[Aaron@MIT]

Alright, everyone. I’m confused.

I have a Map/Reduce script that is scheduled to run on Saturdays and Mondays every week. In fact, here’s a screenshot of how I have it set up.

The actual script file contains logging. The very first line of every function is a log in order for me to track how and when it’s running. But my execution logs are empty. If I go by them, my script has never run. Similarly, the items that are being updated by this script are, well, not being updated. No changes to the fields whatsoever.

Even more confusing is that if I look at the Script Status page for this script, it says that it IS running. I have logs of it running on the 5th, 7th, 12th, and 14th of this month, which are correct given my schedule. Viewing details shows that it called Get Input Data, Map, Shuffle, Reduce, and Summarize, which is also odd since my script only has entry points for Get Input Data and Reduce. The script runs for 4 minutes, so I know it’s not just calling empty functions.

I verified that the script file that it’s calling is indeed the correct file (logging and all) and that there are no errors at all. If I trigger the script manually, it works great; when it’s scheduled, not at all. What is going on? Has anyone experienced this?

Thanks in advance.
This is a cached copy. Click here to see the original post.

[erictgrubaugh]

Is this a sandbox or demo account? The scheduling queues are not processed in these environments; all scripts must be run manually.

Beyond that, I am also very confused.

[chanarbon]

Hi Aaron,

Agreeing with Eric, could you confirm the environment that you have. You may want to add some checkers like why not write/create a file on map and reduce just to test? Also, I would suggest you to create a summarize function which will run through the errors on the mapSummary and reduceSummary

[Chuan Zhuo]

Hi Aaron,

In my mind, Since manually running/trigger Map/Reduce script works great, so I guess the issue part is not in script file itself.

for me, The Schedule tab under script deployment working in confusing mode, I.e.

1. It based on Server timezone.

2. Execution log will be clean up in next day.

3. Yield after Minutes, can be short to 3 min, so it doesn’t goes to another day…

Suggestion: when you doing long term schedule(weekly), do send email in your code, so it tracks execution better. (it might helps on catch detail task execution time)

If it performance differently in some weeks, not all, it must be NS server issue(for schedule task/queue…)

[Author]

Posts