cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
252
Views
0
Helpful
2
Replies

Maximum Number of Registered Devices Per CUCM Node

RL5901
Level 1
Level 1

The maximum number of registered devices per CUCM node is 10000, but this depends on the Deployment Type selected during the installation of the node as a virtual machine. The default setting in Service Parameters is 5000 but can be adjusted. 

We have a 4-node CUCM 11.0 (unsupported) cluster. One of our CUCM Subscribers has 5659 devices registered to it. Service Parameters (Cisco CallManager service) shows the Maximum Number of Registered Devices set to 5000 for this node. 

Questions:

  1. Given the Service Parameter is set to 5000 and there are 659 devices over this limit registered to it, what symptoms would result from this registration overage? 
  2. If I do not have access to the UCS that hosts the VM for this Subscriber node, how do I determine what number was selected for the Deployment Type at installation?
  3. What is the impact on phones, end users, and services of modifying the Maximum Number of Registered Devices for a node, and should other steps be performed if this is the chosen solution to this over-registration? 

 

1 Accepted Solution

Accepted Solutions

jari.suosilta
Level 1
Level 1

See  document "Cisco Collaboration System 11.x
Solution Reference Network Designs
(SRND)" :  Good sections to start:  9-13 and 10-48

2. SSH to servers OR login to system gui, it shows VM size, compare to installation OVA which size it was made:CPU/DISK/MEM.
10K user OVA has 4 CPU, Medium 2.Easy to compare to OVA-file. Download cucm OVA and open with text editor.
note at some point Cucm rls 11.0 or 11.5,  in later OVA-versions memory was increased from 6>8 gb for medium/large sized.
3. Load balancing devices. See that SNDR-document 9-13. 
To spread out evenly among, one can play with DEVICE POOLs> CMG (Call  Manager group) which is order of servers for registrations. simpliest copy devicepool,make new(or use existing)  CallManager group having different server order. Then configure part of device to  use new devicepool to have devices regirster evenly among server.

View solution in original post

2 Replies 2

jari.suosilta
Level 1
Level 1

See  document "Cisco Collaboration System 11.x
Solution Reference Network Designs
(SRND)" :  Good sections to start:  9-13 and 10-48

2. SSH to servers OR login to system gui, it shows VM size, compare to installation OVA which size it was made:CPU/DISK/MEM.
10K user OVA has 4 CPU, Medium 2.Easy to compare to OVA-file. Download cucm OVA and open with text editor.
note at some point Cucm rls 11.0 or 11.5,  in later OVA-versions memory was increased from 6>8 gb for medium/large sized.
3. Load balancing devices. See that SNDR-document 9-13. 
To spread out evenly among, one can play with DEVICE POOLs> CMG (Call  Manager group) which is order of servers for registrations. simpliest copy devicepool,make new(or use existing)  CallManager group having different server order. Then configure part of device to  use new devicepool to have devices regirster evenly among server.

RL5901
Level 1
Level 1

Thank you, Jari. 

In CLI, the Subscriber shows...

VMware Installation:
2 vCPU: Intel(R) Xeon(R) CPU E5-2630 v3 @ 2.40GHz
Disk 1: 110GB, Partitions aligned
8192 Mbytes RAM

 

A list of VM configuration templates and Table 25-4 (Maximum Number of Endpoints Supported Per VM Configuration) shows that this Subscriber was built to support a maximum of 7,500 users:

– 2 vCPU
– 8 GB vRAM
– 110 GB vDisk

Found in the SRND for 11.x chapter "Collaboration Solution Sizing Guidance" section on Using the Cisco Unified Communications Sizing Tool. Table 25-4 Maximum Number of Endpoints Supported Per VM Configuration and the list below it were used.