[lkml]   [1999]   [Jul]   [21]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
    SubjectRe: FS corruption... some help maybe??

    On Wed, 21 Jul 1999, Herbert Wengatz 42850 wrote:

    > +> >
    > +> > Can you tell me, what use rebooting is?
    > +>
    > +> new hardware, sometimes (keyword is sometimes not all the
    > +> time) you may need to reboot for changes in inetd configurations
    > +> to take effect.. /etc/hosts file changes, I have had to.. what
    > +> is your system doing?
    > +> how do you run fsck? can you unmount the system drive and
    > +> run fsck on it? And not reboot? If that is the case then rather
    > +> than reboot I'd do that once a week ..
    > You are naming the exceptions. A system is OK to reboot for the
    > following occasions:
    > - changes in the hardware
    > If you don't have a true HA-solution, replacing a cpu-board is considered
    > severe enough. - Maybe even adding some ethernet cards, or harddisk.
    > But if you *have* a HA-solution, these are NOT considered to be a reason
    > for a reboot.
    > - basic changes in the systems internals (as changes in the /etc/hosts)
    > (You may even change this in singleusermode...)
    > Going into singleusermode is not considered a reboot, since the systems
    > uptime keeps growing...
    > I don't consider a upgrade of any piece of software that does not
    > belong directly to the kernel as being a system internal. I.E. an
    > upgrade of an office-package should be no reason at all to reboot
    > the system. - In fact, where I work, we replaced (updated) the
    > complete Applixware-package overnight on all our customers workstations.
    > - Hasslefree, without any complains and not a single reboot!
    > - after a severe crash: Filesystemschecks - But you only need to go into
    > singleusermode for that! (Same goes for chancges in the /etc/hosts)
    > When you leave these rare and necessary occasions away, what is left?
    > Exact the situations, where NT falls over it's own bootstraps.

    This is the problem. Win95/Win98/WinNT has to be rebooted a dozen times
    just to install a new software package or piece of hardware. And everyone
    gets used to the idea that changing anything (dial-up networking, etc)
    means a "you must now reboot windows for your changes to take effect."

    Unix/Linux isn't like that...


    > +> > At work (where I use Sun workstations) I have three machines
    > +> > (USER-Workstations!
    > +> > Where there are sitting users at the thing, logging in and out
    > +> > and crashing
    > +> > their applications all day long ;) ) - and these three
    > +> > machines have an
    > +> > uptime of far more than 280 days!!!!! (one is over 290 !)
    > +> >
    > +> does Suns filesystems need to be fsck'd?
    > Of course! - But only after a severe system crash. - Which I consider to be
    > an acceptable reason to make an fsck. ;-)
    > As long as you can manage to keep your system free of memmory leaks, zombies
    > and system crashes, there is no dire need for a fsck or reboot at all.
    > +> > Can anybody please show me *any* NT-User-WS with such uptime?
    > +> >
    > +>
    > +> okay now you are dreaming...
    > Why should I?
    > Why should I expect less from NT than from Solaris, when it's *THAT* great,
    > as M$ tries to suggest everybody and his uncle?
    > I'm sitting here with my real life experience and I'm still awaiting
    > M$ to come even a tiny bit closer to what I call a stable (which I call
    > *PROFESSIONAL*) system. - They don't get it. They haven't got it for years
    > now. - Even Linux was completely developed during the time M$ exists, so
    > M$ has a major advantage over Linux in this case. - Is NT more stable? - No.
    > Has it more features? - No. - Is it cheaper? - No. - Does it perform better?
    > - In most cases: No.
    > So why should I use NT?
    > Neutral analysts believe that NT contains 2.3 million bugs. *GULP*
    > And now comes Windows 2000 (aka NT 5.0) - guess how many bugs...
    > The analysts guess it will be around 5 million bugs.
    > I won't dare say Linux is flawless, but I consider it to have a vast amount
    > of bugs less than 5 millions...
    > Just my 0.02 $
    > Best regards,
    > Herbert
    > "Why aren't our NT-people in this meeting?" - "We are talking here about systems
    > with which we have troubles when they are running NOT." Two of my coworkers in a
    > meeting about a scheduled powerdown for crucial servers of a major german bank.
    > -
    > Linux SMP list: FIRST see FAQ at
    > To Unsubscribe: send "unsubscribe linux-smp" to

    To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
    the body of a message to
    Please read the FAQ at

     \ /
      Last update: 2005-03-22 13:53    [W:0.026 / U:78.908 seconds]
    ©2003-2017 Jasper Spaans. hosted at Digital OceanAdvertise on this site