Bug 555756

Summary:	sys-process/cronbase: run-crons: sub-jobs should be run with timeouts
Product:	Gentoo Linux	Reporter:	SpanKY <vapier>
Component:	Current packages	Assignee:	Gentoo's Team for Core System packages <base-system>
Status:	CONFIRMED ---
Severity:	enhancement	CC:	cron-bugs+disabled
Priority:	Normal
Version:	unspecified
Hardware:	All
OS:	All
Whiteboard:
Package list:		Runtime testing required:	---
Bug Depends on:	157547
Bug Blocks:

Description SpanKY gentoo-dev

2015-07-24 05:53:02 UTC

+++ This bug was initially created as a clone of Bug #157547 +++

run-crons has long had behavior where a long running job (perhaps by accident) will starve further execution until that job finishes.  for example, do:
  # printf '#!/bin/sh\nsleep 100000000' > /etc/cron.hourly/long-job
  # chmod a+rx /etc/cron.hourly/long-job

once that job is launched by run-crons, then no more jobs in cron.hourly will be launched until it exits.  this includes both the current `run-crons` as well as follow up launches of run-crons by cron.

we should update run-crons to execute each job with a timeout suitable for the category we're executing.  probably want to make it a command line flag so people can set the timeout higher if they really want this behavior.