High Availability

Workload resilience with replicated storage

Run Skitflow as a Docker Swarm cluster with LINSTOR replicated storage. When a node goes down, your containers restart on another node, and their data is already there, in sync.

Read the docs
Built for uptime

Resilience without a database cluster per app

Docker Swarm handles orchestration, LINSTOR handles your data. Together they keep services and volumes available through a single-node failure.

Automatic failover

Docker Swarm watches every service and reschedules containers onto healthy nodes the moment one drops, no manual intervention.

Replicated block storage

LINSTOR mirrors each volume across nodes over DRBD, so the data a container needs is already present wherever Swarm restarts it.

No data loss on node loss

A replicated volume survives a single-node failure intact, and existing volumes keep serving I/O even while a node is replaced.

Architecture

Your app and its data, on every node

Docker Swarm orchestrates the containers on top; LINSTOR mirrors the volumes with DRBD underneath. Lose a node and Skitflow restarts the app on another that already holds an in-sync replica.

How failover works

What happens when a node dies

Three steps, all automatic.

1 · Storage is replicated

You mount a LINSTOR volume with replicas across the cluster. DRBD keeps every copy byte-for-byte in sync as your app writes.

2 · A node fails

Hardware, network or a reboot takes a node offline. Swarm detects the unhealthy node and your data stays safe on the surviving replicas.

3 · The app restarts elsewhere

Swarm reschedules the container onto a healthy node, LINSTOR routes its I/O to a local replica, and your service is back, with all its data.

Build a cluster that shrugs off failure

Set up Docker Swarm with LINSTOR replicated storage and keep your apps and data available through node failures.

Talk to sales