On 28.11.2014 23:14, Thomas Liske wrote:
> On 11/28/2014 11:04 PM, Sven Hartge wrote:
>> On 28.11.2014 22:57, Thomas Liske wrote:

>>>> So please add a blacklist entry to avoid restarting systemd-journald
>>>> under any circumstances.
>>
>>> We should wait until systemd-journald's RC bug got closed. Maybe we
>>> should add systemd-journald to the override_rc list if the bug is
>>> non-fixable.
>>
>> The bug is non-fixable, as far as I understand the IRC-log.
> 
> So systems will run outdated systemd-journald after (security) updates
> until the systems got rebootet? Since journald owns stdout/stderr of all
> the services there seems to be many (remote) attack vectors. This sounds
> awful to me.

It _is_ awful.

> If journald is restarted the stdout/stderr of running daemons got lost -
> is this a bigger problem than running a vulnerable journald?

The systemd-way of logging is to no longer log to syslog directly but to
just output to STDERR or STDOUT, as far as I understand it.

By restarting journald you risk losing _all_ logging output from all
daemons running the systemd-way.

At least that is the way I understand the situation.

Please ask a systemd maintainer for more input, I might be totally wrong.

Grüße,
Sven.


Attachment: signature.asc
Description: OpenPGP digital signature

Reply via email to