High Availability Login Profiles
Important Notes About High Availability Login Profiles
-
You can use the High Availability login profile tables in this section to configure the upper and lower client re-login values for your presence redundancy group. You configure the upper and lower client login values by choosing , and choosing Cisco Server Recovery Manager from the Service menu.
-
High Availability client login profiles apply only to single cluster deployments. High Availability client login profiles cannot configure the upper and lower client re-login values for the redundancy group if multiple clusters are present. You must perform more tests to discover High Availability client login profiles in multiple cluster deployments.
-
If Debug Logging is enabled for the Cisco XCP Router service, then you should expect increased CPU usage and a decrease in the currently supported logging levels for IM and Presence Service.
-
By configuring the upper and lower client re-login limits on your presence redundancy group based on the tables we provide here, you can avoid performance issues and high CPU spikes in your deployment.
-
We provide a High Availability login profile for each IM and Presence Service node memory size, and for each High Availability deployment type, active/active or active/standby.
-
The High Availability login profile tables are calculated based on the following inputs:
-
The lower client re-login limit is based on the Server Recovery Manager service parameter "Critical Service Down Delay", for which the default is 90 seconds. If the Critical Service Down Delay is changed then the lower limit must also change.
-
The total number of users in the presence redundancy group for Active/Standby deployments, or the node with highest number of users for Active/Active deployments.
-
-
You must configure the upper and lower client re-login limit values on both nodes in a presence redundancy group. You must manually configure all these values on both nodes in the presence redundancy group.
-
The upper and lower client re-login limit values must be the same on each node in the presence redundancy group.
-
If you rebalance your users, you must reconfigure the upper and lower client re-login limit values based on the High Availability login profile tables.
Use High Availability Login Profile Tables
Use the High Availability login profile tables to retrieve the following values:
- Client Re-Login Lower Limit service parameter value
- Client Re-Login Upper Limit service parameter value.
Procedure
| Step 1 |
Choose a profile table based on your virtual hardware configuration, and your High Availability deployment type. |
| Step 2 |
In the profile table, choose the number of users in your deployment (round up to the nearest value). If you have an active/standby deployment, use the node with the highest number of users. |
| Step 3 |
Based on the Number of Users value for your presence redundancy group, retrieve the corresponding lower and upper retry limits in the profile table. |
| Step 4 |
Configure the lower and upper retry limits on IM and Presence Service by choosing , and choosing Cisco Server Recovery Manager from the Service menu. |
| Step 5 |
Check the Critical Service Down Delay value by choosing and choosing Cisco Server Recovery Manager from the Service Menu. The default value is 90 seconds. The lower retry limit should be set to this value. |
Example High Availability Login Configurations
Example 1: 15000 Users Full UC Profile - active/active deployment
You have 3000 users in your presence redundancy group, with 2000 users on one node, and 1000 users on the second node. For an unbalanced active/active deployment, Cisco recommends you use the node with the highest number of users, in this case the node with 2000 users. Using the 15000 users full US (4 vCPU 8GB) active/active profile, you retrieve these lower and upper retry values:
|
Expected Number of Active Users |
Lower Retry Limit |
Upper Retry Limit |
|
2000 |
120 |
253 |
 Note |
The upper retry limit is the approximate time (seconds) it takes for all clients to login to their backup node after a failover occurs. |
Note |
The lower limit of 120 assumes the Critical Service Down Delay service parameter is set to 120. |
Example 2: 5000 Users Full UC Profile - active/active deployment
You have 4700 users on each node in your presence redundancy group . Cisco recommends that you round up to the nearest value, so using the 5000 users full US (4 vCPU 8GB) active/active profile you retrieve the lower and upper retry value based on a number of users value of 5000:
|
Expected Number of Active Users |
Lower Retry Limit |
Upper Retry Limit |
|
5000 |
120 |
953 |
Feedback