Date: Mon, 5 Feb 2018 14:37:20 +0100 From: Polytropon <freebsd@edvax.de> To: galtsev@kicp.uchicago.edu Cc: freebsd-questions@freebsd.org Subject: Re: Response to Meltdown and Spectre Message-ID: <20180205143720.d4d98011.freebsd@edvax.de> In-Reply-To: <53029.108.68.160.114.1517707316.squirrel@cosmo.uchicago.edu> References: <CY1PR01MB12472D916F78A638731ECCE68FFB0@CY1PR01MB1247.prod.exchangelabs.com> <23154.11945.856955.523027@jerusalem.litteratus.org> <5A726B60.7040606@gmail.com> <92120E50-19A7-4A44-90DF-505243D77259@kreme.com> <CA%2BtpaK2o1nbY2W2JVRtogN=P2VM9rag_dodK=GtLWgKwNsYZkg@mail.gmail.com> <F395799E-2C94-47E9-AA1C-5CB075C50076@kreme.com> <044e62f7-69ca-71fe-34a8-5c5cafc06f08@yahoo.com> <slrnp7bpa3.2k8.naddy@lorvorc.mips.inka.de> <0520dd84-c00c-fbf2-da1c-f6ff4c63739d@yahoo.com> <20180203224612.GA10517@milliways.localdomain> <51178.108.68.160.114.1517699531.squirrel@cosmo.uchicago.edu> <slrnp7cmpn.5vo.naddy@lorvorc.mips.inka.de> <53029.108.68.160.114.1517707316.squirrel@cosmo.uchicago.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sat, 3 Feb 2018 19:21:56 -0600 (CST), Valeri Galtsev wrote:
>
> On Sat, February 3, 2018 7:00 pm, Christian Weisgerber wrote:
> > On 2018-02-03, "Valeri Galtsev" <galtsev@kicp.uchicago.edu> wrote:
> >
> >> With all due respect, one person saying, it didn't affect me, doesn't
> >> prove it is not disastrous for somebody else. Even if it is one machine
> >> out of thousand that is "bricked" for some time, it is a disaster for
> >> sysadmin who has that machine as a production server
> >
> > Of course, but who at all is saying that Intel's microcode updates
> > have "bricked" any machines? This appears to be an entirely spurious
> > claim, based on nothing other than grievous exaggeration that turns
> > "higher system reboots" into "bricked". You guys are talking each
> > other into a frenzy of fear over nothing.
>
> 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.
Allow me to explain:
After a device gets bricked, for example by a firmware update
gone bad (bricked by software), or by a power surge coming from
a malfunctioning power supply (bricked by hardware), it cannot
be used for anything anymore. It's like a brick. Usually, it
will be disposed of.
In this specific case, I can fully understand that if an update
(here: CPU microcode) renders a server-class (!) device unpredicatable,
this is often the same as unusable (!) for use as a server. Many
sysadmins will agree that if you cannot rely on a system, it's
not a good fit for a production machine. In this specific case,
it doesn't actually matter if the problem was in there from the
beginning, or added later by an upgrade.
This problem doesn't just occur in CPU or system board firmware
upgrades, but can also affect hard disk firmware or whole appliances.
Basically, any unscheduled downtime (for whatever "accidental"
maintenance reason) is bad in general, and definitely not a sign
of quality for server hardware.
On the other hand, I know many businesses where rebooting servers
is quite common ("Windows"-based installations, of course), which
if often due to software problems, wrongly configured hardware,
broken hardware, or missing technical skills of the "professional
consultants" and "solution experts" - that's what you find out in
the many post mortems. In such settings, rebooting and then accepting
a decrease in performance simply doesn't matter. The costs will be
relayed to the customers, of course, and the final consumer will
pay the costs added on all levels of the trade chain. This is a
typical thing here in Germany - nobody cares.
> I bet I am not the only sysadmin with this point of view. That is why I
> said production server is by no means comparable to sobody's home
> workstation. But I bet many people have the same attitude to their home
> workstations as I have to servers I am responsible for. Well, I for one
> have the same attitude to all my computers, including laptops, and as a
> sysadmin - to all computers of people I support.
You are definitely not alone with that mindset. You probably also
get the gears in your brain working for the next important (and
logical) question: What to buy instead? Which manufacturer can I
trust? Will more money prevent such problems in the future, or will
it just be money spent for marketing BS?
Actually, most users who have experience in professional settings
will want a system that is reliable and secure, for a reasonable
price. Home users, on the other hand, do not care, and they will
buy whatever is on the shelf, in the shiny box, with the "SALE"
tag above it. They have "learned" that permanently rebooting
computers is normal, and usage speed doesn't increase, even though
computing power increases. In our throw-away society, replacing
stuff (with similar stuff to do the same things as before) is
socially accepted and a requirement for technical evolution, and
the "free market" needs that attitude, because without growth there
will be no growth, and no growth is bad for everyone...
> Of course, some people may have different point of view on this, which
> does not deprive me of expecting from my hardware to work without
> "glitches".
This specific mentality is present among those who are not personally
responsible for systems and their reliability (i. e., home consumers
and typical office drones who are paid for being present and visible
in the office, not for getting work done, and surely not for using
their brains).
> My apologies for being edgy. These things make me such.
Working in IT is not fun anymore... ;-)
--
Polytropon
Magdeburg, Germany
Happy FreeBSD user since 4.0
Andra moi ennepe, Mousa, ...
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20180205143720.d4d98011.freebsd>