Mr.PlanB Logo
    NEWDedicated backup hub — PBS, LTO workflows, Kubernetes backup & recovery strategybackup.mrplanb.com ↗
    INDEX / STORAGE — 32 GUIDES · 8 CATEGORIES

    Storage you can reason about.

    Field-tested guides for homelab and lab-adjacent storage — from single-node ZFS to multi-node Ceph, with the network protocols and backup strategy to match. Built for people who read man pages.

    ZFS · Ceph · Longhorn · PBS · Unraid · Harvester · Rook-Ceph · NFS · iSCSI · LTO +7
    storage-pickeranswer 3 questions → get a recommendation
    How many nodes?
    Primary platform?
    What matters most?
    // 0 / 3 answered — recommendation pending
    06Backup Strategy8 guides
    183-2-1 Backup RuleThree copies, two media, one offsite — the foundational backup framework, explained with real architectures.↗193-2-1-1-0 Backup RuleThe ransomware-ready extension: one offline or immutable copy and zero unverified recovery errors.↗20Backup Retention PolicyDaily, weekly, monthly, yearly recovery points — keep enough history without paying to keep everything.↗21RPO vs RTOTurn “how much can we lose” into measurable recovery point and recovery time targets.↗22Backup TestingSeven levels of testing, from job status to a full replacement-cluster restore.↗23Disaster Recovery PlanTurn backups into a working recovery process — priorities, roles, dependencies, runbooks.↗24Backup vs Disaster RecoveryWhy having backups doesn’t mean you can recover — and what closes the gap.↗25Offsite BackupKeep one recovery copy outside the failure zone, with independent access and retention.↗
    08Business Workload Cluster6 guides
    27Kubernetes for Business WorkloadsDecide what actually belongs on the cluster — architecture, availability, storage, security.↗28Legacy Applications on KubernetesMigrate an older workload to Kubernetes without hiding its real operational problems.↗29Stateful Applications on KubernetesDatabases, queues, and search platforms — protect the state, not just the Pod.↗30Customer-Facing Applications on KubernetesKeep the service available, not just the Pods — releases, sessions, autoscaling, dependencies.↗31Multi-Tenant KubernetesShare the platform without sharing every risk — namespaces, RBAC, quotas, node isolation.↗32AI and GPU Workloads on KubernetesUse expensive accelerators without losing control — scheduling, sharing, cost, recovery.↗
    REFERENCE / DEEP DIVES
    Select a topic — full write-ups, no scroll wall.
    ZFS — the gold standard of data integrity

    ZFS — the gold standard of data integrity

    Read the guide ↗

    ZFS combines a file system and volume manager into one, offering end-to-end data integrity, snapshots, replication, and compression. It's the foundation for many homelabs running Proxmox VE or TrueNAS, ensuring your data remains consistent and recoverable even in case of hardware failure.

    +Bit rot protection through checksums
    +Efficient deduplication and compression
    +Native snapshots and replication
    +Ideal for virtual machines and containers
    ChecksumsSnapshotsCompression
    Ceph — distributed and scalable storage

    Ceph — distributed and scalable storage

    Read the guide ↗

    Ceph is a distributed storage system that unifies object, block, and file storage under one architecture. It's perfect for Kubernetes, OpenStack, and large-scale homelabs where high availability and scalability are essential.

    +Self-healing and self-managing design
    +Horizontal scaling without downtime
    +Integration with Rook-Ceph for Kubernetes
    +Excellent performance for multi-node clusters
    Self-healingScales outBlock + object + file
    Longhorn — lightweight Kubernetes storage

    Longhorn — lightweight Kubernetes storage

    Read the guide ↗

    Longhorn, developed by SUSE, is a lightweight, open-source block storage solution for Kubernetes. It offers replication, snapshots, and disaster recovery capabilities with minimal overhead — making it an ideal fit for edge and homelab environments.

    ReplicationDR-readyMinimal overhead
    Proxmox Backup Server (PBS)

    Proxmox Backup Server (PBS)

    Read the guide ↗

    PBS provides fast, deduplicated backups for Proxmox VE environments. Whether you're backing up VMs, containers, or physical hosts, PBS ensures data efficiency and integrity with built-in ZFS and encryption support.

    DeduplicationEncryptionProxmox-native
    Unraid — flexible and user-friendly storage

    Unraid — flexible and user-friendly storage

    Read the guide ↗

    Unraid combines NAS functionality, virtualization, and Docker management in a single platform. It's perfect for homelab users who want an easy-to-manage system with parity-based protection, dynamic disk allocation, and community plugin support.

    ParityDocker + VMsMixed drives
    Harvester — virtualization meets open-source storage

    Harvester — virtualization meets open-source storage

    Read the guide ↗

    Harvester, part of the SUSE ecosystem, integrates Kubernetes, KubeVirt, and Longhorn to deliver a hyperconverged infrastructure for homelabs and edge deployments. It simplifies managing compute, storage, and networking within a single cluster.

    KubernetesKubeVirtHCI
    Rook-Ceph — cloud-native storage orchestration

    Rook-Ceph — cloud-native storage orchestration

    Read the guide ↗

    Rook-Ceph brings Ceph into the Kubernetes ecosystem, automating deployment, scaling, and recovery. It's the preferred choice for those seeking reliable, self-managing block and object storage in containerized environments.

    Operator-managedAuto-recoveryK8s-native
    Network storage protocols: NFS, iSCSI, SMB/CIFS

    Network storage protocols: NFS, iSCSI, SMB/CIFS

    Read the guide ↗

    Different workloads need different network storage approaches. These protocols form the backbone of shared access in both homelab and enterprise setups.

    +NFS — simple, Unix/Linux-native file sharing protocol
    +iSCSI — block-level storage over IP, ideal for virtualization
    +SMB/CIFS — Windows-friendly and cross-platform file sharing
    NFSiSCSISMB/CIFS
    Backup strategies and data protection

    Backup strategies and data protection

    Read the guide ↗

    A strong storage system means nothing without reliable backups. Follow the 3-2-1 backup rule: three copies of your data, two types of media, one offsite or remote backup. Tools like Restic, BorgBackup, and Duplicati make it easy to automate and verify your backup routines.

    3 copies2 media1 offsite
    Choosing the right storage solution

    Choosing the right storage solution

    Read the guide ↗

    Your storage choice depends on your performance, capacity, and redundancy needs. For small homelabs, ZFS or Unraid offer simplicity and reliability. For scalable, cloud-native environments, Ceph, Rook-Ceph, or Longhorn provide flexibility and resilience. Experiment, benchmark, and build the storage architecture that fits your homelab vision.

    Small: ZFS / UnraidScale: CephK8s: Longhorn