uccx upgrade

Answered Question
Apr 27th, 2010

Hi to all,

Need some pointers to do the following:

Have 1 server in production running uccx 5.2 express. (customer not allowing upgrade to 7 directly)

Have built a second server running uccx 7.

I need to migrate everything from the prod server to the new one.

Whats the best option for doing this?

Can i backup and restore to the new server? If so does it need to be on the same rev and ip?

If i build an exact replica of the queues on the prod server will this interfere with existing agents?

How do i point the agent desktop to connect to the new server instead of the old prod one?

Anything else i need to be aware of?

Thanks in advance...

Taff..

I have this problem too.
0 votes
Correct Answer by Aaron Harrison about 6 years 7 months ago

Hi

Ah - I see.

A single node is a primary node.

The only different (I believe) is that when you select 'Single Node' in the initial setup, it doesn't make itself a publisher. When you 'add to cluster' it does it at that point... so you don't need to meddle with the first UCCX node at all.

Regards

Aaron

Correct Answer by Aaron Harrison about 6 years 7 months ago

Hi

OK - we're kind of going in circles here.

Option 1)

You in-place upgrade your UCCX5 to 7, and keep all your data. This doesn't give you an easy rollback. If you do this, you add your new server to the cluster afterwards after building it fresh from media.

Option 2)

You can backup UCCX5, build a new UCCX5 on the new hardware, restore the UCCX5 data to it, then upgrade that to version 7.0.

When you have done this, and switched clients to it, and you are happy you no longer need the original UCCX5.0 server, you can then rebuild the UCCX5.0 server from media and add it to the UCCX7 cluster that contains your upgraded data and data gathered since the upgrade.

Aaron

Correct Answer by Aaron Harrison about 6 years 7 months ago

Hi

Yes - the main risk is that if you do hit problems, and need to roll back... you can't just uninstall the new version.

You would need to build from media, reinstall 5.x, and restore.

Building UCCX is a far less painful than it used to be, so if you have hardware available as an option I would probably take the 'safe' option of a parallel build of 5.x, restore, then upgrade.

Regards

Aaron

Correct Answer by Aaron Harrison about 6 years 7 months ago

Hi

- Have 1 server in production running uccx 5.2 express. (customer not allowing upgrade to 7 directly)

The only options you have are to upgrade from 5 (I presume you mean 5.0(2) ) to 7, or to manually migrate the configuration (scripts, queues, user team and rights assignment etc.

Bear in mind that if you don't upgrade, customer will lose their historical data.

Also consider moving to version 8.0 - as this is a migration to linux you will be able to export the config using tools that also export the historical data and import it to Linux. You would need to verify compatibility for your CCM version and so on.

- I need to migrate everything from the prod server to the new one. Can i backup and restore to the new server? If so does it need to be on the same rev and ip?

As stated above, if you want to migrate everything you need to upgrade. You could optionally do it like this:

- Build the OS off the network with same hostname/IP, install UCCX 5.0(2) and your current SR, but do not run the AppAdmin fresh install routine

- In your maintenance window, take a backup, disconnect the 5.0 server and connect the new server. Run the 'restore' option in AppADmin

- Once complete, upgrade the system to v7.0 and the latest SR

- Your clients should prompt to automatically upgrade when you run them. This requires admin rights, so you need to log into each PC as a local admin.

- How do i point the agent desktop to connect to the new server instead of the old prod one?

If it's a new build on a new IP, you would be advised to remove the old version, and then install the new version from AppAdmin  on the new server.

You may be able to just re-point the IP in the registry at HKLM/Software/Spanlink/SiteSetup/LDAP_HostA

As I say - the main consideration would be maintenance of their historical report data, and the accuracy with which you can migrate their scripts and configuration. If Historical data is important to this customer (it usually is) and if you aren't experienced with the product (which I gather you are not) then your best option would be an in-place upgrade, or the option suggested above of restoring your current version then upgrading to 7.0. The fact that you have a new server is nice as it means you have an easy rollback option of reconnecting the old server.

You would also want to ensure you take a backup of the system, and that you apply the latest supported OS upgrade/SR to ensure you are patched up to date.

It's also a good idea to upgrade a couple of CAD agents and a Supervisor, then test the new system fully before upgrading all their CAD installs; just in case you need to back out.

Regards

Aaron

Please rate helpful posts...

  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 5 (4 ratings)
Loading.
Correct Answer
Aaron Harrison Wed, 04/28/2010 - 00:02

Hi

- Have 1 server in production running uccx 5.2 express. (customer not allowing upgrade to 7 directly)

The only options you have are to upgrade from 5 (I presume you mean 5.0(2) ) to 7, or to manually migrate the configuration (scripts, queues, user team and rights assignment etc.

Bear in mind that if you don't upgrade, customer will lose their historical data.

Also consider moving to version 8.0 - as this is a migration to linux you will be able to export the config using tools that also export the historical data and import it to Linux. You would need to verify compatibility for your CCM version and so on.

- I need to migrate everything from the prod server to the new one. Can i backup and restore to the new server? If so does it need to be on the same rev and ip?

As stated above, if you want to migrate everything you need to upgrade. You could optionally do it like this:

- Build the OS off the network with same hostname/IP, install UCCX 5.0(2) and your current SR, but do not run the AppAdmin fresh install routine

- In your maintenance window, take a backup, disconnect the 5.0 server and connect the new server. Run the 'restore' option in AppADmin

- Once complete, upgrade the system to v7.0 and the latest SR

- Your clients should prompt to automatically upgrade when you run them. This requires admin rights, so you need to log into each PC as a local admin.

- How do i point the agent desktop to connect to the new server instead of the old prod one?

If it's a new build on a new IP, you would be advised to remove the old version, and then install the new version from AppAdmin  on the new server.

You may be able to just re-point the IP in the registry at HKLM/Software/Spanlink/SiteSetup/LDAP_HostA

As I say - the main consideration would be maintenance of their historical report data, and the accuracy with which you can migrate their scripts and configuration. If Historical data is important to this customer (it usually is) and if you aren't experienced with the product (which I gather you are not) then your best option would be an in-place upgrade, or the option suggested above of restoring your current version then upgrading to 7.0. The fact that you have a new server is nice as it means you have an easy rollback option of reconnecting the old server.

You would also want to ensure you take a backup of the system, and that you apply the latest supported OS upgrade/SR to ensure you are patched up to date.

It's also a good idea to upgrade a couple of CAD agents and a Supervisor, then test the new system fully before upgrading all their CAD installs; just in case you need to back out.

Regards

Aaron

Please rate helpful posts...

clougher01 Wed, 04/28/2010 - 00:31

Thanks that was a great reply..

What's the risk of inplace upgrade from 5.02 to 7.01? Rollback options?

Correct Answer
Aaron Harrison Wed, 04/28/2010 - 01:15

Hi

Yes - the main risk is that if you do hit problems, and need to roll back... you can't just uninstall the new version.

You would need to build from media, reinstall 5.x, and restore.

Building UCCX is a far less painful than it used to be, so if you have hardware available as an option I would probably take the 'safe' option of a parallel build of 5.x, restore, then upgrade.

Regards

Aaron

Attachment: 
clougher01 Thu, 04/29/2010 - 00:42

Hi,

Just a couple more questions for you. Our intention is to build a warm

standby cluster utilising the other server.

How do we potentially roll-back to the V5 server if all the clients have

been upgraded at the same time?

How does this become a member of a fail-over cluster if we've upgraded

from existing config? How do we make it the first node in the cluster if

its already built?

What happens with the build of the fail-over server once the V7 server

has gone live? The same with the old server - how do we add it to the V7

cluster?

Thanks for your answers so far...

Aaron Harrison Thu, 04/29/2010 - 02:47

Hi

How do we potentially roll-back to the V5 server if all the clients have

been upgraded at the same time?

- If you have upgraded all clients, then you will need to uninstall all clients and reinstall version 5.0 clients.

- That's why I recommended you test fully with a couple of clients before upgrading the rest.

- What you REALLY don't want is to roll back during business hours in the week, so you really need to test everything properly. Rolling back isn't trivial, and will at least lose any historical data since the upgrade.

How does this become a member of a fail-over cluster if we've upgraded

from existing config? How do we make it the first node in the cluster if

its already built?

- So you are saying that you will have a HA 'primary' cluster, and a warm-standby cluster using the old hardware? Or something else?

-Cold standby clusters are described on page 2-27 of the SRND - (http://www.cisco.com/en/US/docs/voice_ip_comm/cust_contact/contact_cente...)

What happens with the build of the fail-over server once the V7 server

has gone live? The same with the old server - how do we add it to the V7

cluster?

- I'm guessing here that by 'fail-over' server you mean the cold standby cluster? If so, refer to the SRND above.

- Adding the old server to the 7.x HA cluster (i.e. as an active/standby pair with auto failover) requires several things:

-- It needs to be same or better spec than the primary (for example, if the HDD is smaller it won't install)

-- You need HA licensing

-- You need to rebuild it from media, and install the same UCCX version and OS patches etc as the primary. Both nodes need to run full SQL rather than just MSDE.

-- You woule then run AppAdmin and use the 'Add to Cluster' option to provision it as a secondary, and enable all components.

Regards

Aaron

Please help to identify useful content in the forums by marking questions as 'answered' if you have received a suitable response..

https://supportforums.cisco.com/docs/DOC-6212

Attachment: 
clougher01 Thu, 04/29/2010 - 05:13

Ok so is there any benefit in having a spare server for this?

I'd like to upgrade the existing uccx 5 to 7 and make it the primary node in a HA cluster. Then add the spare server as the standby node.

So how do I make the existing node the primary srv? I saw the option in the original build but not sure how to change in appadmin now?

Aaron Harrison Thu, 04/29/2010 - 05:16

Hi

I'm not sure I follow you ?

You have build the new hardware as a UCCX7?

You would then rebuild your UCCX5 and do an 'add to cluster' to make the old UCCX 5 hardware the second node in your cluster.

Aaron

clougher01 Thu, 04/29/2010 - 05:20

The issue is we will lose the data if I make the existing ver 5 srv part of the new 7 cluster. So I will need to upgrade the ver 5 srv to 7 and make it the primary node in HA.

Correct Answer
Aaron Harrison Thu, 04/29/2010 - 05:25

Hi

OK - we're kind of going in circles here.

Option 1)

You in-place upgrade your UCCX5 to 7, and keep all your data. This doesn't give you an easy rollback. If you do this, you add your new server to the cluster afterwards after building it fresh from media.

Option 2)

You can backup UCCX5, build a new UCCX5 on the new hardware, restore the UCCX5 data to it, then upgrade that to version 7.0.

When you have done this, and switched clients to it, and you are happy you no longer need the original UCCX5.0 server, you can then rebuild the UCCX5.0 server from media and add it to the UCCX7 cluster that contains your upgraded data and data gathered since the upgrade.

Aaron

clougher01 Thu, 04/29/2010 - 05:31

Yup its a bit hard to explain via email but option 2 is what we would consider. How do we then make the existing 7 srv the primary HA node if its already built and running? Do we need to kick off a rebuild or is there a way to reset it?

Correct Answer
Aaron Harrison Thu, 04/29/2010 - 06:04

Hi

Ah - I see.

A single node is a primary node.

The only different (I believe) is that when you select 'Single Node' in the initial setup, it doesn't make itself a publisher. When you 'add to cluster' it does it at that point... so you don't need to meddle with the first UCCX node at all.

Regards

Aaron

clougher01 Thu, 04/29/2010 - 12:34

Ok that's all we need to know. Thanks for the answers. You were spot on..

Gabriel Saavedra Thu, 04/29/2010 - 13:57

Don't forget to rate the helpful posts.

All the posts from Aaron in this thread were pretty awesome.

Gabriel.

Actions

This Discussion