Cisco Headend System Release 2.7 Technical References

 

 

 

Field Notice: 

Preventing Failures of the SRM 

Processes 

Our engineers are aware of a condition that may corrupt the memory of the 
System Resource Management (SRM) processes of the Digital Network Control 
System (DNCS). This condition can occur under all system releases of the Digital 
Broadband Delivery System (DBDS) software.  Memory of the SRM processes 
may sometimes become corrupted if the processes are not stopped and restarted 
in a specific order. 
Note:  The dsm, drm, qamManager, pkeManager, camEx, and camPsm processes 
together make up the SRM processes.  
The SRM processes are dependent upon one another. The SRM processes risk 
corrupting system memory should any SRM process be stopped and restarted 
without a stop and restart of the other SRM processes, as well. System operators 
will know that an SRM process has failed when the indicator light on the DNCS 
Control window changes from green to red for any of the individual SRM 
processes. New sessions will not start during the time the SRM processes are 
down. 

If system operators notice that any of the SRM processes are stopped, they should 
ideally stop and restart all DNCS processes.  
However, if a stop and a restart of all DNCS processes is not desirable, then a stop 
and restart of the aforementioned SRM processes ought to be conducted in the 
specific order set forth in the following sections.  Furthermore, the stopping and 
restarting of the SRM processes should be done while within a maintenance 
window. No new sessions can be set up while the SRM processes are down. 

System operators should stop any remaining SRM process that is still running in 
the following order: 
1

  dsm 

  drm