Repository navigation
Recover confirmed lost compute - #667
Merged
Merged
Conversation
SaladDay
marked this pull request as ready for review
October 11, 2026 08:49
SaladDay
force-pushed
the
recover-confirmed-lost-compute
branch
from
October 12, 2026 04:12
a6e0e9c to
7cb3527
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
A crashed or stopped sandbox could leave a settled running allocation owned indefinitely after its node returned. Core now releases that allocation only after the owning node confirms exact-incarnation absence or termination, exact cleanup succeeds, and the initialized retained Session remains eligible for continuation. Device revocation, allocation release and retention serialize under the existing Session lock and compute revision check; offline or uncertain writers remain owned.
Validation passed: real PostgreSQL recovery/guard/concurrency tests, sandbox tests, touched-package vet, documentation/translation checks and independent review. Native microsandbox/NFS acceptance passed for Codex and Claude Code: idle and mid-Turn VM loss, immediate post-acknowledgement loss (kill upper bounds 0.374s/0.385s), identical native identities, retained files, no replay and physical deletion. Codex also qualified the common node-offline, paused-VM and daemon-only negative cases. Missing native history failed explicitly without a new native root; diagnosis remains generic
engine_failedorenvironment_unavailable. Host-crash durability and arbitrary crash-tail completeness remain unqualified.