Forum Discussion

Riadh_R17's avatar
Riadh_R17
Level 4
2 months ago

pairing 2 flex appliances for replication failed

Hi all 

We have 2 flex appliances and while configuring pairing we get two failed jobs on the monitor 

1st job error message : failed to create high availability replication link

2nd job error message  : failed to configure replication service

anyone can help please ?

knowing that the host0 is on the management network

and we configured a bond with two interfaces for replication on a replication network

when starting pairing we used the bond for replication and we gave an address in that network (on dr appliance)

when finnishing pairing (on primary appliance) we used also the same bond configured for replication and we gave an ip in that network

 

  • kiyan's avatar
    kiyan
    Level 6
    2 months ago

    I think you should go through the what L_BR mentioned in the above.
    just check with your network and security team about the network ,vlan, ports (directions), this is a straight forward configuration from Netbackup point.
    also give it a try with one interface in the bond, check if that will work or no.

    • Riadh_R17's avatar
      Riadh_R17
      Level 4
      2 months ago

      what do we mean by the node ip address in the veritas doc below  between ""

      "Make sure that the following prerequisites are met before you begin.

      • The node IP addresses of each appliance must be able to communicate with the console IP address of the other appliance"

       

      do we need an interface other than the replication interfaces for pairing ?

      Thanks

       

       

      • davidmoline's avatar
        davidmoline
        Level 6
        2 months ago

        Hi Riadh_R17 

        For the console or management interface, each flex will have a node IP address and a virtual IP address that is used to access the web console (the console IP is virtual so it can switch between nodes in a FlexHA pair). The requirement is that the node IP on one Flex must be able to communicate with the console on the second (and the second node IP can communicate with the first console IP). From what you have  previously described this seems like the  it should be the case. 

        You need the replication interface created on one of the data NICs (10/25GE) which you configure when you set up pairing, but for replication to function the nodes/console need to be able to communicate as well. 

        Cheers

  • L_BR's avatar
    L_BR
    Level 5
    2 months ago

    While I have had some issues understanding Veritas Admin guides, have you followed this? https://origin-www.veritas.com/content/support/en_US/doc/130821112-160491193-0/v162494716-160491193

    Specifically 

    • You must have a network interface configured to use for replication. Veritas recommends that you do not select a shared interface or an interface that is configured with the same subnet as the appliance node IP addresses.

    I have recently (perhaps 5 weeks ago) configured 5250s with replication and ran into no issues. 

  • davidmoline's avatar
    davidmoline
    Level 6
    2 months ago

    Hi Riadh_R17 

    Firstly - to answer you question "anyone can help please ?" Yes - Veritas support can provide the help you need.

    Can you provide any additional details on the error received and at what stage of the paring process this occurs? 

    What version FlexOS are you using?

    Cheers

    • davidmoline's avatar
      davidmoline
      Level 6
      2 months ago

      Further - is local node able to communicate with the console IP address of the remote Flex? And is there any network controls to restrict port access etc. between the two Flex appliances? 

      • Riadh_R17's avatar
        Riadh_R17
        Level 4
        2 months ago

        Thank you for your answer.

        what do you mean by the local node? do you mean the console ip address of the local appliance or its replication ip address? 

        two appliance console IPs are in the same subnetwork

        but the replication ip address that are given in the pairing process are not in the same subnetwork and does not communicate with console IPs