Hi, In last TSC meeting, resiliency is MUST for all services. As part of this testing, we are doing following for some services we have developed (e.g SMS). Please let us know if this is good enough. SMS has three containers - Secret Service, vault service (which is used by Secret Service) and Consul (which is used by Vault and Secret Service). External services to SMS only call APIs implemented by Secret Service.
This is what we did: - Since scale-out is not mandatory, we set "replica count" to 1 for all three containers. Testing we did: - Configure SS with set of secret domains/secrets. - Kill Consul. K8S bring it up with no user intervention (due to replica count is 1) o Ensure that secrets/domains are still available via SS API - Kill Vault. K8S bring it up with no user intervention (due to replica count is 1) o Ensure that secrets/domains are still available via SS API - Kill SS. K8S bring it up with no user intervention (due to replica count is 1) o Ensure that secrets/domains are still available via SS API - Restart the entire machine where ONAP is installed. o Ensure that secrets/domains are still available via SS API (after ONAP is brought up). Are there any scenarios integration team intends to do for important services that come in the way of volte, vcpe use cases? And we will try to do same for SMS too. Thanks Srini
_______________________________________________ onap-discuss mailing list onap-discuss@lists.onap.org https://lists.onap.org/mailman/listinfo/onap-discuss