-
Notifications
You must be signed in to change notification settings - Fork 443
JobStatus
Vitalii Koshura edited this page Apr 8, 2023
·
2 revisions
Jobs can fail in various ways. Here's a summary of what ends up in the server DB (in all cases, result.outcome is RESULT_OUTCOME_CLIENT_ERROR (3).
what happened | result.exit_status | <message> in result.stderr_out |
app crashed | status returned by OS | "process got signal N" or similar |
app called exit(N), N nonzero | N | "error string - exit code %d (0x%x) " or similar |
app aborted by client, too much time | ERR_RSC_LIMIT_EXCEEDED (-177) | "Maximum elapsed time exceeded" |
app aborted by client, too much disk | ERR_RSC_LIMIT_EXCEEDED (-177) | "Maximum disk usage exceeded" |
app aborted by client, too much RAM | ERR_RSC_LIMIT_EXCEEDED (-177) | "Maximum memory exceeded" |
abort requested by scheduler | ERR_ABORTED_BY_PROJECT (-221) | "aborted by project - no longer usable" |
abort if not started requested by scheduler (because past deadline) | ERR_ABORTED_BY_PROJECT (-221) | |
too many exit(0)s | ERR_TOO_MANY_EXITS (-226) | "too many exit(0)s" |
abort requested by user | ERR_ABORTED_VIA_GUI (-197) | |
input file download failed | ERR_RESULT_DOWNLOAD (-186) | "WU download error: %s" or similar |
output file upload failed | ERR_RESULT_UPLOAD (-187) | description of failed uploads |
client exiting, config.abort_jobs_on_exit | ERR_ABORTED_ON_EXIT | "aborting on client exit" |
scheduler acked active job | EXIT_ABORTED_BY_CLIENT (194) | "Got ack for job that's still active" |
app launch failed | ERR_RESULT_START (-185) | "couldn't start %s: %d" |