UNSOLVED

MarioK1

updated

17 years ago

M

MarioK1

19 Posts

0

1270

January 26th, 2010 05:00

Disaster recovery on remote site from clones???

Hi I have a question regarding Networker architecture and recovery process.

There are two sites with NW server on primary and NW Storage node on remote location.

NW Server and clients - 7.5.1., OS - Windows 2003 Server SP2.

There are two tape libraries and two storage arrays on both sites and synchronous remote replication between storages. We are making regular backup on primary site and clone on secondary site via storage node. The NW server is installed on LUN over the storage array. This LUN is synchronously replicated on remote system.

My question is in case of disaster on main site how to enable NW server on secondary site? I wouldn't like to scan all volumes and recover indexes, having in mind that they are already on the storage array LUN. Any ideas how we can do this in smarter way? Thanks!

Best regards...

  • 524

    0

    Posted February 11th, 2010 04:00

    Hi MarioK

    I'm trying to set up something similar myself for DR purposes (hot site)

    My understanding of this scenario would be that, as you are replicating to the storage node on the remote site

    The SSID's and indexes would be available at that site, (via WAN)

    I would imagine that in the event of a disaster you would re-licence your storage node to be a NW server.

    So you should be back in business almost straight away, cant see any reason why the (clone) server would lose its media index.

    Also, you are replicating the lun of the networker server (accross WAN ?) to your second site.

    This gives you belt and braces replication of the media index but I think still you will have the issue of re-licencing.

    How much of your bandwidth does the clone save set consume ?

    Any advice, links etc you could provide would be very helpfull, feel free to PM me.

  • Jiju1

    181 Posts

    524

    0

    Posted February 11th, 2010 10:00

    HI,

             Is like , you would like to restore specific save sets or need to Dump the whole tape int he remote site .

    Rajesh S

  • ble1

    6 Operator

    •

    14354 Posts

    •

    56186 Points

    524

    0

    Posted February 16th, 2010 05:00

    Why not place the cluster between the sites to manage it?
  • MarioK1

    19 Posts

    524

    0

    Posted September 9th, 2010 01:00

    Hi friends,

    Sorry for delay but I would like to inform you about the solution which we made.

    The cluster is a good idea but the customer doesn’t accept it. So we made a little complicated solution.

    Two sites. Distance is 30 km. Connection between sites is dark fiber with good performance:

    Primary:

    NW server + clients + tape library (TL1)

    Installed directory for NW server is X:\nsr on storage array.

    This LUN (X:\) is replicated synchronous to remote site.

    Remote site:

    NW Storage node + clients + tape library (TL2)

    The solution:

    All clients on primary site are backup via NW server (TL1) and after that the tapes are cloned on remote site via Storage node (TL2).

    The same is on the remote site.

    So we have two copies of the data one on primary and one on remote site.

    Both NW Server and NW SN have configured two libraries TL1 and TL2.

    We make a DR scenario in case of total crash on Primary site.

    The NW install LUN (X:\) is presented to the storage node. The storage node becomes NW server. Disable the TL1 on it.

    We use DR licenses which are provided by licensing.powerlink.com for failover host. License the host as a server. That the way we cope with licensing issue. And we can restore everything.

    Failback process is more or less similar. Stop processes on SN and move X:\ drive to server.

    With this manner we skip long scanner procedure for tapes when we lose nsr db. The main fact is that we have two copies of all data so we use double number of cartridges.

    I hope it helps you.

    Best regards,

    Mario

  • Daniea3

    123 Posts

    524

    0

    Posted September 9th, 2010 02:00

    Good to know. Thanks for sharing the information.

  • MikeToo1

    38 Posts

    524

    0

    Posted September 14th, 2010 08:00

    Hello,

    There should be very little need to scan in tapes if you make regular index backups.  "mmrecov" would be used to restore your media database.

    Please note storage nodes do not have server software installed by default.

    The server identified by it's client ID owns the indexes.

    If there is a DR a server has to be built with the original client ID.

    I will attach the best practice DR guide for the nitty gritty.

    1 Attachment