Issues with Backing up DAG Exchange 2010

 Hi All,


I'm coming over from ArcServe and have been trying to setup our Exchange Jobs in NVBU, but I'm struggling.


Our Exchange environment:

Location A:

 2 CAS Servers

 2 DB Servers

Location B:

  2 CAS Servers

  2 DB Servers

All 4 DB servers are in a DAG.


We have a NVBU server and library at each location.

The plan was to backup the non-active DBs in each location.  So, I setup a Virtual Client in each location comprised of the two DB servers at each location. 

When creating the Backup Selection Set I often get "Error: Failed to get connection to plugin".  I get this error often, but then randomly it'll work.  Although, when I check Access of the Virtual Client it passes.  Then when I check access and firewall access for each client (DB Servers) that's a member of the Virtual Client it passes as well.


EDIT: If I change the Virtual Client IP to the other DB server that's listed as it's member.  It'll work.  I'm able to expand the DAG and see the DBs to select what I want to backup.  But a few days later it'll stop working again and I'll have to flip to the other DB if I want to view the backup selection. 

The exchange backup jobs fail regularly as well. 


Any helpful hints or advice is greatly appreciated. 

  • Hi Paul,
    When it fails with "Failed to get connection to the plugin" - how long does this usually spin before timing out, would you say its about 2 minutes?
    Also, is the DAG configured with an Administrative access point, or is this a IP-less/CNO DAG? By this I mean, is it using a cluster IP or an IP of one of the DAG Members? From the description, it sounds like it might be the latter. If this is the case, then you would have to update the virtual clients IP each time one of those servers loses connection.
    This might be a question better suited to have an Service Request open for, so if you would like to do that, please visit the below link to get a service request started.
  • In reply to Andre.Cooley:

    Hi Andre,

    Thanks for the reply.

    Yes, it gets stuck for a minute or two before it errors out. Then if I switch the virtual client's ip to the other DB member it connects. But only for a few days and then it stops working and I'll swap it again.

    The DAG does have a cluster ip. What I'm doing (or trying to) is splitting up each location into two separate DAGs with each location consisting of 2 DB servers. That's probably my issue. It sounds like this is not possible. Instead I'll probably have to use the DAG cluster IP for the virtual client and then add all 4 DB servers from both locations as members of it.

    I'll raise a service request.
  • In reply to paul.tate:

    Yes, it sounds like you might be right. If they all belong to the same DAG, then they should ideally reside under the one Virtual Cluster node. What you are trying to do could work, but the nature of it, since the actives are switching around, the virtual client would not be consistent unless you constantly updated each time. You may not have to raise a service request for this, but we can have one created anyways in case we need to investigate further.
    If you want, just send me a personal message here with your contact and account info and I can have one created, or have one created through the portal and just request for me in the desription.