Solved: Cisco IP Phones going down when CUCM publisher is turned off

Geo Gon · ‎05-31-2012

Hi all,

Quick question:

I have on my lab environment a CUCM Publisher and a CUCM Subs.

As I understand, the subscriber is only there for redundancy purposes, and the phone should be registered to the subs as opose to the pub.

All my IP phones register to the subscriber, but when I kill my publisher, all the phones are losing its registration.

Aren't they meant to be still registered?

Thanks

nikshah · ‎06-01-2012

you might want to run utils network connectivity on the sub and see if it completes successfully.

any fw between the pub and sub..

its definitely a db replication problem.

if the above tests that i mentioned above are successfull including utils network connectity. you would want to run a stop and roll.

utils dbreplication stop on the sub (wait for it to complete)

utils dbreplication stop on pub

utils dbreplication reset all from the pub.

View solution in original post

Ayodeji Okanlawon · ‎06-02-2012

Geovani,

What you need to do is to use cucm group and device pool for redundancy.

SO in your cucm group you will have your subscriber first and your publish second. You will then assign this cucm group to a device pool and then assign the device pool to the phone.

Now when your subscriber is down, the phones will register to your publisher. This is only good for small deployments. For large deployments you will have multiple subscribers and you will then have first subscriber and second and third. You can only have 3 servers in a cucm group. So the phones can register to secondary and tertiary subscribers when the primary is down

Please rate useful posts

Please rate all useful posts

View solution in original post

Chris Deren · ‎05-31-2012

What is the CUCM group configured like?

Is cm service running on the subs?

Do the phones show registered to the subs?

Sent from Cisco Technical Support iPhone App

Geo Gon · ‎05-31-2012

Hi Chris, thanks for your reply.

I'm still quite new to all this.

CM service is running on the sub.

When the pub and sub are on, it shows that the phones are registered to the pub, even though the tftp address on the ipphones are poiting to the sub.

I have 4 CUCM groups configured, just for lab porpuses, but im only using one of them.

Just noticed that when switching off the pub, all the replicated data that is meant to be on the sub, is not there. Users, Phones etc.. all gone.

??

thanks

robloeber · ‎05-31-2012

I dont think your Pub and Sub are replicating properly. I also think you CM groups are not configured properly. You need to verify whether or not replication is working properly before you look at anything else. I usually do this from the CLI using the following command: utils dbreplication runtimestate

You can also use RTMT if you have it installed. You want to make sure the status comes back as status 2. Any other status means you have replication troubles.

Chris Deren · ‎05-31-2012

Looks like you have db replication issues. Search this forum for dbreplication and you'll find many good threads.

Sent from Cisco Technical Support iPhone App

nikshah · ‎05-31-2012

you might want to login to the cli of the pub and run the following commands that we have a good dbrepln

utils dbreplication runtimestate

utils dbreplication status

utils network host (hostname and ip of the sub), repeat the same from the sub

utils ntp status

see if this all good

Take a dummy phone /test phone change the description like test 1

fro the ssh run the command run sql select * from device where description ='test1'

run it on both the nodes and both the entries should be same.

hope this helps

Geo Gon · ‎06-01-2012

Hi all,

When I type utils dbreplication status that's what I have.

PUB

connect to geovanipub_ccm7_0_1_11000_2 failed

Enterprise Replication not active (62)

command failed -- unable to connect to server specified (5)

SUB

SERVER ID STATE STATUS QUEUE CONNECTION CHANGED

-----------------------------------------------------------------------

g_geovani_ccm7_0_1_11000_2 2 Active Connecting 445296

g_geovanisub_ccm7_0_1_11000_2 3 Active Local 0

-------------------------------------------------

Ayodeji Okanlawon · ‎06-01-2012

Hi,

Can you ping between your publisher and subscriber?

Please rate all useful posts

Geo Gon · ‎06-01-2012

Hi,

Yes, I can ping between them.

nikshah · ‎06-01-2012

you might want to run utils network connectivity on the sub and see if it completes successfully.

any fw between the pub and sub..

its definitely a db replication problem.

if the above tests that i mentioned above are successfull including utils network connectity. you would want to run a stop and roll.

utils dbreplication stop on the sub (wait for it to complete)

utils dbreplication stop on pub

utils dbreplication reset all from the pub.

Geo Gon · ‎06-01-2012

Thanks very much.

I did run all the test you mentioned before, and it was sucessfull.

Will now try to stop and roll, see what happens

Cheers

Geo Gon · ‎06-01-2012

Right:

After doing the above, by nikshah this is what happened:

Pub replication status = 3

Sub replication status = 4

Pub utils dbreplication status:

SERVER ID STATE STATUS QUEUE CONNECTION CHANGED

-----------------------------------------------------------------------

g_geovanipub_ccm7_0_1_11000_2 2 Active Local 0

Status cannot be reported when replication is down on a publisher, or for a cluster with a single active node; aborting status check operation

Sub utils dbreplication status:

connect to geovanisub_ccm7_0_1_11000_2 failed

Enterprise Replication not active (62)

command failed -- unable to connect to server specified (5)

Status cannot be reported when replication is down on a publisher, or for a cluster with a single active node; aborting status check operation

Both the sub and pub can ping each other, and when i check the phones, users, etc on the sub they are all there, as long as the pub is on.

thanks again folks

nikshah · ‎06-02-2012

it seems your replication aggreements never got setup.

It takes a while for the replication to be setup, can you verify the latest status.

sorry couldnt respond earlier.

knovak123 · ‎06-02-2012

Did you change the name of either of the servers?

There is a very specific process for changing the name of a UCM server.

If it is not followed, it may be easier to just delete the sub from the cluster DB and re-add then re-install the sub.

To repair after changing without following the process typically requires TAC to root in and edit all of the hosts, rhosts, sqlhosts, etc.

Thanks,

Kevin

Sent from Cisco Technical Support iPad App

Geo Gon · ‎06-02-2012

Hi Guys, I didn't change the names or anything.

To be fair, I dont even know if it was working before.

Anyway, I have solved the problem by reinstalling the sub and the pub. All good now.

One other thing.

Cisco advises us to register phones with the subs right?

So, on my ip phones, I'm pointing it to my sub, the things is, if I kill my sub, I'd lose all the phones.

How's the redundancy suppose to work?

thanks again for your help