Andy Grimm of Red Hat reports:
OpenShift uses /etc/cron.daily/openshift-origin-cron-daily to run:
/usr/bin/oo-scheduled-jobs run daily &> /dev/null
This in turn runs all the user gears cron.daily content. If these cron jobs
take a long time to run it will prevent further OpenShift gears cron.daily
from being run in a timely manner if at all. The same goes for /etc/cron.weekly/openshift-origin-cron-weekly
This issue was discovered by Andy Grimm of Red Hat.
It looks like this should be set for
Product: OpenShift Online
and not security response. I could be wrong but I don't think the developers will see it in this state.
I see in brew:
Michal was the last person to make a change to that RPM.
I fixed LD_LIBRARY_PATH problem there that cause problem when users have SCLized python/ruby/whatever inside cronjob, that env var was not exported properly.
Kurt: There is a timeout inside the cron_runjob.sh script that is responsible for executing users scripts. This script have 'timeout' command in places as executor. See here:
(In reply to Tim Kramer from comment #5)
> It looks like this should be set for
> Product: OpenShift Online
> Component: Cartridge
> and not security response. I could be wrong but I don't think the
> developers will see it in this state.
This is the CVE bug, what you're describing is the tracking bug https://bugzilla.redhat.com/show_bug.cgi?id=1065045 where the changes can be made.
This was fixed publicly:
For what it's worth, this shipped as part of the OpenShift Enterprise 2.1 rebase.
This issue has been addressed in following products:
RHEL 6 Version of OpenShift Enterprise 2.1
Via RHBA-2014:0487 https://rhn.redhat.com/errata/RHBA-2014-0487.html