Path: csiph.com!v102.xanadu-bbs.net!xanadu-bbs.net!feeder.erje.net!us.feeder.erje.net!newsfeed.fsmpi.rwth-aachen.de!newsfeed.straub-nv.de!news.swapon.de!fu-berlin.de!uni-berlin.de!individual.net!not-for-mail From: Tim Watts Newsgroups: comp.os.linux.misc Subject: Re: How do home users of linux protect themselves against ill-advised changes? Date: Sun, 04 Jan 2015 18:20:20 +0000 Lines: 53 Message-ID: <4sconb-kun.ln1@squidward.dionic.net> References: Mime-Version: 1.0 Content-Type: text/plain; charset=windows-1252; format=flowed Content-Transfer-Encoding: 7bit X-Trace: individual.net HhD0hm4DtCAf5qEPpNOjvw12XnHMWrBDmH14Fq6B4gMSXumXe/ X-Orig-Path: squidward.dionic.net!not-for-mail Cancel-Lock: sha1:wYZVkz90NB8ZiP0YpOQmHyjeRpQ= User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.3.0 In-Reply-To: Xref: csiph.com comp.os.linux.misc:13222 On 04/01/15 17:56, William Unruh wrote: > On 2015-01-04, Tim Watts wrote: >> On 04/01/15 13:33, The Natural Philosopher wrote: >>> On 04/01/15 12:16, Richard Kettlewell wrote: >>>> Jeremy Nicoll - news posts >>>> writes: >>>>> How do home users of linux protect themselves against ill-advised >>>>> changes? >>>>> >>>>> What I mean is, if they install patches or some package whose >>>>> configuration screws things up, and they then can't boot - what then? >>>> >>>> Restore from backup. >>>> >>> The question is pretty carzy to start with. >>> >>> Packages - unless they mess with the kernel/drivers - don't stop the >>> thing booting. >> >> Buggering up libc? > > But that would have come out in rudimentary testing by the distro. > Yes, YOU could bugger up libc by your own playing around with the > source. It is extremely doubtful the distro could. I did not say it wasn't doubtful. I said it was possible - in direct rebuttal to TNP's statement that only a broken kernel would break booting. That's it... And it's not just the distros - and upgrade of libc that is broken mid progress by a power outage, full disk etc could be also at risk of breaking the system. > >> >> Or as I had with ubuntu + ZFS, deciding the hostid was broken (it was >> null) and refusing to mount any ZFS volumes. > > Unless / was zfs that should still boot, and you could remount them when > the system came up (of course with systemd, that is not necessarily true > anymore, unless you put "nofail" as an option for every mounted > partition except / into /etc/fstab) Well - the system was unusable in a normal way as /homes/ was on ZFS. However, because I am not a complete moron, I made sure the OS itself was not on ZFS so recovery was reasonably straightforward once the error had been located. But again, a rebuttal to "only a kernel can break booting". Nope - a moderate number of serious failures in userland can too.