How to restore unity connection publisher?

[srosenthal]

I found a typo in the name of the publisher so I changed it in the OS Administration.  After the server rebooted I can no longer log into the Administration of the unit.  I get database error messages when I try to log in.  The subscriber is operating as primary right now.

So I guess I need to reload the machine completely.  My question is once I start the install process and bring this machine back on line, will it take all of the database information from the subscriber or is there something I need to do during the install to make that happen.  I really do not want to recreate everything that is there even though these units have not gone into production yet.

There are also no backups done since it is not in production.   I know that during the install I get asked if this is the first server in the cluster, do I answer no even though eventually I want this server to be the publisher again?

Thanx, any help is greatly appreciated.

Seth

[David Hailey]

Well, with Unity Connection clusters - at least in 7.1, they are set up to provide HA by having 2 servers configured in a cluster.  The servers are Active-Active in that if one fails, the other can automatically take over.  The SRM (Server Role Manager) is the primary piece of how that works.  How licensing typically works is that with clustering you need a license for each server.  You install one license on the Pub (with MAC address of it's Eth0 interface) and a license on the Sub (with MAC address of it's Eth0 interface).  As far as ports go, the Unity Connection clusters are intended to provide 100% failover (i.e., overprovisioning is technically possible in some scenarios but ideally if you need 72 ports for all your messaging traffic then you should provision 72 ports on each server).  So, the servers share the same phone system integration but each server has it's own set of ports that will register with CUCM.  From there, the CUCM line group configurations control how you distribute calls between the servers.  Ideally, you treat a CUC Cluster like a CUCM cluster in that the Subscriber should handle the majority of all the calls.  If there is burst past it's capacity, you can roll calls to the Publisher via the line group configs.  You also treat the Publisher like a Publisher in that it should be the primary server for handling web traffic (e.g., CU Admin, Cisco PCA, IMAP, etc).

With all that said, there is also a concept of maintaining a 7x warm-standby model.  Basically, you install a spare server somewhere, maintain the DRS backups of the active server somewhere that they can be accessed in the event of emergency/necessity, and then you restore that server with the DRS backup in an emergency.  You can actually maintain this type of setup for a single Connection server or a Connection cluster.  Licensing in this model is a little different.  1) You can purchase a dedicated license 2) You can use the license from the server that has failed and is in the DRS backup.  The MAC addresses won't match up so you have to reboot the standby server every 24 hours because the server is in violation of the license. 3) You can get a replacement license for the server that failed but assign it to the standby server (i.e., with MAC address of the standby's Eth0 interface).

From what you've described, I've assumed you set up a standard Unity Connection cluster.  When you first changed the host name of the Publisher, the Subscriber took over active status, right?  If that's the case, you're doing a normal Unity Connection cluster and what I said in the first paragraph applies.  If you want to send me a private message, I can look at your license files for you if you would like but that's up to you.  I don't know that I really need to based on what I think you're working on.

Hailey

Please rate helpful posts!

[srosenthal]

I was looking over the quote and do see UNITYCN7-HA-24 listed so I  must have missed that somewhere and did not enter the PAK on Cisco to  get the license file.  I will look for it but it might not be until  Thursday.

I created a user on the subscriber and it replicated to the publisher right away.  I ran the show perf query again on the pub and it still says "0".

I then deleted the user from the pub and it was gone from the sub.

I thought all license files get installed on the pub and would be automatically applied to the subscriber like the call manager does.

Btw, what brought all this on was I was trying to troubleshoot mwi not working and not being able to call a phone to record prompts.  I can of course force mwi on and off by dialing the numbers from the phone.

Just thought I would throw that in.

Seth

[David Hailey]

So you have HA for up to 24 ports, that's what that product code is. There is a command to attempt to repair replication on a node; however, I'm not sure how much good that will do currently. So, let's proceed as follows:

A few questions for you to answer:

1) Do you have RTMT loaded and do you know how to look at the replication state in RTMT? If not, I can tell you.

2) Which server is active in the cluster? Go to CU Serviceability > Tools > Cluster Mgmt and it will tell you.

3) For clusters, this is how licensing works: When a Cisco Unity Connection cluster (high availability) is configured, two licenses are required. The license that has the MAC address of the publisher server must be installed on the publisher server. The license that has the MAC address of the subscriber server must be installed on the subscriber server. So, there is likely an issue here that you need to straighten out with TAC regarding the licensing on the cluster. I would do that first and foremost if you can.

4) Let's get the cluster working properly and then I can help you with MWI. Is this a SCCP or SIP integration with CUCM (I recommend SCCP if you're not already using it).

5) Are there any other PBX's involved here where you're using TIMG or PIMG for integration?

Hailey

Please rate helpful posts!

[srosenthal]

I do have RTMT loaded but could not find where to look at the replication state.

Under CU Serviceabiltiy - Tools - there is no Cluster Mgmt option.

I think the connection to CUCM is sccp, don't remember choosing SIP when we set the ports up.  No other PBX's, just the call manager.

I do notice in RTMT I am getting a lot of  SyslogSeverityMatch messages.

I will be out on the road again tomorrow most of the day.  Thanx again for all the help.  I guess I will have to open a TAC case on Thursday morning.

[David Hailey]

Ouch. No cluster management option is a bad indicator. I hate to say

that. There are still some things I could try to help you with but

you should get the TAC case open for sure. Can you still web into the

Publisher? If so, under Cluster do you have the Subscriber defined as

a second node? You wouldn't have been able to install the second node

without it but check again. In the interim, hit me up when you get

back to working on this. In the end, even if you have to rebuild from

scratch I can guide you with the right procedures to ensure the 2nd

time is a charm. You're welcome for the help...that's what the forums

are for anyways.

Hailey

Please rate helpful posts!

Sent from my iPhone

On Mar 16, 2010, at 10:46 PM, srosenthal

[srosenthal]

Ok, my bad, I was looking under Unified Serviceability and not Unity Connection Serviceability.

So I tried to make the pub primary and it did not work.  It looked like it was making the switch but in the end the subscriber took over as primary again.

I do have full access into both boxes, via web or cli.

I really need to have these boxes stable by the end of the week so I am seriously thinking of just going ahead and re-installing the OS on both units, making certain I get the spelling right.  They are going to need to go live on Monday.

My guess is then all I will have to takle is the MWI on/off issue and I am trying to locate the license for the HA as it was ordered but I have not received it yet.

Any tips on doing the reload would be much appreciated.

Thanx, Seth

[David Hailey · Accepted solution]

Well, if you want to go the reload route then this is the way to do it:

1) Shutdown the Subscriber.

2) Rebuild the Publisher via DVD which will overwrite the hard drive. The DVD comes with the server. Do not try to use the "Recovery" CD, just the straight up application CD - CUCM version (whatever you had shipped).

3) Before you install the Subscriber, make sure you verify NTP sync on the Pub (critical).

4) Once you have the Publisher rebuilt, add the Subscriber in CU Admin as part of the cluster just as you normally would. Use the IP address of the server (recommended).

5) Rebuild the Subscriber from DVD as you did the Publisher. It should be added as the second node in the cluster so make sure you have DNS and IP connectivity to/from all resources needed BEFORE STEP 2 ever occurs.

6) Once both servers are up and running, check CUC Serviceability and look at the cluster status.

7) Then, go to the CLI on both servers and run "show tech network hosts". The /etc/hosts file should have a loopback address and both cluster servers included in the file.

8) Make a test user on the Pub, verify it replicates to the Sub. Delete it from the Sub, make sure it deletes from the Pub.

9) Test failover then test failback by going to CUC Serviceability and putting the Publisher back as active.

10) You can also look at the cluster status on both servers from the CLI - show cuc cluster status is the command to do that.

11) If all that checks out, get your license files loaded and do your configurations.

12) Make sure you set up DRS backups on a schedule (for both servers) and perform an initial manual DRS backup of both as well.

13) I would also recommend that you upgrade both systems to the latest SU for 7.1.3 which is 7.1(3b)SU2. You can download it from CCO.

As for MWI, there is a specific service that enables MWI to work properly. If that service didn't start on the Sub for whatever reason when it failed over, then that may be the cause. Get the servers rebuilt and let's go from there. Sound good?

Hailey

Please rate helpful posts!

[srosenthal]

Oh happy day!  It looks like third time is a charm!  I reloaded both machines and all is working, even the MWI.

Hailey, thank you so much for all the help!

I am in Richmond, VA, if you are close I still want to buy you that cold one.

One last question as I am too tired to look it up right now.  What is the web address for the end user to log on to unity to their personal mailbox?  For example I know of CM it is www.callmanager/ccmuser, what about unity connections?

Seth

[David Hailey]

That would be Personal Communications Assistant - PCA: https:///ciscopca

No problem, man. Glad I could help. I'm in DC but have friends in Richmond and come down often. I'll hit you up next time I'm down and we can grab a cold one (or 10)...but you don’t have to buy. Like I said, just glad you got it worked out.

Hailey

Please rate helpful posts!

[srosenthal]

I often get up that way also for work, if you want, you can email me at kf4djx@yahoo.com.

Take it easy!

[srosenthal]

Ok, here is a question regarding the HA license.

I did install it earlier, did not realize thats what it was, but I think I did it incorrectly.  I registered the PAK to the Publisher and installed the license there?  The screen show right amount of mailboxs but shows twice the number of ports with half being used by the publisher and none being used by the subscriber.

Should I have registered the PAK with the subscriber?

Seth

[David Hailey]

You should definitely have a separate license on the Subscriber. What does it show?

[srosenthal]

The subscriber only shows the demo license.

I guess I need to call Cisco to get them to re-issue the license with the subscibers mac address.

Seth

[David Hailey]

You need a license on each server homed to each individual NIC (eth0

only).

Sent from my iPhone

On Mar 22, 2010, at 6:44 PM, srosenthal

[Rob Huffman]

Hi Seth,

Sorry to hear your having problems with this, but at least you've got Hailey

with lots of great tips (+5 Hailey!)

I thought you might find these docs helpful;

Replacing Cisco Unity Connection 7.x Servers

http://www.cisco.com/en/US/docs/voice_ip_comm/connection/7x/upgrade/guide/7xcucrug040.html#wp1052015

Renaming Cisco Unity Connection 7.x Servers

http://www.cisco.com/en/US/docs/voice_ip_comm/connection/7x/upgrade/guide/7xcucrug060.html#wp1053641

Cheers!

Huff