Description of problem:

The environment had some problems and as such the amount of logs that were produced were in the range of 1300 Gigs in a very short span of time in an hour so. 
The logrotate utility is configured by cron to run daily and keep only 140M logs but like here in case the logrotate didnt have a chance to run and the customer was in a fix before that

/var/log/neutron/*.log {
    rotate 14
    size 10M
    missingok
    compress
    copytruncate
}


Version-Release number of selected component (if applicable):

RHOS 7.0

How reproducible:

Always

Steps to Reproduce:
1.By having runaway logs on a RHOS deployment
2.
3.

Actual results:
The disk space filled out completely

Expected results:
Some defence mechanism to handle such things

Additional info:

The logs get compressed become very small like a 250 GB log got compressed to 4.4G, see below

656G    openvswitch-agent.log
4.4G    openvswitch-agent.log-20160121.gz ->~250GB
12G     openvswitch-agent.log-20160122.gz ->~684GB
635G    openvswitch-agent.log-20160123
3.2M    openvswitch-agent.log-20160124

but if the disk does not have space for compression to succeed, the problem arises with huge logs and no disk to write to them or do anything/


The logrotate utility is working fime and the cron is configured to run it daily and it is doing as expected