Response to Meltdown and Spectre

Polytropon freebsd at edvax.de
Mon Feb 5 16:32:34 UTC 2018


On Mon, 05 Feb 2018 16:19:43 +0000, Frank Leonhardt wrote:
> On 2018-02-05 15:57, Matthias Apitz wrote:
> > El día lunes, febrero 05, 2018 a las 02:37:20p. m. +0100, Polytropon 
> > escribió:
> > 
> >> > For all production server I run any reboot that is not scheduled by admins
> >> > is ultimate disaster, so it is equivalent to "bricked" machine. That
> >> > hardware can not be further used as production server, but "mere" fact or
> >> > reboot is ultimate disaster itself.
> >> 
> >> While this is not the established meaning of "bricked", it it
> >> definitely an understandable (!) interpretation of the term.
> > 
> > For me "bricked" in addition to "unusable" means: there is no software
> > way any more to change the fault because the hardware does not boot up
> > into a state where it could read(...) any attempts to change somethinb. 
> > The
> > device is now "dark" and does not communicate anymore by no means.
> > 
> > Only a hardware change (for example replace some chip) would help.
> 
> I'd go further - something is only bricked to me when I can't fix it 
> using a rework station. It's been done...
> 
> HOWEVER, Polytropon et al make a very good point - if the software 
> update means the device is no longer usable for its intended purpose 
> then it might as well have bricked it. The effect is the same.

That's exactly what I wanted to express: A device like a server
is a machine that is specified and expected (!) to work within
certain margins. Those can include processing power, I/O throughput,
in conclusion downtime, reliability, performance (power utilized
per $) and so on. If a software update changes this situation -
and not needlessly turns it into an unbootable piece of electronic
garbage -, the machine leaves those margins and can no longer be
trusted in a production environment. It's not bricked per se (does
still boot, does still run), but can be treated as if it was - it
needs to be removed from the installation and replaced by something
else. Basically, it's up to your defintion of "works as intended".
Especially in production server environments, downtime is a very
important thing.

NB: Downtime always means "unplanned downtime", because planned
downtime is "maintenance period". :-)



-- 
Polytropon
Magdeburg, Germany
Happy FreeBSD user since 4.0
Andra moi ennepe, Mousa, ...


More information about the freebsd-questions mailing list