Snapshots and backups for a cloud VICIdial box

A Server snapshot is one of the most convenient tools a cloud provider gives you: a full image of the VM's disk at a point in time, created in minutes and restorable just as fast. For a VICIdial box this is genuinely useful, but it covers a narrower set of failure modes than most admins assume.

This post covers when snapshots protect you, where they fall short, and how to build a backup strategy that actually covers your data if something goes wrong at the provider level.

What a cloud snapshot actually is

A Server snapshot captures the entire disk state of a VPS at the moment you trigger it. On most cloud providers this is a copy-on-write image stored in the same datacenter as your live server. You can boot a new server from that image in a few minutes, which makes it excellent for a fast rebuild after a botched update, a corrupted config file, or an Asterisk version upgrade that goes sideways.

What it does not protect you from is a datacenter-level event. If the storage backend that holds both your live disk and your snapshot image has a hardware failure, neither the live data nor the snapshot survives. This is rare, but real. It is also worth noting that snapshots include every Call recording on disk at the time, which means the snapshot grows as your recording archive grows, and snapshot storage costs scale accordingly.

When to take snapshots

flowchart TD
  A[Before any system change] --> B[Take snapshot]
  B --> C{Change successful?}
  C -->|Yes| D[Keep snapshot for 7 days]
  C -->|No| E[Restore from snapshot]
  D --> F[Delete old snapshots]
  G[Weekly scheduled snapshot] --> H[Retained for 30 days]
  H --> I[Off-box recording export]
  I --> J[Object storage / external backup]

Take a snapshot before every significant change: OS updates, VICIdial upgrades, new Conf file rollouts, or carrier trunk migrations. This gives you a clean rollback point without touching production. Beyond pre-change snapshots, a weekly automated snapshot with a 30-day retention window is a reasonable baseline for most operations.

Keep in mind that a snapshot of a running VICIdial box is technically a crash-consistent image, not an application-consistent one. Active calls, open database transactions, and Asterisk's in-memory state are not flushed to disk before the snapshot runs. For most recovery scenarios this does not matter, but it means you should not rely on a snapshot to recover a specific active call or an in-progress campaign report. The database will come up cleanly from InnoDB's crash recovery, so the practical risk is low.

Off-box backup for recordings

The most durable backup strategy for Recording retention requirements is to export recordings off-box on a daily or hourly schedule. An S3-compatible object storage bucket in a different region from your VPS gives you a copy that survives provider-level incidents. A simple rclone sync or aws s3 sync cron job pointing at /var/spool/asterisk/monitor/ is enough to automate this.

Database backups deserve the same treatment. A nightly mysqldump piped to gzip and synced to object storage is the standard approach. For Failover purposes, a database backup plus a snapshot is a recovery path that covers most realistic disaster scenarios.

Recovery time expectations

A full restore from snapshot typically takes 5 to 15 minutes depending on disk size and provider. After the new server boots you may need to update DNS and let the Provisioning settle before agents can log in. If you have exported recordings and a DB backup separately, restoring those on top of the snapshot brings you to a fully consistent state.

For the rebuild-from-scratch approach, see treating a cloud VICIdial box as rebuildable, which covers keeping your configuration reproducible so you are not dependent on a snapshot at all.

VICIfast snapshot handling

VICIfast is Single tenant Managed hosting, which means each customer's box is independent and snapshot operations do not affect other tenants. The platform manages snapshot scheduling and integrates with the cloud provider's snapshot API so you do not have to SSH in to trigger one. The full reasoning for choosing a managed cloud path over self-hosting is in the VICIdial cloud guide.

To see what backup and snapshot features are included in each plan, visit the pricing page and compare the add-on options.