Forum Discussion

KeirL's avatar
KeirL
Level 6
11 years ago

EV Building Block Configurations

Hi

Is there a very good clear document for what the required infrastructure needs are to be able to use Building Blocks for DR. I can find 'bits' of information but not enough to get a clear understanding.

for example, it appears that all the resources need to be shared - which is fine for SQL as I have a separate SQL server and also the vault stores are via cif shares so that should be fine, but do Indexes need to be a shared resource too? My indexes are on fast SAN disk (as indexes on cifs isn't recommneded) presented to each EV server so how does this bit work? 

kind regards

KL

Enterprise Vault
Information Governance
Windows Server (2003-2008)
  • GertjanA's avatar
    GertjanA
    11 years ago

    Keir,

    We're reasonably large (I think), and use Robocopy to copy indexes to the DR server. We have a 3 to 1 (3 prod, 1 dr) server, and use mountpoints for the indexes. this way we have on the prod servers a folder called:

    f:\enterprise Vault\EVSRV001\Indexes\Index-VMP01\index1 (etc)

    f:\enterprise vault\EVSRV002\indexes\index-VMP01\index1 (etc)

    These are replicated to the DR-server, which then has the appropriate folders. When performing a DR, as the folder structure is unique, there is no issue.

    We also have environment where they use SAN-disks, which has san-replication obviously. I think the main thing is that IOPS is ok.

  • KeirL's avatar
    KeirL
    Level 6
    11 years ago

    Thanks Rob

    Yeah - I saw those two - 

    But I think the show stopper for me is that EV indexes need to be on shared storage (presumably CIFs) as sharing block storage would need some kind of clustering overlay.

    So building blocks are only for very small environments where the customer is happy to go against Symantec recommendations and put the EV indexes on network storage?

    Seems a bit bizzare that Symantec off building blocks at all in my view then?

    Am I missing something :o)  ?

     

    cheers

  • GertjanA's avatar
    GertjanA
    Moderator
    11 years ago

    Keir,

    We're reasonably large (I think), and use Robocopy to copy indexes to the DR server. We have a 3 to 1 (3 prod, 1 dr) server, and use mountpoints for the indexes. this way we have on the prod servers a folder called:

    f:\enterprise Vault\EVSRV001\Indexes\Index-VMP01\index1 (etc)

    f:\enterprise vault\EVSRV002\indexes\index-VMP01\index1 (etc)

    These are replicated to the DR-server, which then has the appropriate folders. When performing a DR, as the folder structure is unique, there is no issue.

    We also have environment where they use SAN-disks, which has san-replication obviously. I think the main thing is that IOPS is ok.

  • KeirL's avatar
    KeirL
    Level 6
    11 years ago

    Thanks Gertjan

    Can you just clarify how you do your DR?

    Are you saying you use 'building blocks' for DR and use robocopy to replicate your indexes? Such that your production indexes are on fast storage but robocopy them to NAS so the building block approach can work?

    how does this work when you run the 'update service locations' command?

    sorry for so many questions, but it very useful for me to understand what others do to achieve HA and DR solutions.

    kind regards

    Keir

  • John_Santana's avatar
    John_Santana
    Level 6
    11 years ago

    Yes, I'm interested to know if there is any best practice guidance in building the EV for DR site.