Query on CUCM migration from 8.5.1 to 10.5.1 using PCD

[Solomon Rufus Kavala]

Hello Team,

We are migrating our Physical cluster (Pub & Sub) 8.5.1 SU2 to UCS 10.5.1 using Cisco PCD. We have installed PCD and discovered existing cluster to it. While performing migration task, can we proceed with Pub & Sub at same time or you want us to do one at a time. I need to know, when we start migrating from existing cluster to new cluster, it wont create any outage right??? Please make me understand these queries, so that we will have a clean migration.

Thanks in advance,

Solomon.

[Aman Soi · Accepted solution]

Hi Solomon,

I understand that u are using same IP addresses for new CUCM 10.

If u are using different then, existing cluster would keep on working as it is and there would be no outage .Once u new setup is established, u can change tftp IP for new one so that phone scan get regsitered to CUCM 10

I am capturing the snapshots from  link which will help for understanding

Migrate Cisco Unified Communications Manager to Release 10.5(1)

The first Unified Communications Manager publisher is migrated. Prime Collaboration Deployment shuts down the current Unified Communications Manager publisher automatically.The migration task pauses once the publisher is successfully migrated to Unified Communications Manager 10.5(1).

Add a Migration Task

In the Specify Migration Procedures section, set the sequence for the migration task. Following is the recommended sequence:

a) Migrate Publisher node. Add Pause.

b) If dedicated TFTP nodes are deployed in your environment, migrate dedicated TFTP nodes.

c) Migrate all backup subscriber nodes. Add pause.

d) Migrate all primary subscriber nodes.

When you create a migration task, Prime Collaboration Deployment presents the default sequence, which has

one server in each install step. You can use the editing tools to place more than one server in a step. For best

results, do not include more than six servers in any one step.

• Pencil icon—Opens the Edit Step window.

• Page icon—Adds a new step after the current step.

• X—Deletes the current step.

• Up arrow—Moves the step up to be executed earlier.

• Down arrow—Moves the step down to be executed later.

Note In the Edit Step window you can add nodes to bemigrated fromthe list of available nodes:

• The step to which each node is assigned displays next to the node. If a node is not assigned to any step,

the node shows as unassigned.

• The Pause task after step completes option pauses the task after completion of this step. If you choose

this option, you must start the next step manually to continue the task.

Step 10 Click Next.

Step 11 Review your selections. In the Review section, you also have the option to add notes to your new migration

task.

Step 12 Click Finish to add the new migration task. The task appears in the table on the Migrate screen.

Step 13 Perform a DRS backup of your Unified Communications Manager cluster.

http://www.cisco.com/c/en/us/td/docs/voice_ip_comm/cucm/ucmapMigrate/10_5_1/CUCM_BK_M24251C0_00_migrate-procedure-for-cucm_1051/CUCM_BK_M24251C0_00_migrate-procedure-for-cucm_1051_chapter_00.html

regds,

aman

[Solomon Rufus Kavala]

Aman,

Thank you very much for your advice. As per your suggestion, we went ahead and performed migration, but for some reasons PCD did not take the password credentials which we gave for our existing 8.5.1 cluster, be it platform admin credentials or Security Passwords. So we are forced to give new credentials. After PCD Installation, we added migration task and then added old & new cluster information. Immediately we ran CUCM console at our ESXi Host to see the progress of PCD script. But to our surprise, it stopped over the step where we give admin credentials, we gave the same credentials which we gave in PCD instead of existing CUCM cluster. Then we ran all the remaining steps manually. At the end we see CUCM 10.5.1 installed successfully. But old cluster information was not migrated to the new one. I have attached the screenshot of PCD for your reference. As per the attachment, second step was taking ages to get completed. I suspect some issue and because of that it is not progressing further. Need your expert thoughts. So that we will complete our migration successful.

Thanks in advance,

Solomon.

[Aman Soi]

Hi Solomon,

the problem would be due to bug .Were the password changed before u started migration .I mean to say to u need to use the credentials which u entered while installing 8.5 cluster first time.

PCD CUCM Data export fails due to permission admin user permission issue

CSCup43027

Description

The customer cannot continue with Migration of CUCM cluster to a new version

Symptom:Data export fails from CUCM cluster

Conditions:The credential used to discover the CUCM cluster was not the original credential created during install of the CUCM cluster.

Workaround:The site should use the credential that was created during the CUCM cluster install

regds,

aman

[Solomon Rufus Kavala]

Aman,

Thanks again for your quick response. So If I understood it correct, I need to use the same credentials (be it platform admin, Application admin or Security passwords), which I have used for my existing 8.5.1 old cluster. Is that right??? 

In fact, I gave the same credentials but PCD installation throw-ed me an error stating select proper password and did not allowed me to use old password, which I have used it in my existing 8.5.1 cluster. So in this scenario, how should i proceed further.

Solomon. 

[Aman Soi]

Hi Solomon,

I can share my experience:

we did one of the migrations for call manager version 7 and entered the credentials  but did not work.We opened up TAC case and came to know that we were entering incorrect credentials.

Findings of TAC :

Initally the server was build with version call manager  5 and migrated to version 7 .After that password was changed.TAC took the root access.

you can login PCD and run the command if any error comes up.

file get activelog tomcat/logs/ucmap/log4j/*

I would recommend you to open TAC case.

regds,

aman

[Solomon Rufus Kavala]

Thank you Aman, Will do the same. Or else i also can proceed with fresh installation of CUCM 10.5.1 on the ESXi Host and export/import all the necessary files through BAT from existing cluster. What do you say?

Solomon.

[Aman Soi]

I suggest opening case.

I am confident problem would get resolved by TAC.Once done, migration would be very smooth.Since this problem is coming, it is taking too many hours.

Do not go for export/import.

regds,

aman

[Aman Soi]

Hi Solomon,

Is the issue resolved?

regds,

aman

[Solomon Rufus Kavala]

Hi Aman,

We did opened TAC case as per your suggestion and TAC engineer asked me to change admin credentials of old Physical server (8.5.1) to match with the new server. After changing those credentials, We have rebooted the old cluster and then rediscovered Old cluster with new admin credentials in PCD. Then created a new migration task and found PCD successfully migrated the old database to new 10.5.1. 

Now I am yet to change the IP Address of CUCM (TFTP) under DHCP scope. Tell me one thing, when we change the TFTP IP either from local gateway DHCP scope or from DHCP Server, will all the IP Phones automatically register to new cluster after bouncing switch ports connected to IP Phones or do we need to do a factory reset of all phones to get them register to new cluster.

Need your expert advice.

Regards,

Solomon.

[Aman Soi]

Hi Solomon,

I can suggest few options:

a.try RESET for one IP phone not Restart from Old Call Manager server.

b. In switch , u can run command show ip dhcp binding and then, run clear ip dhcp x.x.x.x [ where x.x.x.x is IP address of one IP phone]

c. Lastly, reset switch.

No need to do factory reset for IP phones.

regds,

aman

[Solomon Rufus Kavala]

Hi Aman,

Thank you for your suggestion. After updating the CTL Update on both clusters, Phone registration went on smooth from old cluster to new one. I have another question. We have around 10 SRST branch sites, which are running 15.X IOS. Can you please let me know, is there any specific version of IOS need to install on SRST Gateways for CUCM 10.5.1.

Based on the below link:

http://www.cisco.com/c/en/us/td/docs/voice_ip_comm/uc_system/unified/communications/system/versions/IPTMtrix.html

I have found Cisco Unified Survivable Remote Site Telephony (SRST)  need to get upgraded to 5.3(3)M3 or 15.4(1)T1

Under Communication's infrastructure Cisco 2801, 2811, 2821, 2851, 3825, 3845 Integrated Services Router to be 15.1(4)M7 and Cisco Integrated Services Router (ISR) 2900 and 3900 Series to be 15.3(3)M3 or 15.4(1)T1. 

Need your valuable thoughts to proceed further, We have a mixture of 28XX & 29XX routers. 

Regards,

Solomon.

[Aman Soi]

Hi Solomon,

yes, u need to upgrade the IOS of router whether ISR or ISR G2 to compatible version.

http://www.cisco.com/c/en/us/td/docs/voice_ip_comm/cusrst/requirements/guide/srs10spc.html

http://www.cisco.com/c/en/us/td/docs/voice_ip_comm/cucme/requirements/guide/33matrix.html

Do check for the DRAM and Flash before upgrading router.

regds,

aman

[aneesh.abraham]

Hi,

    Can we directly migrate from cucm 8.5.1 to 11 using PCD? or we need to upgrade 8.5 to 8.6 before migrate. 

Regards,

Aneesh

[Jaime Valencia]

See table 1

http://www.cisco.com/c/en/us/td/docs/voice_ip_comm/cucm/pcdadmin/11_0_1/CUCM_BK_P00E3B55_00_pcd-admin-guide-115/CUCM_BK_P00E3B55_00_pcd-admin-guide-115_chapter_011.html