Cron - Job was killed by signal #15

*johnfl68* · 02-07-2013, 03:52 PM

I just got a bunch of these messages from one of my scripts:

"Job was killed by signal #15"

I am not really finding anything in searching as to the meaning of this message.

Does anyone have any possible clues?

Thanks!

John

**Bob** · 02-07-2013, 04:02 PM

John,

What server was the site on? If the TAZ Server then one of the technicians killed some hanging processes to try and troubleshoot a potential issue.

-Bob

*johnfl68* · 02-07-2013, 04:10 PM

Bob:

Thanks. Yes, I am on TAZ.

I don't think it would have been anything of mine, as none of my cron jobs run for very long.

If you find anything suspect, please let me know.

John

**Bruce** · 02-07-2013, 05:23 PM

The problem wasn't strictly your scripts, it was more of a network problem. Network traffic to a few points outside of our network was taking a very long time to respond (minutes to complete connections).

From my reading, your cron job fetches a number of URLs in sequence. This in itself is not a problem, but it also lacks a timeout. This left it hanging waiting for network input.

As a result, about a dozen copies of your script were running at once, none of which were making any progress after running for as much as a half hour. Since they weren't making any progress, we killed them to prevent further problems.

*johnfl68* · 02-07-2013, 08:19 PM

Hmmm:

I will take a look at that to see if I can correct then as soon as I can. I know some elements of those scripts are already doing that (I get timeout warnings on occasion when uploading to Amazon S3, for example).

I will just need to figure out which elements do not already have timeout indications built in (and figure out a way to test).

Sorry.

John