opensource/coder
3
0
Fork 0
mirror of https://github.com/coder/coder.git synced 2026-09-24 15:04:27 +08:00
Code Issues Packages Projects Releases Wiki Activity

fix: allow lifecycle code path to retry failed stop jobs (#26203)

Browse Source 
Co-authored-by: Mux <mux@coder.com>
This commit is contained in:
Callum Styan
2026-06-10 16:29:00 -07:00
committed by GitHub
co-authored by Mux
parent 0b99e67ce7
commit 4e2c9f9cae
4 changed files with 91 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files
coderd/autobuild/lifecycle_executor.go
+9 -6
View File
@@ -576,7 +576,7 @@ func getNextTransition(
return database.WorkspaceTransitionStop, database.BuildReasonAutostop, nil
case isEligibleForAutostart(user, ws, latestBuild, latestJob, templateSchedule, currentTick):
return database.WorkspaceTransitionStart, database.BuildReasonAutostart, nil
case isEligibleForFailedStop(latestBuild, latestJob, templateSchedule, currentTick):
case isEligibleForFailedCleanup(latestBuild, latestJob, templateSchedule, currentTick):
// Use task-specific reason for AI task workspaces.
if ws.TaskID.Valid {
return database.WorkspaceTransitionStop, database.BuildReasonTaskAutoPause, nil
@@ -688,14 +688,17 @@ func isEligibleForDelete(ws database.Workspace, templateSchedule schedule.Templa
return eligible
}
// isEligibleForFailedStop returns true if the workspace is eligible to be stopped
// due to a failed build.
func isEligibleForFailedStop(build database.WorkspaceBuild, job database.ProvisionerJob, templateSchedule schedule.TemplateScheduleOptions, currentTick time.Time) bool {
// If the template has specified a failure TLL.
// isEligibleForFailedCleanup returns true if the workspace is eligible to be
// stopped due to a failed build. A failed start is cleaned up by stopping it,
// and a failed stop is retried by issuing another stop. In both cases the
// remediation is a stop build.
func isEligibleForFailedCleanup(build database.WorkspaceBuild, job database.ProvisionerJob, templateSchedule schedule.TemplateScheduleOptions, currentTick time.Time) bool {
// If the template has specified a failure TTL.
return templateSchedule.FailureTTL > 0 &&
// And the job resulted in failure.
job.JobStatus == database.ProvisionerJobStatusFailed &&
build.Transition == database.WorkspaceTransitionStart &&
(build.Transition == database.WorkspaceTransitionStart ||
build.Transition == database.WorkspaceTransitionStop) &&
// And sufficient time has elapsed since the job has completed.
job.CompletedAt.Valid &&
currentTick.Sub(job.CompletedAt.Time) > templateSchedule.FailureTTL