Cinder: Resilient Edge Storage
Cinder is EmberNet's storage service. It gives your edge workloads durable block volumes, keeps replicas of them on other nodes where the cluster has more than one, and gives you a single view of capacity, volume health, and backups, with no external storage array required.
What it does
- Replicated volumes, where Cinder runs replicated. On a cluster with more than one node, a volume keeps copies on different nodes, so losing a disk or a node does not lose the data. Volumes created from EmberCORE default to three copies. A single-node site has nowhere to put a second copy, so its volumes are not replicated, and backups are what protect them.
- Automatic rebuilds. When a replica is lost, Cinder rebuilds it on healthy capacity on its own.
- Data stays close to the work. Cinder keeps a replica on the node running the workload wherever it can, so reads stay local.
- Honest capacity. Usage is the sum of what each volume actually holds on disk, not nominal volume sizes, so what EmberCORE shows is what you are really consuming.
- Guardrails. Cinder warns you when a disk is running short of headroom, because that is the point where a rebuilt copy may have nowhere to go.
Backups
Cinder backs volumes up to an external target of your choice: Amazon S3 or any S3-compatible store, Azure Blob Storage, or NFS. After you configure a target, use Test connection to confirm EmberNet can reach it with the credentials you gave it before you rely on it. Saving a target does not run the test for you.
CinderSilo, EmberNet's own object store for backups, is not available yet. It shows in EmberCORE as coming soon.
Storage alerts
Cinder watches volume health and capacity continuously and raises alerts on its own:
| Alert | Severity | When |
|---|---|---|
| Volume degraded | Warning | A volume has been running with fewer healthy copies than it should |
| Volume faulted | Critical | A volume has no healthy copy and cannot serve data |
| Capacity guardrail | Warning | A disk is past its capacity guardrail |
Every storage alert rings the bell in EmberCORE. Critical ones also go out by email and as a phone push notification, to your organization and to Fireball. Alerts resolve themselves when the problem clears. See Alerts & Notifications.
Using Cinder
Cinder is part of the platform, so there is nothing to deploy. Admins manage it from the Storage view in EmberCORE, which is documented in detail under EmberCORE Storage.
Next steps
- Storage (EmberCORE view): capacity, volumes, backups
- Flux · AnvilMQ