Automatic cron runs stop working: "No channels ready to be executed, skipping cron." on every run. [#1952624]

My cron jobs suddenly stopped working. I looked into the database log, and saw this on every cron report in the log:
"No channels ready to be executed, skipping cron."

I went to admin/config/system/cron and pushed the "Run cron" button and all of the cron tasks executed properly then and from then on.

There were no errors in the log. Why would this happen? Why would all of my cron jobs stop executing, and then why would manually initiating a cron run make them work again?

Comments

Comment #1

travis uribe commented 25 March 2013 at 21:39

At first glance, this may seem to be a duplicate of this issue:
http://drupal.org/node/1748106

But it is not. In that issue, pressing run on the Elysia Cron status page continued to produce "skipping cron" messages, and there was no report of cron tasks suddenly stopping working.

Comment #2

gotheric commented 27 March 2013 at 21:42

This is strange.

A badly crashed job could produce something like this (elysia cron - and the standard cron - things that the channel is running, because a job crashed and left and inconsistent enviroment), but this should be only a temporary state (there is an option to tell the "stuck time", and it should be some minutes if you have not changed anything in that).

If this happens again you could consider installing the 7.x version (it has several improvements and fixes, and should be stable enough...) and see if this solves the problem.

Comment #3

travis uribe commented 30 March 2013 at 19:26

I do have the 7.x version, but I'm assuming you meant the 2.x dev branch.

This same bug did happen again, last night around midnight, or so the staff tell me. But the logs for that time range were not visible this time. But I can tell from the timestamps of published nodes that the scheduled tasks did indeed stop running - or there was an error each time in attempting to publish them.
This time, the cron runs just started working again on their own. I don't know if one of the staff did anything to make it start again, and they're unlikely to know either, but if they did anything I'm certain they did not mess with cron.
I'll give the dev branch a try and report back. This is indeed a very strange issue. I can't for the life of me think what could cause this to happen.

Thanks.

Comment #4

vinmassaro commented 5 June 2013 at 15:59

What I found worked was to change the weight of the tasks I wanted to run to -50, -49, -48. These got moved to the top, and then they started running correctly from the UI or from Drush.