There is now a fix for businesses experiencing mysterious crashes while using Cisco’s Network Services Orchestrator (NSO) software.
A multivendor system called NSO connects applications in business and service provider contexts with network automation and orchestration tools, as well as the underlying physical and virtual infrastructure.
The networking behemoth recently issued a field notice describing a procedure in the programme that results in a leak of resources and a subsequent crash; the interval between crashes is only dependent upon the resources allotted to the software.
Twenty NSO 5 versions, ranging from 5.6 to 5.8.2, are affected.
The issue is that subscribers can access an API named cbd-diff-iterate, but Cisco also uses it internally for licencing and notification needs.
Customers that utilise Smart Licensing, notification-kicker, or who directly contact the API through C, Python, Java, or econfd are impacted by this.
According to Cisco’s field notice, “Any change under /kickers/notification-kicker or /devices/device will activate an internal subscription, which results in the leak.”
There is no way to disable the internal subscribers.
“Once sufficient transactions have been completed, the Cisco NSO process will be halted owing to resource exhaustion… The resources allotted to Cisco NSO will determine how many transactions may be completed before termination, according to Cisco.
NSO 5.6.7.1, 5.7.5.1, and 5.8.2.1 are the three versions that have been made available to address the problem.
Cisco advises frequent restarts for administrators who are unable to implement the upgrades quite yet.
