I need shared storage for a mix of Linux VMs and containers, with snapshots and backup retention. What would you recommend?

For a mix of Linux VMs and containers, with shared storage plus snapshots and backup retention, I’d recommend one of these two approaches depending on your scale and tolerance for complexity:

Best all-around recommendation: ZFS-based storage

Use ZFS on a storage server or appliance and export it via:

• NFS for Linux VMs and container hosts
• iSCSI if you need block storage for VM disks
• optionally SMB if Windows clients ever appear

Why ZFS fits your needs

• Snapshots: fast, space-efficient point-in-time recovery
• Replication: easy incremental send/receive to a backup system
• Data integrity: checksumming and self-healing
• Retention: built-in snapshot scheduling and pruning
• Shared storage: works well for VM datastores and container persistent volumes

Good setup pattern

• Primary storage: ZFS pool on redundant disks/SSDs
• Snapshots: frequent local snapshots, e.g. every 15 min / hourly / daily
• Backups: replicate snapshots to a second system or object storage
• Retention policy: keep short-term local snapshots and longer-term backups

When to use what protocol

• NFS: simplest for Linux VM images, container volumes, shared files
• iSCSI: better if your hypervisor wants block storage
• NFS + ZFS is usually the easiest and most flexible for Linux-heavy environments

---

If you want an integrated storage platform: TrueNAS

If you want something easier to manage than building everything by hand, consider:

• TrueNAS SCALE if you want Linux-based, apps/containers, ZFS
• TrueNAS CORE if you prefer the classic FreeBSD-based option

Why

• Web UI
• Snapshot scheduling
• Replication jobs
• Shares for NFS/SMB
• Good fit for backups and retention policies

---

If you want enterprise distributed storage: Ceph

Choose Ceph if you need:

• scale-out storage across multiple nodes
• high availability
• block, file, and object storage
• tolerance for node failures

Tradeoffs

• More complex to deploy and operate
• Usually overkill unless you have several servers and need serious availability/scalability

---

What I would avoid as the primary answer

• Plain ext4/XFS on a single server without snapshots/replication
• Generic NAS without strong snapshot/replication support
• RAID alone as a “backup” solution

---

Practical recommendation by size

Small to medium environment

ZFS + NFS + snapshot replication

• simplest strong solution
• excellent for Linux VMs and container storage

Need a polished appliance

TrueNAS SCALE

• good balance of features and manageability

Larger / highly available environment

Ceph

• only if you truly need distributed, fault-tolerant shared storage

---

Backup note

Snapshots are not backups by themselves. For retention, I’d do:

• local snapshots for quick rollback
• off-box replication for disaster recovery
• periodic backup copies to another system or object storage
• keep at least one backup copy offline or immutable if possible

If you tell me:

1. how many VMs/containers,
2. your budget,
3. whether you need high availability,
4. what hypervisor/orchestrator you use,

…I can suggest a more precise architecture.