We are wanting to implement Unity, however are concerned about Disaster Recovery. We have multiple locations and if we place a Unity servers at each location it will be quite costly. However, if we place one or two Unity servers in a single location and that location experiences a disaster, where as Unity is not available, the remote sites can not get there VM anymore. Is there a happy medium somewhere? If so, can someone explain or get me a link that explains?
What PBX(s) and how are they distributed in your network (that you plan to integrate with Unity)? Will you be using Unity voicemail with Exchange/Notes or Unity Connection (its own message store)? If using Unity with Exchange message store, you must plan for your domain and global catalog server locations (Unity will want to be in the same subnet as its partner Exchange server, unless Exchange is onbox). Knowing this will help you deploy Unity in your network. Disaster recovery then is a business issue for you. Does your customer service level agreement require 24X7 access to voicemail? There is an obvious cost to having a spare server hardware with licensing waiting for a failure to occur. For us, we elected not to do failover. We have the same SLA for voicemail as email (we are a unified messaging installation with Exchange) so we have a 4-hour turnaround to get Unity operational. If you have redundant/hot-swappable components, i.e. dual-CPU, dual RAID1 mirrors, you mitigate the need for a fully redundant server. In the 5 years we have been running 3 Unity servers and one Bridge server, we have not had an unplanned outage. Here is a link to the Unity design guides - http://www.cisco.com/en/US/customer/products/sw/voicesw/ps2237/products_implementation_design_guides_list.html
Start with 4.0 and then review the addendums for 4.1 and 4.1. Notably chapter 2 - Network and infrastructure considerations.
- SLA requires 24x7 access, with special exceptions during scheduled maintenance windows.
- WAN Topology
* Loc1 - 25Mbps PPP MPLS
* Loc2 - (Backup to Loc1)25Mbps PPP MPLS
* Loc3-9 - 3Mbps PPP MPLS
Currently, if Loc1 dies, all phones still work because of Loc2's subscriber servers and if Loc2 is gone as well, SRST gives the users are the remote locs basic functionality of their phones. However, if we went to Unity for voice messaging, without unified messaging (because of some pretty stringent email requirements and various other issues), how could we get the same type of redundancy? Could we put one Unity server in Loc1 and a second in Loc2, each with redundant message stores or something like that? Ultimately it seems that we would have to have a message store in Loc1 and a message store in Loc2, if in deed this is the case, what would it take to replicate the 2 for failover purposes.
From the design guide, it looks like you could put a Unity voicemail server in loc1 and another in loc2. Failover support would require another Unity server in each loc1 and loc2 (to provide support if the Unity server had a problem). If your loc3-9 sites need access to voicemail or auto attendant functionality in the event of a WAN failure, you could install Unity Express on each loc3-9 router. Unity Express can be networked with your other Unity servers to provide identified subscriber messaging (check digital networking in Unity). These are just conceptual ideas for you to consider. I would definitely contact your local Cisco SE to validate any design considerations.
I have similar kinda question to ask you as well. Would be a great help if get any answer.
We have Unity Connection 2.1 installed at our HQ with CCM6.1 and Unity Express with CCME installed at one of our remote sites. We are planning to convert CCME to SRST and register all remote site users to the central cluster of CCM6.1 so that all remote site users can get access to the CCM 6.1 featureset. In addition, we are also planning to put all remote site voice mail users on Unity Connection 2.1 server instead of separate Unity Express module. So here is my question, I have been asked if we could use Unity Connection 2.1 as a Primary voice mail solution for the remote site users and use unity express with SRST as a secondary solution, should the WAN link between HQ and remote site is down.
I have never come across this kinda situation or configuration so if there is any document available explaining the scenario i.e. how to configure unity connection and express in primary/secondary mode, what key points needs to be considered , any additional licences required to make them working in primary/secondary mode, etc would be a great help.
Depends on a lot of things. If you are strictly a VM user, and have no need for UM with Unity, Unity Connection in Active/Standby will work. It basically uses the same concepts as CUCM and clusters. (but only pairs does it work, not multiply like CUCM)
For Unity 5.x, Failover can be split, but limited to roundtrip time, bandwidth and distance. There is SQL replication going on that has requirements before you try and split a pair. (short distance, high bandwidth seems to be fine with Cisco)
In Unity 5.x, you also have to deal with Exchange and how will this be managed in a DR situation.
I have one customer who has a pair of Unity servers with 5.x and failover and message store. In the event the totally redundant datacenter is burned to the ground, there is a cold standby Unity server at a remote location. They do run DIRT weekly to update the subscribers greetings, recorded names etc. CUCM simply has a third hunt group for the DR server, when primary and failover are no longer available (highly unlikely) it will go to the DR Unity server.
I'm not able to access my old voice mail messages all of a sudden. The recording says something like 'the message is currently not available'. This has never happened before in all the years I have been using this system. I have t...
If you have 2 ISR routers, one acting as Failover, do we need to have both the same number of SRST licenses on the 2 routers?
No. You will only need the SRST licenses on the primary router. Because this feature...