A policy change is being run against configurations that have over 120 interfaces, with one change, and is experiencing timeouts and failures. The network device logs show there is a delay of 30 seconds between each block, 15 seconds before running a command, and 30 seconds before running the next block. The network device TTY timeout is configured for 1 hour of inactivity timeout, not the duration of the session. The NCM upload script is configured with additional parameters ‘login_timeout’ set to 10 and ‘command_timeout’ set to 15. Running NCM policies against 17 devices with configuration changes takes over 4 hours.
So there is a need to decrease the timeouts (or wait times), remove the capture of running/startup configurations before doing the upload, and prevent the NCM process from starting a new session (leaving old sessions open) when making changes to a configuration causing previous session timeouts?
DX NetOps Spectrum Release : 20.2 or later
The cause of the delay was that the command timeout was set to 15 in the additional script parameters for the upload script.
Decreasing the command timeout in the additional script parameters for the upload script from 15 seconds to 2 seconds resolved this issue.