Dagger cache volume "no space left on device" in CI

By Daniel Zoghalchali·Latchkey

The Dagger Engine keeps a layer cache and persistent cache volumes on disk. On a CI runner with a small volume, these can fill up mid-run, and a step fails with "no space left on device".

What this error means

A step fails with "write ...: no space left on device" or the engine reports the cache volume is full during an otherwise valid pipeline.

Error: input: container.withExec: failed to solve: write /var/lib/dagger/... : no space left on device

Common causes

The engine cache grew past the disk

A long-lived engine accumulates cached layers and cache volumes that are never pruned, exhausting the disk.

Large cache volumes on a small runner

Persistent cache volumes for dependencies can grow large and outstrip a modest runner disk.

How to fix it

Prune the engine cache

Reclaim space by pruning the Dagger Engine's build cache.

dagger core engine local-cache prune

Free disk and reset volumes

1. Remove unused preinstalled tooling to reclaim runner space.
2. Prune the engine cache between jobs on persistent runners.
3. Reset oversized cache volumes when they stop paying off.

docker exec dagger-engine buildctl prune

How to prevent it

• Prune the engine cache on persistent runners.
• Bound cache volume growth for large dependency sets.
• Use runners with enough disk for the cache.

Frequently asked questions

Related guides

References

• Docker documentation
• Docker build cache
• GitHub Actions documentation