- Docs
- Storage
Storage
- Overview The storage backends on a Syself Autopilot cluster, block, local, shared, and object, and when each one fits.
- Access modes and volume binding What ReadWriteOnce and ReadWriteMany mean on a Syself Autopilot cluster, and why the pod is scheduled first and the storage follows.
- Storage classes reference Every StorageClass on a Syself Autopilot cluster with its provisioner, access mode, reclaim policy, expansion, and binding mode.
- Choose storage for a workload Pick the right storage on Syself Autopilot from two questions, how many writers and what happens to the data when a node is replaced.
Block storage
- Use Hetzner Cloud volumes Create a PVC on the standard StorageClass, mount it in a pod, and understand why a Hetzner Cloud volume binds only when a pod is scheduled.
- Expand a Hetzner Cloud volume Grow a Hetzner Cloud volume in place by editing the PVC size, and understand why a volume can only grow, never shrink.
- Volume attach constraints and reclaim policies Cloud volumes attach to one cloud node at a time and are kept, not deleted, when the PVC goes away, so clean them up to stop paying.
- Run a single-writer workload on block storage Wire a StatefulSet or a Recreate Deployment to a cloud volume so it survives node drains without deadlocking on the RWO handoff.
Local storage
- Set up local NVMe with TopoLVM Install TopoLVM, prepare each bare metal server's data disks with LVM, and get the local-nvme, local-ssd, and local-hdd storage classes.
- Local PV lifecycle Follow a local PVC from thin logical volume to reclaimed space, and free the disk that a Retain policy leaves behind after you delete the claim.
- Data survival across reprovision Syself Autopilot rewrites the OS disk on every reprovision and never touches your data disks, so local data survives in place unless the workload moves to another server.
- Performance of local vs network storage Local disks skip the network hop, so latency stays low and predictable, which is why databases belong on them instead of network-attached storage.
- Longhorn and node replacement Keep Longhorn data when Syself Autopilot replaces nodes. Switch Longhorn pools to manual node replacement and evict each node before you delete it.
Shared storage
- ReadWriteMany with JuiceFS Stand up a shared ReadWriteMany filesystem on Syself Autopilot with JuiceFS, backed by S3-compatible object storage and a Redis or PostgreSQL metadata engine.
- Shared storage with NFS Serve a ReadWriteMany filesystem on Syself Autopilot with NFS, using an in-cluster server or the NFS subdir provisioner for dynamic PVCs.
- Choose an RWX backend When to share data through object storage instead of a shared filesystem on Syself Autopilot, and how to choose JuiceFS or NFS when you do need ReadWriteMany.
Object storage
- Run SeaweedFS for S3-compatible object storage Deploy SeaweedFS inside a Syself Autopilot cluster as your own S3-compatible object store for buckets, backups, and JuiceFS.
- Use S3-compatible storage from workloads Point an application on Syself Autopilot at an S3-compatible endpoint for uploads, artifacts, and exports instead of a block volume.
Data protection and DR
- Backup layers and what needs protecting The two kinds of data on a Syself Autopilot cluster, declared and runtime, and the backup each one needs.
- Back up cluster state with Velero Install Velero against your object storage bucket and capture Kubernetes resources plus PVC data with the node agent.
- Volume snapshots and their limits Why Syself Autopilot backs up PVC data by file copy instead of CSI volume snapshots, and what that means for your data.
- Run a restore drill Rehearse restoring a Syself Autopilot backup into a scratch namespace so you know it works before a real outage.
- Offsite and immutable backups Keep a copy of your Syself Autopilot backups in a second region and lock it with object-lock so ransomware cannot delete it.
- Disaster recovery for stateful data A concrete plan to rebuild a lost Syself Autopilot cluster and bring every stateful workload back to a known good point.
Storage operations
- Encrypt data at rest with LUKS Encrypt local disks at rest on Syself Autopilot with LUKS to meet ISO 27001 and similar requirements. Encrypt once, unlock the disks at boot, and manage your own key.
- Monitor storage capacity Watch PVC usage, LVM thin-pool fill, and TopoLVM node capacity so you catch a full disk before a workload does.
- Tune storage performance Get more out of the disk you have with the right storage class, filesystem, thin-pool sizing, cache, and resource requests.
- Migrate data between storage classes Move a workload from cloud volumes to local NVMe, or the reverse, by copying into a new PVC on the target class without losing data.
- Troubleshoot a stuck PVC Every reason a PVC sits Pending, a pod will not mount, or a PVC will not delete, with the fix for each symptom.