[CentOS] Systemd debug logging turned on in CentOS 7

[Thomas Eriksson]

On 02/28/2017 08:55 AM, Brian Mathis wrote:
> On Tue, Feb 28, 2017 at 10:49 AM, Valeri Galtsev <galtsev at kicp.uchicago.edu>
> wrote:
> 
>>
>> On Tue, February 28, 2017 9:22 am, Rob DeSanno wrote:
>>> Last time I saw it, I had just upgraded my CentOS 7 box with the
>>> 3.10.0-514 kernel and it rebooted already configured into debug mode.
>> Not
>>> sure if this is a “feature† of the newer kernels or not but glad to
>>> see that i’m not the only one who had noticed this.
>>>
>>> # awk -F\' '$1=="menuentry " {print i++ " : " $2}' /etc/grub2.cfg
>>> 0 : CentOS Linux (3.10.0-514.6.2.el7.x86_64) 7 (Core)
>>> 1 : CentOS Linux (3.10.0-514.6.2.el7.x86_64) 7 (Core) with debugging
>>> 2 : CentOS Linux (0-rescue-7b37bcbe36eb420fb6426976c41b0aaf) 7 (Core)
>>> 3 : CentOS Linux (0-rescue-7b37bcbe36eb420fb6426976c41b0aaf) 7 (Core)
>> with
>>> debugging
>>
>> I am not certain if there is real harm to have kernel with all debug stuff
>> running on production machines. Probably no harm security wise, the only
>> unpleasant stuff is: you really would prefer to run as slim kernel as
>> possible on production systems. If I'm wrong about "no harm", somebody
>> chime in, I then will be really eager to address it on my boxes.
>>
>> Valeri
>>
>>
> 
> Main issue I've seen is that logs grow by an order of magnitude larger than
> when it's off, due to systemd being systemd and now running in debug mode.
> Other than disk space, it would affect any central logging system you have
> with lots of unnecessary traffic, and would also add a lot of IO, amplified
> if you have many machines running on a VM host.
> 
> ~ Brian Mathis
> @orev


Just to put the record straight; it's not related to kernel debugging
being enabled or not. It's systemd debugging that is being turned on
for all menu entries, kernel debug or not.

Anyway, I think I have found a pattern. Only those machines that were
updated from 7.2 to 7.3 using the CR repo are showing this behaviour.
New 7.3 installs are fine. I'll just clean up the machines affected
and move on. There must have been some debug config left in an
installation script in one of the CR rpms.

thanks to everyone responding,

	Thomas