- Le plus récent
- Le plus de votes
- La plupart des commentaires
Hi Casey,
as you mention, the cloud to ShadowManager edge sync is based purely on the configuration of the ShadowManager component. You could do as you point out, include all possible DeviceId in the Shadow Manager configuration but that will imply: 1/ all Greengrass Cores sync all the shadows with an impact on local storage 2/ depending on the number of devices you might hit a deployment document size limit
My suggestion would be to deploy just the ShadowManager component using each core device as target. In this way each configuration can be tailored to reflect the devices which are attached to the Greengrass Core device. I would recommend to implement some automatic logic to manage these deployments.
I also encourage you to create a feature request in https://github.com/aws-greengrass/aws-greengrass-shadow-manager.
Cheers,
Massimiliano
Contenus pertinents
- demandé il y a un an
- demandé il y a un an
- Réponse acceptéedemandé il y a 7 mois
- AWS OFFICIELA mis à jour il y a 2 ans
Awesome, thank you.
I haven't experimented much with distributing deployments across targets, so I just want to make sure I understand. We can have a deployment that targets a group with the bulk of the component configuration, minus the ShadowManager, then create unique deployments targeting each GG-Instance with just the ShadowManager. Then Nucleus will handle merging those deployments together to create the desired effective-configuration?
Just tested with just that and it seems all good. The
effectiveConfig.yaml
looks like it's keeping track of which Components came from which Deployment, which is cool.For reference, I created this as well: https://github.com/aws-greengrass/aws-greengrass-shadow-manager/issues/125