CUCM Redundancy

[alshahiry]

Dears,

I have two Cisco Call Managers want them to work as redundany for my Cisco IP phones. Essentially I install them as a 1st node server and both are upoloaded with Node and IP phones licenes to ensure hardware/software redundany as per the requirements. Now on the DHCP pool created for this network i do mention both CCMs IP addresses as TFTP1 and TFTP2.

Now each IP phone had both TFTP servers but the wired issue that this scenario worked normally with one clinet as when i disconnect the 1st CCM it takes 1 minute and 2nd CCM hanlde the whole IP phones while my other client it does not work. Kinldy is there any paremeters should i take care of and check, kinldy advise?

Thanks.

[Aman Soi]

Hi,

you are correct, we cannot edit that.

If u make thefirst  tftp address as 172.16.4.12 in pool , are ip phones getting registered?

regds,

aman

[Vadym Maliarov]

1) When you say CUCM 1 is disconnected. How do you "disconnect" it? did you use power button? Did you enter "utils system shutdown" command?

2) Do you have SUB and PUB under CUCM Group?

3) How long did you wait for phones failover?

4) Are all phones effected?

5) Do you face the same issue with SIP and SCCP? Or with only one particular protocol?

6) What do you have under CUCM Service parametrs:

- Station and Backup Server KeepAlive Interval

- Station KeepAlive Interval

- SIP Station KeepAlive Interval

Also make the following test:

1) Utils system shutdown on PUB

2) Wait for the phones to failover

Vadym

[alshahiry]

Dear Vadym,

Answers are as below:

1. Power off the 1st CUCM through power buttom.

2. No its starnd alone servers for each CUCM.

3. Waited almost more than 5 minutes.

4. All IP phones get affected.

5. same issue with SIP and SCCP.

6. values are as below:

- Station and Backup Server KeepAlive Interval = 30

- Station KeepAlive Interval = 10

- SIP Station KeepAlive Interval = 60

Thanks

[Volodymyr Morskyy]

You are trying to failover phone to a different cluster. (each CUCM publisher in your case is a different cluster). Technically you can do this (just as proof of concept) by configuring same doamin name and then using DNS server to have redundant entry for the domain name.

But what you need is re-install one of the servers and build it as a subscriber, add to existing cluster. Then you will be able to see both servers under "CallManger servers" menu, youi can add then both to CUCM group.

[alshahiry]

Dear Volodymyr,

Unfortunately we can not do re-installaion of CUCMs! From other hands how we can confirm the steps you mention:

1. Configure same domain?
2. Using DNS server to have a redundat entery?

can you specify the steps please?

[Volodymyr Morskyy]

Hi alshahiry,

You asked strange question - you getting strange responce )). Please do not use in production and do not expect any support from cisco on this setup. Theoretically if you assign same name to CUCMs (but they will have different IP) you would be able to achieve redundancy using DNS. Let say you use name "cucm_pub". Configure DNS entry for "cucm_pub" specifying primary IP of CUCM1 and secondary IP of CUCM2. In DHCP server for the phone use domain name "cucm_pub" for specifying TFTP option. So the phone will go to CUCM1. If the CUCM1 goes down DNS should tell the phone IP address of CUCM2 and phone should go there. For this to work you will also need completely to replicate the setup from CUCM1 on CUCM2 (some sort of HSRP).

Again it is just to proove that this should be possible to make.

[Volodymyr Morskyy]

Could you please explain this frase: "...Essentially I install them as a 1st node server..." Couse for me it looks like you are trying to build cluster of 2 publishers. Cluster can have only one publisher, all other servers should be subscribers(MOH, TFTP, etc). Then you can add you both servers to CUCM group, create a device pool and assign previously created group to this pool. Phones will be registered to the first server in CUCM group, second will be used as a fallback option.