[cisco-voip] CallManager DR scenario

[Ed Leatherman]

The way I configured my device pools was something like:

pub at main site.

sub A, sub B at main site
Sub C, sub D at secondary site

I split campus up into two areas, phones in one area go to the main site and
in the other go to the secondary.
CCM Groups: ABC, BAD, CDA, DCB

then based on the primary line of a phone:
ABC = even numbered DN's in main area
BAD = odd numbered DN's in main area
CDA = even numbered DN's in second area
DCB = odd numbered DN's in second area

this way either site can cover if the other goes offline, and the loads are
relatively well balanced on each server. have a seperate TFTP Server at the
main site that could turn into a subscriber if needed, and a spare server at
the secondary site that could be pub or sub if needed. Has worked out well
so far, havent had to rebuild any servers yet though (jinx!)

On 6/26/06, Carter, Bill <bcarter at sentinel.com> wrote:
>
> Actually it isn't a bad idea to have drives available for cold standby.
> IPCC Express cold standby servers will be located at the DR site.  The
> subnet will be extended to the DR site via metro fiber.  When the rotate
> drives for IPCC they can have standby drives for the DR site.
>
> I need to think about server placement and groups.  I was thinking of
> the Publisher being a "Glass House" server.  CM groups would be
> configured with primary and secondary as HQCM02 and DRCM04, or HQCM03
> and DRCM05.
>
> Maybe I could do this:
> Group A - HQCM02, DRCM04, HQCM01
> Group B - HQCM03, DRCM05, HQCM01
>
> In this scenario, during a prolonged outage, I could put HQCM01 drives
> into the DRCM05 server.  Then devices would register to either DRCM04 or
> HQCM01 server.
>
>
-- 
Ed Leatherman
IP Telephony Coordinator
West Virginia University
Telecommunications and Network Operations
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://puck.nether.net/pipermail/cisco-voip/attachments/20060626/9bb6b790/attachment.html