Job Site Outcomes¶
Each site in a job has its own progress and, once its part of the job ends, its own outcome. The Status column on the Sites tab of the job's details shows both. The outcome decides whether the site counts against the failure limit, and whether a rerun of the job covers it.
A Site's Status While It Runs¶
| Status | Meaning |
|---|---|
waiting to start |
The site has not begun its steps |
step N (type) |
The site is on step N |
step N (type), no answer from the site since ... |
The Hub is waiting for an answer about step N; see When a Site Does Not Answer |
waiting until ..., step N |
The site is at a Wait or a Wait until a time, which ends at the time shown, in your browser's time zone |
at the checkpoint, step N |
The site is waiting at a Checkpoint for someone to continue the job |
waiting for a free place after step N |
The site is ready to carry on after a checkpoint or a clock-time wait, and is waiting for an in-progress place |
finishing |
The site has run its last step and ends on the next pass |
How a Site's Part Ends¶
| Status | Counts as a failure | A rerun covers it |
|---|---|---|
finished |
No | No |
step N (type) failed: ..., step N (type) was refused: ... or step N (type) was not sent: ... |
Yes | No |
unknown: no answer from the site about step N |
Yes | No |
offline, not run, not ready, not run (...), not permitted, not run (...) or not run: step N never reached the site |
No | Yes |
stopped, not run or halted, not run |
No | Yes |
rejected, not run |
No | No: a rejected job cannot be rerun, only copied |
stopped after step N or halted after step N, with a reason when one applies |
No | No |
A site that failed keeps whatever its steps changed. A site the job's import was sent to can be rolled back once its part has ended, whatever its outcome; see Rolling Back a Job Site. A rerun makes a new draft for the sites that were not run; see Copying and Rerunning a Job.
Sites That Are Not Run¶
When the job reaches a site, before its first step, the Hub checks that the steps can run there. A site that fails a check is not run, and its status says why:
| Status | Why |
|---|---|
not ready, not run (the site is disabled or deleted at the Hub) |
The site is disabled at the Hub, or was deleted |
offline, not run |
The site is not connected to the Hub; the job does not wait for it |
not ready, not run (the site has not reported its Hub settings) |
The Hub cannot tell whether the site takes job steps, for example because it runs an older QIE version or has not reconnected since it was upgraded |
not ready, not run ('Use Hub' is off at this site (System Configuration, Remote Management Hub, Hub Sign-In tab)) |
Use Hub is off at the site (see Hub Sign-In) |
not ready, not run (This site's version does not run job steps) |
The site's QIE version does not take job steps |
not ready, not run (The Hub role '...' matches no 'Role Map' entry at this site (System Configuration, Remote Management Hub, Hub Sign-In tab)) |
No Role Map entry at the site matches the preparer's Hub role |
not ready, not run (The person has no Hub role, so no 'Role Map' entry at this site matches (System Configuration, Remote Management Hub, Hub Sign-In tab)) |
The preparer has no Hub role, or is no longer a Hub user |
not permitted, not run (the preparer does not hold the Jobs grant here) |
The preparer does not hold the Jobs grant at the site. When it is the approver who does not, the text says the approver does not hold the Jobs grant here |
not ready, not run (developer license) |
The site has a developer license, and the job has a start, resume or restart step |
The Hub judges a site's settings from what the site reported when its tunnel last connected, and a change at the site reaches the Hub at its next connect. The site still makes every check itself when a step arrives, and a step it refuses fails the site.
Before the job starts, the Warning column flags the sites the Hub already expects to skip; see Job Details.
When a Site Does Not Answer¶
The Hub sends a step again only when it never reached the site, or when the site has no record of it:
- A step that never reached the site is sent again with a new statement, after 10 seconds at first and less often the longer the site stays out of reach, down to once a minute. That is the case when the Hub could not reach the site at that moment, and when the site turned it away because the preparer was over the site's request limit. When a Hub node relaying to the site answers in its place, the step may or may not have reached the site, so the Hub asks the site about it as below.
- A step that reached the site without a complete answer coming back is followed by a question to the site, by the statement the step was sent with. When the site says the step finished, the Hub records its result as if it had answered. When it says the step is still running, the Hub asks again on later passes. When it has no record of the step, the Hub sends the step again with a new statement.
- A step whose work at the site was cut off, for example by a restart of the site, is reported by the site as failed and interrupted.
- A step the site accepted but could not start, for example because its upload was cut short, is recorded by the site as never started. So is one that still has not started 15 minutes after it arrived. Nothing ran, so the Hub sends the step again with a new statement, up to three sends in all. After the third, the site fails with
step N (...) did not start at the site after 3 sends.
While no answer comes, the site keeps its in-progress place, and its status shows since when it has not answered. Once every other site has ended, or is waiting at a checkpoint or a clock-time wait, a site that has gone five minutes without answering ends as unknown. A site that has gone 30 minutes without answering ends as unknown whatever the other sites are doing, so it gives up its in-progress place to a site waiting to begin. The Hub keeps asking it until then. An unknown site counts as a failure. A site the Hub never managed to send anything to ends instead as not run: step N never reached the site, at the same times, where step N is the first step meant for the site, after any waits and checkpoints before it: nothing ran there, so it is no failure and a rerun includes it.
When Someone Loses Access¶
A site where the preparer or the approver no longer holds the Jobs grant goes no further. While nothing has been sent to it, even after a wait or checkpoint at the start of the job, it is not permitted, not run (...); after that, it ends as stopped after step N: the preparer does not hold the Jobs grant here, or the same for the approver. Neither counts as a failure. See Running a Job.