IPCC Express with High Availability

Unanswered Question
Apr 2nd, 2007


We have customer who purchased two servers for IPCC Express 4.x

installation with High Availability,

does any body have proper procedural document. for installation.

1. Is CRS A and CRS B both server will be having any connectivity

between them through crossover cable for redundancy.

2. While installing CRS B server does it going to use the same profile

name what CRS A had used,

3. Do i need to configure the same configuration again what i configured

in CRS A.



I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 4.1 (9 ratings)
mmelbourne Mon, 04/02/2007 - 08:46

I don't think a crossover cable between the spare pair of Ethernet ports is a supported configuration; it would also limit your options for Voice monitoring/recording. To avoid any issues with name resolution, it may be easier to just configure one interface, and let the replication occur over this. It's important to note that both servers must be in the same building and plugged into the same switch; effectively this is server resiliency, not network resiliency. We have two data centres separated by a resilient Gigabit Etherchannel VLAN trunk over dark fibre and siting the servers in different data centres wasn't possible. If the network becomes partitioned, then both servers could elect themselves as 'Master', and when the servers are brought back together, there is no mechanism to reconcile the databases.

Basically, you configure CRS A as you would normally - so you can test scripts, CAD (etc) before introducing the second node.

Ensure that both servers can ping each other and the CallManager(s) by name. If necessary add static entries to the hosts and lmhosts files. You install CRS B using the same profile name and the installer spots that it's the second node because all the cluster information, including the license details, are stored in LDAP (DCD). With IPCCX 4.0(5), I had to go back to CRS A and run the installer again to Update the node. MS SQL (Disk 4) needs to be installed on both nodes as a requirement for HA. You then go through the component activation stage on Node 2 and activate the required components.

A second set of CTI ports will be created when the CRS Engine component is activated. The CTI ports will use the next DN following on in the range used for the CTI ports on CRS A. Make sure these DNs don't overlap with any other numbers. Also, you'll need UCCX 4.0(5) if you want to have CTI ports with leading zeros.

You may need to check the default language on the CRS B after installation; I found it didn't inherit this from the CRS A, and so following failover it couldn't find any documents in the Repository and the scripts failed.

Also, when installing CRS on new boxes, build both using the OS Build CD/DVDs independently; don't try and short-cut the process by using a set of mirror disks to quickly build the base OS of the second server. This doesn't work as internal Computer ID values are identical which stops the Microsoft Distributed Transaction Co-ordinator from starting.

Please rate helpful posts.

shivkumargadad Mon, 04/02/2007 - 23:00


That above information provided was really helpful for me,

could you please tell me whether i need to configure the the Resource groups CSQ in CRS B server after CRS A configuration and testing.

Do i need to create different CTI ports and CTI RP on CRS B server, as what ports used in CRS A server.

If you have any cisco document on HA Installation that will be very much useful for me and will clear my all doubts.



mmelbourne Tue, 04/03/2007 - 00:57

You don't need to configure Resource Groups, Skills, CSQs (etc) on CRS B, as all this information is stored in LDAP and imported into CRS B at installation time.

Only different CTI ports are required on CRS B which are created during component activation on CRS B. Differnet CTI RPs are not required as these are controlled by the active CRS Engine.

There isn't an HA installation/best practice document which I do think is an omission, as it would be good to have an understanding of the failover architecture and behaviour under different failure scenarios.

The UCCX 4.0 SRND is here:


shivkumargadad Tue, 04/03/2007 - 03:37


Thanks for the useful information provided,

i am cleared of my doubts,

only and last question,

do i need to install SQL after os installation or after CRS installation.



mmelbourne Tue, 04/03/2007 - 05:01

CRS is installed after the main CRS installation. If I recall, there's a tick box to elect to install MS SQL2k as part of the component selection when installing CRS. However, it can be manually installed afterwards.

Note that the autorun on the MS SQL2k CD will stop the database on the CRS server, even before you've clicked the Install button to proceed with the installation!

Please rate useful posts!

Sandy Lee Tue, 04/03/2007 - 12:01


I'm in the process of upgrading and I want to have HA as well. Actually, we have IPCC 3.5(4) on one server and we want to upgrade to 4.0(5) and eventually add a server for redundancy.

In the SRND, it says that "Both servers must be collocated in the same building with the CallManager nodes running the CTI managers"

We have a CCM cluster over our WAN and I have the Pub as the primary CTI manager and one Sub as the backup CTI Manager, but they're in a different building and of course, different vlans.

Can I change the CTI managers in my cluster to make another Sub (located in the same building as the Pub) as the backup CTI ? If so, how do I do that, without having huge problems ? Or do I leave the CTI Managers like this ?

Thank you very much and your post was very helpful!

mmelbourne Tue, 04/03/2007 - 12:27

I suppose this would depend on the bandwidth and latency across the WAN. Our CallManagers running the CTI Manager are in different buildings (but the same VLAN, with effectively LAN characteristics). You can choose which CTI Managers are used, by changing the servers specified in the JTAPI Provider configuration in CRS AppAdmin. The first server in the list is used; if that is unavailable CRS will register the CTI ports and RPs with the next CTI Manager server in the list.

An upgrade from 3.5(3) is what we did, and the actual upgrade itself went fairly smoothly. We did debate the virtues of an upgrade versus a clean install, but we didn't want to lose historical data, so an upgrade was the only real option.

You do need to revisit your scripts and modify any which access files (e.g. XML files and dynamic prompts) directly. These files are now located in the Repository, which is replicated through the SQL2k Publisher/Subscriber operation.

Apart from that, all agents, Resources, CSQs, Skills (etc) are migrated. The CTI ports and RPs get recreated because in UCCX 4.0 any JTAPI configuration is performed through CRS AppAdmin, not in CallManager. Although new phones still need to be manually associated with the RMJTAPI user. The upgrade uses the existing RMJTAPI user, but creates new users for the JTAPI Provider.

The upgrade process is quite time consuming as the CRS boxes need to be re-imaged in 4.0(5) and effectively the configuration is restored using BARS during the Restore process. The process is well documented, and the use of BARS and the Recovery process itself become clearer when you undertake the upgrade.

Please rate helpful posts.

wenqianyu Tue, 04/03/2007 - 18:28

Hi there,

I just installed a CRS B (IPCCX4.0(5)). It seems can become active if the active CRS is down. My question is as following:

Do I need to run the installer in CRS A?

Is this the first CRS CD?

Do I need to choose "Install" or "Repair" when run the installer?

Will that cause any problem if I do not run installer in CRS A?


mmelbourne Wed, 04/04/2007 - 01:08

When I installed the second node, towards the end of the installation a "Synchronise Servers" message appeared, which stated to re-run the installer on Node A.

The node appeared as invalid in AppAdmin until this was done. When the installer was run on Node A, the only installation option was "Update Node". Install or Repair were not available.

The installer should autorun when Disk 1 is inserted.

wenqianyu Wed, 04/04/2007 - 15:21


I did not get this kind of message. Also the failover seems working for the standby server. Maybe a different version of IPCCX.

shivkumargadad Fri, 04/06/2007 - 06:04


I have installed CRS A server installed SQL and restarted then run setup as cluster.

and uploaded the license.and activated the CRSengine and all other components.

and after installing CRS B server installed SQL message appeared as synchronising

then setup run in server setup mode. and it didn't asked me to put the license, and when i selected the componenets like CRS engine its giving me an error, attached is the error.

It had took all the details like script jtapi rmuser, and csq resource group details,

but CRS Engine is not showing me in Control center, and the second attachment show the component activation for CRSB where CRS engine is unchecked, when i try to activate i am getting error shows in 1st attachement.

Please get back to me, is it because of license reason,

Please get back to me some body its very urgent for me.



mmelbourne Fri, 04/06/2007 - 07:42

Do you have an HA licence installed?

Are there any errors in the MADM/MIVR/MCVD log files?

shivkumargadad Sat, 04/07/2007 - 00:26

I dont know whether the license we have is for HA, how to verify whether the license consist of HA.

and where to find MADM and MIVR, MCVD log files.

mmelbourne Sat, 04/07/2007 - 03:31

What does your CRS AppAdmin License information page show (a link in the Control Centre pages)?

The logs are in C:\Program Files\wfavvid\logs

shivkumargadad Tue, 04/10/2007 - 04:45


There was problem with License, I uploaded the Correct License on the CRS A server and restarted the Engine, then tried uploading the license on the CRS B server but that was still there as Addon server, I reintsalled the CRS B server installed SQL then it run in to Server setup and joined to the cluster.

i just activated all the componenets, started working and all the configuration got replicated.

After reboot SQL replication happened automatically i didn't go to manually CRS update tool.

Every thing working fine including the failover.




shivkumargadad Tue, 04/10/2007 - 21:26


I am using Extension mobility,and IP Agent Phone service for agents to login,

I have two queries here,

1. Here every agent has to login to Extension mobility and then IP Agent Phone service. Is there any way where once the user login to Extension mobility should make him to login IP Agent phone also automatically.

2. As we were using IP Agent phone service, We have defined IP Agent Service as below in call Manager Publisher server.

Here in service i have defined only the CRSA server ip address in the Service,

what if my CRSA server goes down and how the agents will be logging in to IP Agent Phone Service. does it automatically takes over the CRSB server. and Agents will be Ready in IP Agent Phone Service.

do i need to mention CRSB server IP Address next to the CRSA server ip address in IP Agetn Phone Service URL.

mohitbawa2574 Tue, 04/10/2007 - 21:42

You will need to define two IPPA services . One for each IPCC Server. On a failure IPPA does not function like CAD . CAD automatically failsover in one minute but IPPA does not. Agents have to manually login to the secondary service.

I hope this helps!!!


Mohit Bawa

shivkumargadad Wed, 04/11/2007 - 05:21


Thanks for the information,

We have this setup newly installed its pages

are opening very slowly, like if i click Scrip management it takes almost 40 seconds to open, any page in CRS appadmin, in both the CRS A and CRS B sserver ,

does any body knows the reason.

Both of the CRS Servers are there in one workgroup even other two Call Managers also there in the same workgroup.

Rebooted both the servers , still its very slow. there's not cpu or memory utlization.

mmelbourne Wed, 04/11/2007 - 06:15

I'd first check name resolution. Also, in the browser on the CRS boxes, I've seen "automatically detect" proxy settings enabled which will slow down the first connection, but subsequent requests are usually OK.

Are the servers specified in LDAP configuration reachable (particularly the first in the list).

shivkumargadad Wed, 04/11/2007 - 06:57

I have edited the hosts files in both CCM's and both CRS server about the Hostname and IPAddress,

i am able to ping hostname from any server to any server.

here when i open crsadmin page first time it comes fast and logins also fast, but when i click Script management option or Prompt management it take about 20 seconds

In CRS LDAP Information both CCM's IP's mentioned.

shivkumargadad Thu, 04/12/2007 - 01:20


I have checked the LDAP Configuration in CRS Server both the CCM's ip address is mentioned.

The Master CRSA server has become very slow in operating the Appadmin page, when i open the Appadmin page it comes very fast but if i go for any script uploading it opens the page after 20 seconds, and even if i click the Prompt Management its very slow takes about 20 seconds any options if i click taking lot of time,

I have edited the hosts file, and both the CRS servers and CallManagers are there in one workgroup.

and all these servers are terminated on one 2950 switch, any server to any server ping response is about 10ms delay is there, and when i checked the CPU Utilization and Memory Utilization is its consuming very less,

both the servers i rebooted once, now the CRSA server is Master,

Could you tell me why its happening.

we have hardly configured about 3 scripts and about 10 agents,



mmelbourne Thu, 04/12/2007 - 01:33

Are you opening the CRS AppAdmin pages on the master server? Which version of IPCC Express? Does it make any difference if the standby server is disconnected? It sounds like the issue is with the Repository (where script, prompts and documents are stored). I suggest you raise a TAC case with your support provider.

shivkumargadad Thu, 04/12/2007 - 04:24

Yes I am opening CRS appadmin pages on the master server, its 4.0.5 version of CRS Server, I tried failover test and tried calling the ctiroutepoint i didn't faced any problem, I am storing the Scripts in


and Prompts i am storing in


do i need to store them in


mmelbourne Thu, 04/12/2007 - 04:31

Script and prompts (and documents) need to be uploaded to the Repository using the AppAdmin interface, rather than storing them in specific locations.

shivkumargadad Thu, 04/12/2007 - 05:27

I have uploaded the prompts using appadmin page, but the ivr script when i created i stored in the below path.


Do i need to do something here,

and when application manager service, and Manager manager service in partial service what needs to be done.

wenqianyu Sat, 04/07/2007 - 01:01

Hi there,

I have a problem with CRS database configuration. To use DB Read step in CRS Editor, you need to configure a database in CRS. Before CRS Database configuration, need to create a Database Source Name (DSN) in CRS server (in Data Source ODBC).

I have a DSN created in CRS A server, and created a CRS Database subsystem. On CRS A, database subsystem with CRS Engine has been started. However, the CRS B's database subsystem is not started. Failure message in the log was "No DSN Configured".

My question is: when create DSN in CRS A, do I also need to create the same DSN in CRS B? Is there a way that these two servers replicate their DSN information? If only configured in CRS A, and CRS A is down, then CRS Editor connecting to CRS B will not be able to find the database in DB Read. I guess I miss something in my configuration for two CRS servers. Please help.

mmelbourne Sat, 04/07/2007 - 03:34

I think you probably need to recreate the DSNs on CRS B, if you're sure everything else has been replicated and the Database Subsystem is still out of service, It's an interesting point and not one I've seen documented anywhere. I wonder whether the same is true of the e-mail Subsystem?

wenqianyu Mon, 04/09/2007 - 15:26

Yes, you are right. As soon as I configured DSN on CRS B, then the Database Subsystem under CRS Engin in CRS B became active. Is there any way to configure DSN on one CRS server and automatically replicate to the other?


This Discussion