lei.chen/myxce
1
0
Fork 1
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

fix(apiserver): stop job-control panic on int/int64 type mismatch

Browse source 
xcecli jobs kill/stop/resume/checkpoint failed for every job with an
empty error body. models.JobInfo.JobStatus/.ArrayIdx are plain Go
`int`, but validateJobControl/foundJob asserted them as `int64` via a
bare type assertion, which always panicked; Gin's Recovery()
middleware swallowed the panic and returned an empty 500.

Fixes gitea#45 / bug-1742.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Y4M5MJM2zK6nU43da267Ku
This commit is contained in:
Rongsong Shen 2026-09-01 06:37:04 +00:00
parent 83d40fbb33
commit 343639fc7e
No known key found for this signature in database
GPG key ID: 5BCDAAE0123E9FE2
3 changed files with 128 additions and 10 deletions
Download patch file Download diff file Expand all files Collapse all files

39
src/frontend/apiserver/controllers/jobs.go
View file

@ -3010,8 +3010,24 @@ func validateJobControl(jobinfo utils.OrderMap, action string,
actionType = models.JOB_KILL
}
// bug-1742: array_idx is a plain Go `int` on models.JobInfo, not `int64`
// -- utils.IntOrInt64 accepts either, unlike a bare `.(int64)` assertion.
arrayIdx, arrayOk := utils.IntOrInt64(jobinfo.Data["array_idx"])
if !arrayOk {
logger.Error("array_idx is not an int/int64", zap.Any("array_idx", jobinfo.Data["array_idx"]))
failedJobs = append(failedJobs, map[string]interface{}{
"unique_job_id": jobinfo.Data["unique_job_id"].(int64),
"job_id": jobinfo.Data["job_id"].(int64),
"array_idx": arrayIdx,
"msg": fmt.Sprintf("Invalid array_idx"),
})
return successJobs, failedJobs
}
if rawStatus, exists := jobinfo.Data["job_status"]; exists {
if jobStatus, ok := rawStatus.(int64); ok {
// bug-1742: job_status is likewise a plain Go `int`, not `int64`.
if jobStatusInt, ok := utils.IntOrInt64(rawStatus); ok {
jobStatus := int64(jobStatusInt)
db_job_status := int(jobStatus)
job_status := models.JobStateStr(db_job_status)
jobStatus = int64(models.JOBSTAT[job_status])
@ -3021,14 +3037,14 @@ func validateJobControl(jobinfo utils.OrderMap, action string,
failedJobs = append(failedJobs, map[string]interface{}{
"unique_job_id": jobinfo.Data["unique_job_id"].(int64),
"job_id": jobinfo.Data["job_id"].(int64),
"array_idx": int(jobinfo.Data["array_idx"].(int64)),
"array_idx": arrayIdx,
"msg": fmt.Sprintf("job cannot be <killed> or <signaled>"),
})
} else {
successJobs = append(successJobs, map[string]interface{}{
"unique_job_id": jobinfo.Data["unique_job_id"].(int64),
"job_id": jobinfo.Data["job_id"].(int64),
"array_idx": int(jobinfo.Data["array_idx"].(int64)),
"array_idx": arrayIdx,
"msg": fmt.Sprintf("job is being <%s>", action),
})
}
@ -3037,7 +3053,7 @@ func validateJobControl(jobinfo utils.OrderMap, action string,
successJobs = append(successJobs, map[string]interface{}{
"unique_job_id": jobinfo.Data["unique_job_id"].(int64),
"job_id": jobinfo.Data["job_id"].(int64),
"array_idx": int(jobinfo.Data["array_idx"].(int64)),
"array_idx": arrayIdx,
"msg": fmt.Sprintf("job is being <%s>", action),
})
} else {
@ -3045,16 +3061,16 @@ func validateJobControl(jobinfo utils.OrderMap, action string,
failedJobs = append(failedJobs, map[string]interface{}{
"unique_job_id": jobinfo.Data["unique_job_id"].(int64),
"job_id": jobinfo.Data["job_id"].(int64),
"array_idx": int(jobinfo.Data["array_idx"].(int64)),
"array_idx": arrayIdx,
"msg": fmt.Sprintf("job cannot be <%s>", action),
})
}
} else {
logger.Error("job_status is not an int64", zap.Any("job_status", rawStatus))
logger.Error("job_status is not an int/int64", zap.Any("job_status", rawStatus))
failedJobs = append(failedJobs, map[string]interface{}{
"unique_job_id": jobinfo.Data["unique_job_id"].(int64),
"job_id": jobinfo.Data["job_id"].(int64),
"array_idx": int(jobinfo.Data["array_idx"].(int64)),
"array_idx": arrayIdx,
"msg": fmt.Sprintf("Invalid job_status"),
})
}
@ -3063,7 +3079,7 @@ func validateJobControl(jobinfo utils.OrderMap, action string,
failedJobs = append(failedJobs, map[string]interface{}{
"unique_job_id": jobinfo.Data["unique_job_id"].(int64),
"job_id": jobinfo.Data["job_id"].(int64),
"array_idx": int(jobinfo.Data["array_idx"].(int64)),
"array_idx": arrayIdx,
"msg": fmt.Sprintf("job_status not found for job"),
})
}
@ -3078,8 +3094,11 @@ func foundJob(jobid int64, array_idx int, jobinfos []utils.OrderMap) ([]utils.Or
jobs := make([]utils.OrderMap, 0)
for _, jobinfo := range jobinfos {
if array_idx > 0 {
if jobinfo.Data["job_id"].(int64) == jobid &&
int(jobinfo.Data["array_idx"].(int64)) == array_idx {
// bug-1742: array_idx is a plain Go `int` on models.JobInfo, not
// `int64` -- utils.IntOrInt64 accepts either.
jobinfoArrayIdx, ok := utils.IntOrInt64(jobinfo.Data["array_idx"])
if ok && jobinfo.Data["job_id"].(int64) == jobid &&
jobinfoArrayIdx == array_idx {
jobs = append(jobs, jobinfo)
}
} else {

73
src/frontend/apiserver/controllers/jobs_test.go
View file

@ -119,3 +119,76 @@ func TestPreCheckJobRequest_EmptyQueue(t *testing.T) {
t.Errorf("empty queue should not trigger reserved-name error: %v", err)
}
}
// TestValidateJobControl_HandlesRealJobInfoTypes is a regression guard for
// bug-1742 (gitea #45): models.JobInfo.JobStatus and .ArrayIdx are plain Go
// `int` (see the JobInfo struct in models/jobs.go), but validateJobControl
// asserted them as `int64` -- a mismatch invisible to any test that
// hand-built its utils.OrderMap with int64 literals instead of going through
// the same store.StructToOrderMap() reflection path production actually
// uses. That mismatch made every job-control call panic with "interface
// conversion: interface {} is int, not int64", surfaced to xcecli as an
// empty `failed: []` (Gin's Recovery() middleware swallows the panic).
func TestValidateJobControl_HandlesRealJobInfoTypes(t *testing.T) {
resetJobSubmitState()
jobInfo := models.JobInfo{
Unique_JobID: 73932643848159233,
JobID: 1788186846,
ArrayIdx: 1,
SubmitUser: "admin",
QueueName: "normal",
JobStatus: models.JOBSTAT["RUN"],
}
orderMap := store.StructToOrderMap(jobInfo)
successJobs, failedJobs := validateJobControl(orderMap, "kill", "")
if len(successJobs) != 1 || len(failedJobs) != 0 {
t.Fatalf("expected exactly one success job for a killable RUN job, got successJobs=%v failedJobs=%v", successJobs, failedJobs)
}
if got := successJobs[0]["array_idx"]; got != 1 {
t.Errorf("array_idx = %v (%T), want 1 (int)", got, got)
}
}
// TestValidateJobControl_RejectsNonKillableStatus verifies a job in a
// terminal state is reported as a failure (not silently dropped, and
// without panicking) -- this path also reads array_idx before returning.
func TestValidateJobControl_RejectsNonKillableStatus(t *testing.T) {
resetJobSubmitState()
jobInfo := models.JobInfo{
Unique_JobID: 1,
JobID: 1,
ArrayIdx: 0,
JobStatus: models.JOBSTAT["DONE"],
}
orderMap := store.StructToOrderMap(jobInfo)
successJobs, failedJobs := validateJobControl(orderMap, "kill", "")
if len(successJobs) != 0 || len(failedJobs) != 1 {
t.Fatalf("expected exactly one failed job for a DONE job, got successJobs=%v failedJobs=%v", successJobs, failedJobs)
}
}
// TestFoundJob_MatchesRealArrayIdxType is a regression guard for the same
// bug-1742 type mismatch inside foundJob's array-index lookup path.
func TestFoundJob_MatchesRealArrayIdxType(t *testing.T) {
resetJobSubmitState()
jobInfo := models.JobInfo{
JobID: 1788186846,
ArrayIdx: 1,
}
orderMap := store.StructToOrderMap(jobInfo)
found, err := foundJob(1788186846, 1, []utils.OrderMap{orderMap})
if err != nil {
t.Fatalf("unexpected error: %v", err)
}
if len(found) != 1 {
t.Fatalf("expected to find 1 job, got %d", len(found))
}
}