UCCE Clustering

[bilalghayad]

Dears;

I have a disaster recovery site, and I need to use clustering principle (for IP Telephony and the Contact Center with the CVP), I will have site A and site B and I need to apply cluster principle for them.

Currently the main site need to be upgraded, and then we will apply cluster with the disaster recover site. The question is: Can I start the implementation of the Disaster Recovery Site and once finished then I can upgrade the main site and apply the cluster? Or I should upgrade the main site and set it in cluster mode and then to start the implementation of the Disaster Recovery Site?

Any advise?

Regards

Bilal

[david.macias]

Bilal,

What is cluster mode?  Take a look at the UCCE SRND it will give you an idea of the failover/redundancy that UCCE has, not really a cluster when talking about ICM or CVP.  The only UCCE product I would call a cluster is UCM.

david

[geoff]

I reiterate what David says. These systems are hard enough to understand, so let's use the terms in the SRND.

There is no such animal as a "cluster principle". UCCE works on a totally different paradigm to a cluster. For example, there is side A and side B - but no side C. It's a somewhat tricky and sophisticated synchronization model that you need to get a grasp of.

Regards,

Geoff

[bilalghayad]

Dears;

Please cool down.

I am talking about two physical sites, one in a city and another one in another city.

We got design from the ATP that saying they are going to do a cluster solution (City one has UCCE "which has side a and side b" and the other City has UCCE "which also has site a and site b"), and the both cities are working based on clustering. As I understood, that while the installation of the IPCC, we can select if the system to be cluster.

Now, I am same as both of u .. I am wondering how this is existed with UCCE, did anyone here about a possibility to run UCCE in clustering mode?

I know that it is possible in Call Manager ... but the question, how it will be with UCCE? Did any one here about this?

Regards

Bilal

[david.macias]

I think you should have the ATP read the SRND too.  What you're discribing is possible, but it has nothing to do with clusters.

david

[bilalghayad]

Sorry friends, it has been explained for me in a wrong way.

Actually the design is to use side A in one city and side B in another city (and not to have two IPCCE clusters).

But what is the maximum acceptable delays between these two sites to have a good implementation? Is it important to apply layer 2 connection between these two sites or it can work even if we applied an IP Routing between those two sites.

Any special consideration in such implementation that u would recommend or advise me to take care for it?

Note: There will be also CVP and Call Manager in site A and in site B.

Your kindly advise is highly appreciated.

Regards

Bilal

[bilalghayad]

In other words:

Because there is a private network and public network, so for private network, they have to be in the same IP address network range (same subnet mask) or we can deploy them in different cities and using a routing connection, then they can work even if both are from different subnet mask?

Now, regarding to side A which will be located in City A, how can I have a redudancy in the same site (in case side A down, and we do not need to use side B .. then what could have a kind of redundancey to recover this shortage)? This could happen because the link between both sides are down, and at the same time we might have upgrade for side A, so we are thinking what kind of redundancy that can be used to recover this?

Thanks for the kindly advise.

Regards

Bilal

[david.macias]

Have you read the SRND?  Everything you seek is in there.

david

[bilalghayad]

Dear;

I read the SRND, in case of using the WAN, it is still talking about private network.

My question not about the needed bandwidth for the Message Delivery Service (MDS), but it is about the IP address schema.

For example, if the IP address of the side A was:

192.168.x.x

Then, can I use IP address for side B to be:

10.10.x.x

Or they have to be in the same subnet mask?

We can guarantee a bandwidth between these two sides (A and B) and a fiber connection, but is it a condition to be in the same IP address range (in other words, is it a condition to be in the same VLAN?)

Your kindly help is high appreciated.

Regards

Bilal

[david.macias]

The IP address doesn't matter, it's the bandwidth which makes all the difference.

david

[bilalghayad]

Fully thanks for the help

In case I have side A and side B in the same geographical area, and then we decided to build a disaster recovery site in a new geographical area. Is it possible to convert side B to become side A (so to have two PGs for each interface and two central controllers), this to make side A it self redundant. Because we might have enough servers to build the disaster recovery site "the new geographical area" (to be side B) without use any of the servers in the first geographical area.

I know that I am talking about something complex (to have side A it self redundant .. so to have two services of each service and all of this to be side A) .. but I am thinking that this could be needed to have the geographical area #1 redundant it self (because in case of have a fail in the servers in geographical area # 1, still we do not need to switch over to the geographical area # 2 which is the disaster recovery area), that is why I am thinking to make side A redundant.

I read about the parent/child, but I did not c that it suite my case specially that I do not need to use system CCE or CCX. Maybe there are other possible scenarios ... but did not get until now something suite what I am looking for (to make side A it self redundant while side B to be in another geographical area which is the disaster recovery area).

Special thanks for all those helped.

Regards

Bilal

[bilalghayad]

If IP addresses will not effect, so there should be a place where I can enter the IP address of the private network to let the servers know how to communicate with each other.

I mean, where I can enter the IP address of the private network of RGR-B at RGR-A and same thing to enter the IP address of the private network of RGR-A at RGR-B? Same thing for PGs and HDS?

Thanks for the help.

Regards

Bilal

[david.macias · Accepted solution]

You can put the public, private, high IPs when you run setup.  Check out the installation guide.

david

[bilalghayad]

Thanks for the help and explaination.

So whenever is needed to change the IPs, I have to run the setup? Can not be changed from any other place?

Regards

Bilal

[david.macias]

Yes, in the registry, but I recommend you stick to setup.  Any reason why you don't want to run setup?

david