[RFC] Fixing the machine id
mzqohf at 0pointer.de
Wed Mar 9 10:36:51 PST 2011
On Wed, 09.03.11 13:25, Colin Walters (walters at verbum.org) wrote:
> > So, here's what systemd now does at boot: when it finds /etc/machine-id
> > populated, then everything is fine. If it is not populated, it tries to
> > initialize it from /var/lib/dbus/machine-id, with a fallback on
> > /dev/urandom.
> My one concern is that if systemd has a /dev/urandom fallback, and
> it's silently used in the case of OS misconfiguration, it could make
> consumers of it unreliable. Someone posted a DBus patch to get rid of
> the file and just autogenerate it each time, but one of the major
> points of this ID *is* that it's persistent; if your app or library
> needs to store machine-specific state, you really don't want it to be
> lost on reboot.
> Or is systemd going to create /etc/machine-id from /dev/urandom
> persistently? In that case, sounds good.
The rule is: if there's a way to make it persistant, we make it
persistant. But usually there isn't if we boot from CDROM, for
Lennart Poettering - Red Hat, Inc.
More information about the dbus