external storage
    properly.

    You would then use those groups to configure you clustered whatevers....in
    this case clustered file shares.

    You can still 'fix' this, it would take some time, but it is 'do-able'.

    --
    Chuck Timon, Jr.
    Microsoft Corporation
    Windows Server 2008 Readiness Team
    This posting is provided 'AS IS" with no warranties, and confers no rights.
     
    Chuck [MSFT], Jul 31, 2007
    #6
  7. DHamblin

    DHamblin Guest

    Actually, I have inherited this. It was orginally put in by Dell/EMC as part
    of a project last year. After the SANS part melted down back in March I
    found that it had not been optimally configured.

    Now working with the Cluster part; I am beginning to get the same feeling
    again.
     
    DHamblin, Jul 31, 2007
    #7
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.