Why does loadjob savedsearch return artifacts from a failed job?

[bhooker_axcient]

I have a dashboard which loads the results of a saved search to speed up the load times.

The saved search is scheduled to run frequently, and keeps results from the past 7 or 8 searches in it's history.

Often times, the dashboard gets 0 results back from the loadjob command, and when I check the latest jobs, the most recent job status will be failed.  To fix it, I just manually delete the failed job and let it roll back to the previous Done job.

Documentation for loadjob has an ignore_running option, but I am not seeing a way to ignore failed jobs, which would be nice.

** As an aside, I have noticed that I can make the saved search fail if I repeatedly call loadjob from a search (hit search 7 or 8 times without letting it finish) while the job is running.  I suspect that something like this is happening due to dashboard loads while this job is running which causes the failed job.

[bhooker_axcient]

On the dashboard, there were too many calls to loadjob with the savedsearch parameter.  I found that if a user loaded that page while the saved search had a job in progress, the job would fail.   

I corrected the problem by instead making the call once in the dashboard:

<search id="cached_results">
  <query>| loadjob savedsearch="my_job_name"</query>
  <done>
  <set token="tok_cached_result">$job.sid$</set>
  </done>
</search>

Then I can load those results each time I needed to reference them:

<search>
<query>| loadjob "$tok_cached_result$"</query>
...
</search>

I hope this helps someone!