seamless NSO HA

[kbarve]

What is the option for seamless NSO HA? Is there any document with configuration details?

[KJ Rossavik]

Please see the documentation that comes with the Tail-f HCC component.

[kbarve]

Thanks Kjetil.

I got the NSO deployment guide in tail-f HCC package.

I am able to bring up the HA configuration, Fail over from Master to Slave is seamless but Fail-back (Reverting to original master) is manual procedure

i.e. role-override on master node from master to slave and then after sync-up role-revert on both nodes to bring back the system in original state.

My Question: is there any automated way to do Fail-back instead of this manual procedure of role-override / role-revert ? Or any suggestion to automate it?

Regards

Kirti

[Jan Lindblad]

This is by design in HCC. If you would like another policy, you can either add that to your own copy/branch of the HCC package, or use a different High Availability Framework (HAFW) solution.

[kbarve]

Hi Jan,

Thanks for the reply.

Could you please point to other HAFW options where Failover/Failback can be automated ?

Regards

KIrti

[Jan Lindblad · Accepted solution]

Chapter 8 in the NSO 4.4 Admin Guide talks about the NSO HA Architecture and APIs. If you would like to use another High Availability Framework (HAFW) than HCC, you can implement one with your own policies. This can be as simple as a 20-line bash script with keepalived that fails back and forth automatically, all the way to an integration with a commercial 5-digit dollars HA product from a third party.

[kbarve]

Another Question.

When there is a switchover between Master and Slave, is there a way get a notification to external system? (may be in form of snmp/log/...) ?