Along with software upgrades, you might have to upgrade the hardware components in the IPT network. The hardware upgrades include upgrading the memory on the servers, voice gateways, and routers, and upgrading the server platform. This section discusses the upgrades to the server platform.
Three critical pieces of server resources are CPU, memory capacity, and hard disk size. Monitoring these resources should be a part of the day-to-day monitoring task for the CallManager administrators. Use RTMT in CallManager 4.0 and above to monitor the performance counters listed in Table 9-1. If you are using older versions of CallManager, include the performance counters listed in Table 9-1 in your monitoring list. Also, remember that you have to monitor these parameters for all the CallManager servers in the cluster.
Performance Object | Counter |
|---|---|
Memory | Available Kilobytes |
Available Megabytes | |
Cache Bytes | |
Demand Zero Faults/sec | |
Page Faults/sec | |
Page Reads/sec | |
Page Writes/sec | |
Pages Input/sec | |
Pages Output/sec | |
Pages/sec | |
Paging file | % Usage - _Total Instance |
% Usage Peak - _Total Instance | |
Logical disk (all instances) | % Disk Read Time |
% Disk Time | |
% Disk Write Time | |
Current Disk Queue Length | |
Disk Read Bytes/sec | |
Disk Write Bytes/sec | |
Free Megabytes | |
Physical disk (all instances) | % Disk Read Time |
% Disk Time | |
% Disk Write Time | |
Disk Read Bytes/sec | |
Disk Write Bytes/sec | |
% Disk Read Time | |
% Disk Time | |
% Disk Write Time | |
Processor (all instances) | % Processor Time |
Cisco CallManager | CallManagerHeartBeat |
CallsActive | |
CallsInProgress | |
FXOPortsActive | |
FXOPortsInService | |
FXSPortsActive | |
FXSPortsInService | |
PRIChannelsActive | |
PRISpansInService | |
RegisteredHardwarePhones | |
RegisteredMGCPGateway | |
TranscoderOutOfResources | |
TranscoderResourceActive | |
TranscoderResourceAvailable | |
Process (instance CCM) | Private Bytes |
Virtual Bytes | |
Virtual Bytes Peak | |
Working Set | |
Working Set Peak |
When you are monitoring the performance counters, if you notice too many peaks for page faults and page writes/reads, the server needs more RAM (provided that you have ruled out the possibility of memory leaks with other applications). As long as you have provisioned the system within the Cisco guidelines of device weights and dial plan weights, you should not encounter resource issues.
A replacement to server hardware is required in the following cases:
If the server reaches its End of Service (EoS) period as specified by Cisco or the hardware vendor
If a nonrectifiable component failure occurs on the server
If the call-processing capacity of the server needs to be increased
In all three cases, you can replace the hardware without issues if you have a full data backup available. Follow the same procedures described earlier in the "Recovery Using Backup Data" section to recover the data onto the new server.