Forum Discussion

mokkan's avatar
mokkan
Level 6
11 years ago

Proxy Vs Phantom

I am trying to understand how Proxy and Phanton works.  From  my understanding , if we want to use NIC resources in diffrent SGs, You create a dedicated SG for the NIC and use that Resource in other applications as dependency.  Am I right? 

 

How about Phantom works?

 

 

  • Basic rules of a service group is that you must be able to offline it and online it, and this is not possible if you don't have any OnOff resources in the group, in which case you could add a dummy "FileOnOff" resource, but the Phantom resource was designed for this.

    Also note, never try to manually offline or online the Phantom resource - always offline or online the group.  I has a customer who tried to offline the Phantom resource and it hung VCS, so had to force stop the cluster.

    Mike

  • Hi,

    Not sure if someone can share a real scenario from any customer, but based on above explanations, its pretty clear on when are you going to use each resource type. VCS Bundled agents guide talks about typical examples & depictions

    1. Proxy resource -  If you have limited network cards in your cluster nodes, you would want to configure NIC resource in 1 group or multinic resources in parallel service group while configuring proxy resource in other service groups on same cluster. Typical example & depiction is shown here

    https://sort.symantec.com/public/documents/sfha/6.1/solaris/productguides/html/vcs_bundled_agents/ch06s03s01.htm

    2. Phantom resource - as explained above, you would need a phantom resource to help VCS to understand the status of any group which only contains persistent resource like NIC. Phantom is used mostly in parallel service groups only. Here is a typical scenario of phantom being used

    https://sort.symantec.com/public/documents/sfha/6.1/solaris/productguides/html/vcs_bundled_agents/ch06s04s01.htm

     

    G

     

    G

7 Replies

  • Yes, as you mentioned the Proxy resource allows multiple service groups to monitor the same network interface. This reduces the network traffic that would result from having multiple NIC resources in different service groups monitoring the same interface. Pl. see: http://www.symantec.com/business/support/index?page=content&id=TECH6205

    Phantom: A VCS SG shows an online status only when all of its nonpersistent resources are online. Therefore, if a parallel service group has only persistent resources (or None type resources like network interface --None type means VCS has no control over either starting or stopping), VCS considers the group offline, even if the persistent resources are running properly. By using a phantom resource, the status of the service group is shown as ONLINE. Use Phantom resouce only for parallel SGs..

  •  

    Hi, Phantom  only for  parallel  sgs.

      and it help vcs  determine the status of  the sg.

     

      Sometimes, with only   None type resource, and OnOnly resource in  sg,  vcs  can't assess the state of the sg.

     

  • Basic rules of a service group is that you must be able to offline it and online it, and this is not possible if you don't have any OnOff resources in the group, in which case you could add a dummy "FileOnOff" resource, but the Phantom resource was designed for this.

    Also note, never try to manually offline or online the Phantom resource - always offline or online the group.  I has a customer who tried to offline the Phantom resource and it hung VCS, so had to force stop the cluster.

    Mike

  • Thank you very much for your explanation. Can you please provide an example where people use in real senario or in production?

  • Hi,

    Not sure if someone can share a real scenario from any customer, but based on above explanations, its pretty clear on when are you going to use each resource type. VCS Bundled agents guide talks about typical examples & depictions

    1. Proxy resource -  If you have limited network cards in your cluster nodes, you would want to configure NIC resource in 1 group or multinic resources in parallel service group while configuring proxy resource in other service groups on same cluster. Typical example & depiction is shown here

    https://sort.symantec.com/public/documents/sfha/6.1/solaris/productguides/html/vcs_bundled_agents/ch06s03s01.htm

    2. Phantom resource - as explained above, you would need a phantom resource to help VCS to understand the status of any group which only contains persistent resource like NIC. Phantom is used mostly in parallel service groups only. Here is a typical scenario of phantom being used

    https://sort.symantec.com/public/documents/sfha/6.1/solaris/productguides/html/vcs_bundled_agents/ch06s04s01.htm

     

    G

     

    G

  • Thank you G.  Why can' we use proxy resrource to find the satus of SG instead of phantom?

     

     

  • A Proxy resource is a MonitorOnly resource and as in my last post 

    Basic rules of a service group is that you must be able to offline it and online it, and this is not possible if you don't have any OnOff resources in the group

     

    So you need to a OnOff resource so adding a Proxy resource does not help, so Phantom is the recommendation, so that you can determine the status of the SG.

    Below is example extract from main.cf:

    group app1-sg (
        SystemList = { r55v61a = 0, r55v61b = 1 }
        AutoStartList = { r55v61a, r55v61b }
        )
    
        IP app1-ip (
            Device = eth0
            Address = "192.168.56.55"
            NetMask = "255.255.255.0"
            )
    
        Proxy app1-pub-pxy (
            TargetResName = eth0-pub
            )
    
        app1-ip requires app1-pub-pxy
    
    group app2-sg (
        SystemList = { r55v61a = 0, r55v61b = 1 }
        AutoStartList = { r55v61a, r55v61b }
        )
    
        IP app2-ip (
            Device = eth0
            Address = "192.168.56.56"
            NetMask = "255.255.255.0"
            )
    
        Proxy app2-pub-pxy (
            TargetResName = eth0-pub
            )
    
        app2-ip requires app2-pub-pxy
    
    group networks (
        SystemList = { r55v61a = 0, r55v61b = 1 }
        AutoFailOver = 0
        Parallel = 1
        AutoStartList = { r55v61a, r55v61b }
        )
    
        NIC eth0-pub (
            Critical = 0
            Device = eth0
            )
    
        Phantom networks-ptm (
            Critical = 0
            )

     

    Mike