Migrating to v0.62.0
Before: historical release instructions
Historical release snapshot. For the current command surface, follow the latest migration and unreleased changes.
Local run retries
runs retry refuses local runs because a new attempt cannot recover the
original execution environment. Captured submission environments are deleted
when execution starts.
Queued local retries without an execution environment also fail before dispatch.
Previously:
sparkwing runs retry --failed --run run-fictional
Submit the pipeline again from its checkout, with the intended environment and original inputs:
sparkwing run build --sw-detached
This creates a new run; it does not resume the failed run's node outcomes. Controller-backed retries keep using their configured execution context:
sparkwing runs retry --failed --run run-fictional --profile prod