Forum Discussion

tunisGharbi's avatar
10 years ago

Cluster GEO integration with VVR oracle database : steps order

I have 2 sites with the same config : 2 servers and storage.

Each cluster have HA failover with Oracle database.

I need to install Cluster GEO VVR to replicate Oracle data.

My question is how to automate this configuration?

Did i test first manually  the VVR then I intagrate with the cluster

Any Idea about the order and steps to automate this configuration

 

Best regrads.

  • This is how I do it

    1. Create replication group in VVR with virtual replication IP, but not RVG resource and online group
    2. Configure VVR (create RDS) with virtual IP configured in VCS
    3. Add RVG resource in replication group 
    4. Add RVGPrimary resource to application group
    5. Configure global cluster
    6. Configure application group as a global service group
    7. Test you can switch global service group locally and across sites

    Just to add you can automate all these steps if you want as they can all be done from the command line - best thing to do is to do manually and then look at /etc/VRTSvcs/conf/config/main.cmd for commands to create GCO objects (remote cluster, heartbeats, ClusterList service group attribute, hares -add/modify))

    Mike

  • Hi,

    In a normal scenario, diskgroup & filesystems will be imported on primary site while secondary site will only have diskgroup imported but you can't mount volumes. If all the VVR related daemons & vxconfigd is working on secondary site, using vradmin commands you could configure VVR without needing to mount filesystem manually.

    For IP address, yes you need to have IP address plumed & working on secondary site in order to communicate with primary site.


    G

9 Replies

  • Hi,

    First thing, do you have license already for VVR/GCO ? I would recommend to install that if in case you don't have it handy.

    Secondly, you will need to configure VVR first where you would need to create replication between two sites.

    There is a lot of documentation available for this, best I would recommend would be to follow below

    https://sort.symantec.com/public/documents/sfha/6.1/solaris/productguides/html/sfhas_dr_impl/ch07s02.htm

    Once you have replication setup, you need to connect the two clusters using GCO, you can find steps here in section "setting up global clusters"

    https://sort.symantec.com/public/documents/sfha/6.1/solaris/productguides/html/vcs_admin/ch16.htm

    When you say automate the steps ? you mean you want to automate the configuration steps ? If thats the ask, I would say I haven't seen automated steps for configuring VVR, you would need to do manually. You can though run GCO Wizard to from Veritas Java console to configure GCO

     

    G

     

  • thanks Gaurav

    License already installed for  VVR/GCO.

    I have already  install VCS on each site with Oracle database SG failover.

    II have to  Configure VVR between the 2 sites.

    Should first test manually the replication then second Install the global Cluster and third  integrate this configuration with GCO?

    Besrt regards.

  • This is how I do it

    1. Create replication group in VVR with virtual replication IP, but not RVG resource and online group
    2. Configure VVR (create RDS) with virtual IP configured in VCS
    3. Add RVG resource in replication group 
    4. Add RVGPrimary resource to application group
    5. Configure global cluster
    6. Configure application group as a global service group
    7. Test you can switch global service group locally and across sites

    Just to add you can automate all these steps if you want as they can all be done from the command line - best thing to do is to do manually and then look at /etc/VRTSvcs/conf/config/main.cmd for commands to create GCO objects (remote cluster, heartbeats, ClusterList service group attribute, hares -add/modify))

    Mike

  • Hi,

    There is no separate software for GCO, its just a GCO configuration wizard.

    The way I prefer is, configure replication between two sites using VVR, (steps were given in link above)

    Once done, test the replication & ensure your both sites are in sync, if you have window available, would prefer to test the DR before automating the failover in cluster. Once you are sure that your application works on primary & DR both, you can run GCO wizard & configure GCO between the two sites. This wil add global service groups which are failover capable across sites. As you would have tested the DR before this, you would be confident that application wuold work on DR site.

     

    G

  • thanks Gaurav

    thanks Mike

     

    In fact i have no choice in my case!

    I work with oracle administrator  and application administrator.

    i have to test on each site the HA failover for application and oracle then I will work a lone with relplication.

    I can disable the cluster ofter sucees test and mount manually file system and IP virtual to install/configure the replication

  • Hi,

    In a normal scenario, diskgroup & filesystems will be imported on primary site while secondary site will only have diskgroup imported but you can't mount volumes. If all the VVR related daemons & vxconfigd is working on secondary site, using vradmin commands you could configure VVR without needing to mount filesystem manually.

    For IP address, yes you need to have IP address plumed & working on secondary site in order to communicate with primary site.


    G

  • I am afraid there is no such docs which will have screenshot wise steps ... however if you look at VVR admin guide, or refer to advance cluster section in VCS admin guide, you should find lot of important steps well documented...

    Link is there in Mike's response above for SFHAS guide

     

    G