SafeDisk AI

Self-Hosted macOS iOS Build Runner Disk Full

When a Mac build provider, iOS build gateway, or self-hosted runner fails with errno=28, write64() failed, or No space left on device, treat it as an operational policy problem: find the safe cleanup boundary before deleting Xcode state, simulator runtimes, archives, or customer build artifacts.

macOS runner storage check

Copy the runner check before deleting Xcode state.

Paste the job link, ENOSPC line, or read-only output and we reply with the safe first step. The free SafeDisk AI deletion trial is only for unclear delete / confirm / protect decisions such as simulator runtimes, archives, DerivedData, or customer build artifacts.

Add command output optional

Free trial step. We reply first with a delete / confirm / protect judgment or one read-only detail we still need.

Preview CI policy Ask AI about one file

Do Not Collapse These Buckets

Provider Recovery Order

  1. Mark the provider unhealthy and stop routing new builds to it.
  2. Capture read-only disk evidence and the exact path that failed to write.
  3. Remove current-job temp artifacts owned by the failed job first.
  4. Clean rebuildable Xcode and package caches only after no job needs them.
  5. Review simulator runtimes and required toolchains separately before deletion.
  6. Return the provider to the pool only after a post-cleanup scan shows enough headroom for the largest expected build.

Turn one ENOSPC build into a runner policy

Copy the read-only size check first. If the output still needs judgment, send the result for review before pruning volumes or resetting state.

Free AI deletion trial

Need a delete / confirm / protect answer for this runner?

Submit the form first; the failing job link or runner storage result can follow. We check whether the free first step is enough before asking for the free SafeDisk AI deletion trial.

Add output optional

Start with email. Output can follow after the first reply.