|
|
|
Specifies the maximum amount of time allowed between HRTT measurements before a prefix is determined to be unreachable through a service provider link.
outage-detection silence { enable | disable |
tslm_threshold <time> <multiplier> }
no outage-detection silence { enable | disable | tslm_threshold }
Must be in engine or policy configuration mode (config-engine or config-engine-policy), which requires privileged-level access.
CNA will consider a link no longer able to reach a prefix if both of the following conditions are met:
An excessive number of prefixes categorized as I (insufficient measurements) in output from the show prefix command can be a symptom of premature declaration of outages for low-traffic prefixes and at least one link categorized as o (outage detected), especially in systems with few service provider links. To avoid this condition, increase the value of time to a value larger than the typical inter-measurement time for the prefix.
When configured from engine mode, the command is applied globally, to all prefixes.
When configured from policy mode, the command is applied only to those prefixes which are associated with the policy.
For active probes, you should create a decision policy and set the multiplier value to 3. Setting the value any lower can cause premature declarations of outages and is not recommended.
NOTE: The per-link outage-detection algorithm compares measurement times across each of your service provider links to determine if there is an outage on any one of them. But if the currently asserted route experiences an outage, the single-pixel .gif that is used in the passive user traffic test will never be served to users in the prefix, and therefore no measurements will be collected. CNA must use a different mechanism to detect that condition. See the outage-detection total-silence command.
Use the no form of the command to revert to default values for time and multiplier.
|
|
|