Linux-Misc Digest #494, Volume #18 Wed, 6 Jan 99 18:13:08 EST
Contents:
Re: Anti-Linux FUD (Floyd Davidson)
Re: Does Linux support plug & play ("LORD Ni'Kon")
Re: Netscape eats up *all* the swap (Floyd Davidson)
Re: Linux OS NOT preemptive multitasking ? (Stephen E. Halpin)
Re: Linux OS NOT preemptive multitasking ? (Stephen E. Halpin)
Re: Stupid Wordperfect Question (garv)
Re: Anti-Linux FUD (Floyd Davidson)
Re: help me choose license ([EMAIL PROTECTED])
Re: Anti-Linux FUD (David O'Bedlam)
Re: Printing only odd/even pages of postscript (Michael Powe)
Re: Partitions (Michael Powe)
html editor (Indiana)
Re: Anti-Linux FUD (Matthew Malthouse)
Re: Linux: Fight for survival or on victory march? (jedi)
Re: Is Microsoft a nasty company ? I'm asking you this question. (Randal)
Re: high uptime server designs? (Bob Deep)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (Floyd Davidson)
Crossposted-To: uk.comp.os.linux
Subject: Re: Anti-Linux FUD
Date: 6 Jan 1999 12:45:05 GMT
Reply-To: [EMAIL PROTECTED]
David Damerell <[EMAIL PROTECTED]> wrote:
>Floyd Davidson <[EMAIL PROTECTED]> wrote:
>>David Damerell <[EMAIL PROTECTED]> wrote:
>>>Floyd Davidson <[EMAIL PROTECTED]> wrote:
>>>>David Damerell <[EMAIL PROTECTED]> wrote:
>>>>>Floyd Davidson <[EMAIL PROTECTED]> wrote:
>>>>>>Nix <$}xin{$@esperi.demon.co.uk> wrote:
>>>>>>>Oh, and steering clear of PATHization also lets you install things
>>>>>>>without requiring everyone to log out and log in again (ugh, shades of
>>>>>>>Windows).
"install things without requiring everyone to log out and log in again."
>>>>>>So why would anyone log out to make it effective? That is
>>>>>>absolutely not necessary.
"That is absolutely not necessary"
The rest of this discussion is nothing but obfuscation. It simply
is not necessary. It *could* be done, but it 1) does not have to
be, 2) is not desirable to. Hence the response above is correct.
>>>>>Consider Joe Luser on a large multiuser system. He doesn't know what
>>>>>you're doing; he isn't going to get the new system PATH until his next
>>>>>login. But he will get immediate access to symlinked binaries.
>>>>What makes that bad?
>>>Look at the question I am answering. I am answering the question as to
>>>whether a logout is necessary. I didn't even start to say this was good or
>>>bad, mauve or orange; and talking about that is a pretty lousy distraction
>>>from having been wrong when you said 'So why would anyone log out to make
>>>it effective? That is absolutely not necessary.'
>>That is *still* absolutely not necessary. Your suggestion is not only
>>not prefered, it isn't desirable at all.
>
>Perhaps you missed the bit where I said that I am not discussing whether
>this is good or bad (or preferred or desirable.) Let me summarise;
>
>Nix claims that PATHization forces people to log out and in again.
>You disagree.
>I point out a situation in which it might be necessary, and would not be
>necessary with symlinked binaries.
You did not describe such a situation. It is NOT necessary.
>Please do not introduce this good/bad diversion; confine yourself to the
>points being discussed. In what sense would our naive user not need to log
Then please confine yourself to the points being discussed. Do
you really mean to suggest that you are the sole arbitrator of
which points any usenet discussion, much less this one, is
about? Get real.
>out and in again to access the new binaries (in the fashion he expects -
>without absolute paths) in the situation where the system PATH is amended?
>
>I must emphasise _again_ that I am not discussing which of these
>approaches is good or bad.
And again, I'll point out that a bad solution is a non solution
when good solutions are available. Your example is a
non-solution with a contrived requirement that does not exist
in reality.
Floyd
--
Floyd L. Davidson [EMAIL PROTECTED]
Ukpeagvik (Barrow, Alaska) [EMAIL PROTECTED]
Pictures of the North Slope at <http://www.ptialaska.net/~floyd>
------------------------------
From: "LORD Ni'Kon" <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux,alt.uu.comp.os.linux.questions
Subject: Re: Does Linux support plug & play
Date: Wed, 6 Jan 1999 02:19:15 -0600
You have a problem, linux cant use comm5 with its default setings. You need
to look for a kernel that can.
Jerry Lynn Kreps wrote in message <[EMAIL PROTECTED]>...
>Jerry Lynn Kreps wrote:
>>
>> Thomas F. Ewald wrote:
>> > Luc:
>> > Sounds interesting and useful. My PnP modem is on COM5 (via WIN95) and
>> > Linux can't find it. Could you
>> > run through a setp-by-step on how to use these utilities, please?
Thanks.
>> >
>> > Tom
>>
>> I had that problem with my Sony VAIO. Redirected it to COM4 and got 8
>> bit mono on some programs, but not all. Could be that my modem was a
>> winmodem and required code from the M$ OS to operate in the 16 bit
>> mode. Anyway,I download the OSS sound software from 4front and
>> installed it. Works perfectly now.
>
>I forgot to say that besides COM5 my modem also used COM10, which
>USUALLY marks it as a Winmodem.
------------------------------
From: [EMAIL PROTECTED] (Floyd Davidson)
Subject: Re: Netscape eats up *all* the swap
Date: 6 Jan 1999 06:22:29 GMT
Reply-To: [EMAIL PROTECTED]
Alexander Viro <[EMAIL PROTECTED]> wrote:
>Floyd Davidson <[EMAIL PROTECTED]> wrote:
>[snip]
>>>I've found all versions of Netscape to be pigs. :(
>>
>>A total of 46Mb of virtual memory is far too little. Doubling
>>both the RAM an the swap would be a good idea. If nothing else,
>>at least up the swap to provide something more than 70-80Mb of
>>virtual memory.
>
>Hmm... If that doesn't shout 'pig'... BTW, what are reasons to
I am hard pressed to see where that shouts 'pig' at all. It
shouts "versatility" for one thing. Ability, for another.
I would NOT care to return to a "computer" that could only
address 52K of program space. I spent a bazillon hours doing
8080/8085/Z80 assembler coding because that was the only way
to fit a program into the addressable memory. I prefer C
and Linux by *far*. (I remember when X was first announced,
and I basically thought, so what! I'll *never* be able to
afford the hardware that requires. I learned a lesson...
>use it? I'm really curious and it's not a trolling. For what
>I've seen the thing is memory-,CPU- and IO-thirsty, makes a
Up to this point you are showing bias, or ignorance. I can't
tell which.
>poorest newsreader (weaker than telnet new.foo.org nntp and
>about equally efficient), allows HTML in email and news,
>is buggy as hell (wake me up when lynx will die with SIGBUS)...
However, you have hit several nails *directly* on the head
with that one! I simply do not understand why anyone uses
a browser to read usenet or email either one. They all
tend to emulate the msdos way of doing things for those
features: the stupid way!
>It's, erm, editor is dumber than even pico. It looks like
>a fundamentally non-UNIXish program - monolitic, heavy-weight,
>inflexible... What makes people use the sucker?
To look at URLs! To browse the net. It works like a charm!
(BTW, I didn't even realize there was a *name* for that
excuse for an editor! :-)
But since I use XEmacs for editing and reading email, and
trn for reading Usenet, I really don't think the comments
about IO, memory, etc. are valid.
Floyd
--
Floyd L. Davidson [EMAIL PROTECTED]
Ukpeagvik (Barrow, Alaska) [EMAIL PROTECTED]
Pictures of the North Slope at <http://www.ptialaska.net/~floyd>
------------------------------
From: [EMAIL PROTECTED] (Stephen E. Halpin)
Crossposted-To: alt.folklore.computers
Subject: Re: Linux OS NOT preemptive multitasking ?
Date: Wed, 06 Jan 1999 08:33:28 GMT
On 5 Jan 1999 21:56:17 GMT, [EMAIL PROTECTED] (Douglas W. Jones,201H
MLH,3193350740,3193382879) wrote:
>On Tue, 5 Jan 1999 06:05:00, [EMAIL PROTECTED]
>(Matthew Kirkcaldie) wrote:
>
>> >>hrm... poor phrasing on my part. RT OS == a superset of a strictly
>> >>multitasking OS. it fills the same specs as a multitasking OS, and then
>> >>some.
>>
>> Then it's a subset, if it's more specific, unless I misunderstood. All
>> RTOSs are MT OSs; there are some MT OSs which aren't RT.
>
>Take a look at the Amulet kernel, http://www.cs.uiowa.edu/~jones/amulet/
>
>This kernel is distinctly a real-time kernel, and it is absolutely not
>preemptive nor is it based on a tasking model! I prefer preemptive
>multitasking myself, but the point is, there are other models.
>
>The Amulet model is based on Charm, a distributed processing model where
>thereads are very short lived, and where each thread claims the entire
>resources of its CPU until that thread terminates. The short lifetime of
>threads allows all scheduling decisions to be made between threads, so
>there is never any reason to save or restore the state of a thread.
>
>Amulet extends this idea with real-time primitives. When you send a
>message to a method of an object, thereby firing up a short-lived thread
>to handle that method, you have the option of specifying when the message
>should be sent, and when you do this, you specify a real-time window,
>giving the earliest acceptable delivery time and the deadline, so that
>soonest-deadline first scheduling may be used.
Real-time means a guaranteed response time to specific events.
In your system, the 'soonest-deadline' task may be of the least
importance (lowest priority,) causing you to starve higher
priority tasks, demonstrating that you cannot provide real-time
response at the kernel level. This does not mean that you cannot
build a real-time >system< on top of this kernel by carefully
characterizing user level code, but it does mean that the kernel
is not a real-time kernel.
>The result is not purely object oriented -- methods never return results,
>so if you want a result returned to the object from which a method was
>called, you've got to include, as a parameter to the method you call, the
>identity of a method of your object to be called with the result. Sure,
>this is clumsy, but it's not new. It's called programming with explicit
>continuations by the folks in programming language theory, and it is a
>feature (or misfeature) inherited from Charm.
These basic concepts have been appearing throughout event driven
systems for decades - Charm appears to have been started in 1987.
Still, its a great way to build event driven systems.
> Doug Jones
> [EMAIL PROTECTED]
-Steve
------------------------------
From: [EMAIL PROTECTED] (Stephen E. Halpin)
Subject: Re: Linux OS NOT preemptive multitasking ?
Date: Wed, 06 Jan 1999 07:15:11 GMT
On 04 Jan 1999 09:42:37 -0500, Frank Sweetser <[EMAIL PROTECTED]> wrote:
>[EMAIL PROTECTED] (Floyd Davidson) writes:
>
>> A "real-time" OS. Which might not really be what anyone wants Linux
>> to be, but they do have their uses... telephone switching systems
>> for example.
>
>well, then you've got a real time OS, not a multitasking OS. and even
>then, the RT kernel won't be pre-empted by a task, it simply makes
>guarantees about scheduling characteristics that a plain multitasking OS
>can't make.
This is false. Real-time kernels are reentrant, running various
services at different priority levels. It is not only possible
but indeed a requirement that a real-time user level task be able
to preempt portions of the kernel. The network stack is perhaps
the most simple example, as a real-time kernel has no control
over the activity of a network interface, and therefore could
give no >reasonable< delay characteristics for user level tasks
if the kernel could preempt those tasks to deal with any amount
of network traffic. There are examples of timesharing OSs which
also provide for non-kernel tasks to preempt the kernel.
>--
>Frank Sweetser rasmusin at wpi.edu fsweetser at blee.net | PGP key available
>paramount.ind.wpi.edu RedHat 5.2 kernel 2.2.0pre3 i586 | at public servers
>Let us be charitable, and call it a misleading feature :-)
> -- Larry Wall in <[EMAIL PROTECTED]>
-Steve
------------------------------
From: garv <[EMAIL PROTECTED]>
Subject: Re: Stupid Wordperfect Question
Date: Wed, 06 Jan 1999 21:28:00 GMT
gfonseca wrote:
> This is a really stupid question...
>
> After installing Wordperfect 8 onto /usr/progs... I can't get it to run!
>
Did you change to the wpbin directory, then type ./xwp ? (dot important).
on mine cd /opt/wp70/wpbin $ ./xwp
------------------------------
From: [EMAIL PROTECTED] (Floyd Davidson)
Crossposted-To: uk.comp.os.linux
Subject: Re: Anti-Linux FUD
Date: 6 Jan 1999 12:36:02 GMT
Reply-To: [EMAIL PROTECTED]
David Damerell <[EMAIL PROTECTED]> wrote:
>Floyd Davidson <[EMAIL PROTECTED]> wrote:
>>David Damerell <[EMAIL PROTECTED]> wrote:
>>>Floyd Davidson <[EMAIL PROTECTED]> wrote:
>>>>Having sbin directories in a normal users path is not required in
>>>>order for a normal user to know the system's IP address.
>>>What's a better way to do it than 'ifconfig', then?
>>There are at least two ways. The most obvious is to run
>>ifconfig with a full path name.
>
>Yes, yes, you think anyone doesn't know that.
It appears to be the case. Why else would anyone want to put
sbin directories into every normal user's path, and then sit
here and adamantly claim there is no other way to use the progs
in those directories.
>>The most obvious way for normal users to find *any* ip address,
>>is nslookup.
>
>Rubbish. nslookup does something completely different, as you ought to
>know.
You need to re-read the stated purpose above. It does NOT ask
how to find the IP address of an interface. It does ask how to
find the IP address of the "system". That is indeed what
nslookup is for.
>>>mkdosfs is in an sbin directory.
>>But that 1) does not mean that sbin directories should be in
>>a user's path, nor does it mean mkdosfs is needed at all!
>>A variety of tools are available, and it varies from one
>>system to another. I use /usr/bin/fdformat, ymmv.
>
>MM suggests that formatting a formatted floppy in order to make a file
>system on it is pointless.
Then use any of a number of other tools. There are several, and
some of them are not in sbin directories. Of course you clearly
must by now, finally, understand that a full path name will get
a normal user access to the sbin directory... ;-)
>>>>>[By this, I refer to someone who obviously could have root if they wanted
>>>>>it, but is a housemate or family member or suchlike who doesn't want or
>>>>>need it; such a person may have a user account with privileges which
>>>>>equate to being root.]
>>>>But that is NOT a "normal" user
>>>On a home system - where an increasing proportion of Linux users are -
>>>that is a normal user. _Anyone_ with physical access to the box can give
>>No that is NOT a "normal" user.
>
>Repeating yourself does no good. Given that it is necessary to trust
>anyone who lives with the box, in what sense is such a person not a normal
>user?
"Normal" user logins do not do systems admin tasks.
Repeating myself seems to be necessary, since sometimes the same
question requires answering more than once, and the answer has
not magically changed since the last time. The right answer is
still the right answer.
Floyd
--
Floyd L. Davidson [EMAIL PROTECTED]
Ukpeagvik (Barrow, Alaska) [EMAIL PROTECTED]
Pictures of the North Slope at <http://www.ptialaska.net/~floyd>
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: help me choose license
Date: Wed, 6 Jan 1999 21:35:00 GMT
I wrote:
> There are arguments both ways on this, but I don't think they'd risk it.
> Even if they won in court they would take a huge PR hit.
David Steuber writes:
> Well, a company beginning with M lost a court to with Stac
> Electronics for stealing disk compression code. They are even bigger
> now. Where is Stac?
I have no idea, nor is it relevant. That was two corporations fighting
over legal arcana. A company that deliberately violated the GPL would be
shunned by the Linux community (I was asssuming we were talking about Linux
products). With free software as fashionably as it now is, they would
probably take a hit in the proprietary market that would wipe out any
advantages. Do you want to be the spokesperson who has to explain to the
press why your company is stealing software written by individuals in their
spare time and given to the community? And then there are the legal
expenses. They would be fighting teams of pro bono attorneys all the way
to the Supreme Court.
--
John Hasler
[EMAIL PROTECTED] (John Hasler)
Dancing Horse Hill
Elmwood, WI
------------------------------
From: David O'Bedlam <[EMAIL PROTECTED]>
Subject: Re: Anti-Linux FUD
Crossposted-To: alt.destroy.microsoft,comp.os.linux.advocacy,comp.os.os2.advocacy
Date: Wed, 06 Jan 1999 09:06:45 GMT
Uh, people, WITF is a "FUD"? I must've been sick that day.
TheDavid
--
"no statement, no matter how innocent, can be guaranteed to
be uncontroversial." --Talysman <[EMAIL PROTECTED]>
------------------------------
From: Michael Powe <[EMAIL PROTECTED]>
Subject: Re: Printing only odd/even pages of postscript
Date: 06 Jan 1999 13:49:16 -0800
=====BEGIN PGP SIGNED MESSAGE=====
Hash: SHA1
>>>>> "Bruno" == Bruno Barberi Gnecco <[EMAIL PROTECTED]> writes:
Bruno> Hello ALL! Does anyone know how to print odd and even
Bruno> pages separatedly of a postscript file? If possible,
Bruno> someway to use ghostscript, because I don't have a
Bruno> PS-compatible printer.
Well, you can use the psutils. Psselect cuts up ps files according to
your needs. So, `psselect -r -o file.ps file.new.ps' takes out all
the odd-numbered pages in file.ps and puts them in reverse order in
file.new.ps. `psselect -e file.ps file.even.ps' takes out all the
even-numbered pages the same way. Print the first one, put the pages
back in the paper tray, print the second one -- et voila, two-sided
printing.
mp
8<---------------how-easy-is-it-to-demunge-an-address?------------------->8
#! /usr/bin/perl # if you are [EMAIL PROTECTED] (Another Luser):
while ($line = <>){ if ($line =~ m/^\s*$/ ){ last; }
if ($line =~ m/^From: (\S+) \(([^()]*)\)/){ $from_address = $1; } }
if ($from_address =~ m/\S+NOSPAM\S+/){ $x = index($from_address, NOSPAM);
substr($from_address, $x, 6+1) = ""; printf("The real address is %s\n",
$from_address);}else { printf("No munge, just plain %s\n",$from_address);}
printf("\nBrought to you by the Truth In Mail Headers Foundation\n");
8<-----------------------here's-one-example------------------------------>8
- --
Michael Powe
[EMAIL PROTECTED] http://www.trollope.org
Portland, Oregon USA
=====BEGIN PGP SIGNATURE=====
Version: GnuPG v0.9.0 (GNU/Linux)
Comment: Encrypted with Mailcrypt 3.5.1 and GNU Privacy Guard
iD8DBQE2k9oB755rgEMD+T8RAjW9AJ40pkda1osxBZdXnu+pzl/GxAP+JgCePrgh
vcqPf0xaADx1YCdB6FErT7s=
=ZSL9
=====END PGP SIGNATURE=====
------------------------------
From: Michael Powe <[EMAIL PROTECTED]>
Subject: Re: Partitions
Date: 06 Jan 1999 13:39:40 -0800
=====BEGIN PGP SIGNED MESSAGE=====
Hash: SHA1
[posted and mailed]
>>>>> "Tom" == Tom <[EMAIL PROTECTED]> writes:
Tom> I've got a 512Meg HD. Can anyone suggest what partions I'll
Tom> need and of what size in order to install Linux? I was
Tom> planning on 40Megs for my Root, 300Megs for Usr and the rest
Tom> for X windows. Also when I install Linux, how does the
Tom> installation program choose the partion to install to? Are
Tom> the names I give them relevant?
Here's my setup: obviously, more than 500Mb but you can see the
proportions.
Filesystem 1k-blocks Used Available Use% Mounted on
/dev/hdc1 101075 42627 53229 44% /
/dev/hdc2 101093 64807 31065 68% /home
/dev/hdc3 1041638 700604 287217 71% /usr
/dev/hdb3 495746 182304 287839 39% /usr/local
/dev/hdb6 201011 92251 98380 48% /usr/src
Most of your space is going to be used up in /usr and /usr/local, so
that's where you want the bulk of your space. I probably made root
bigger than it needed to be, 50M is probably sufficient -- but if you
make it that small, you probably should add a /tmp partition, also.
Some installations will require that you install a separate swap
partition:
233 $ --> cat /proc/swaps
Filename Type Size Used Priority
/dev/hdb5 partition 18108 1268 -1
Others will let you just use a swap file. Either way, you should have
some swap space available.
mp
8<---------------how-easy-is-it-to-demunge-an-address?------------------->8
#! /usr/bin/perl # if you are [EMAIL PROTECTED] (Another Luser):
while ($line = <>){ if ($line =~ m/^\s*$/ ){ last; }
if ($line =~ m/^From: (\S+) \(([^()]*)\)/){ $from_address = $1; } }
if ($from_address =~ m/\S+NOSPAM\S+/){ $x = index($from_address, NOSPAM);
substr($from_address, $x, 6+1) = ""; printf("The real address is %s\n",
$from_address);}else { printf("No munge, just plain %s\n",$from_address);}
printf("\nBrought to you by the Truth In Mail Headers Foundation\n");
8<-----------------------here's-one-example------------------------------>8
- --
Michael Powe
[EMAIL PROTECTED] http://www.trollope.org
Portland, Oregon USA
=====BEGIN PGP SIGNATURE=====
Version: GnuPG v0.9.0 (GNU/Linux)
Comment: Encrypted with Mailcrypt 3.5.1 and GNU Privacy Guard
iD8DBQE2k9fi755rgEMD+T8RAi3bAJ9VeVRSM/TakV3Zt0tNwt7xsQK/BgCfYWqw
k2BmPAQNhI0EShgXQMqKPfk=
=qvqg
=====END PGP SIGNATURE=====
------------------------------
From: Indiana <[EMAIL PROTECTED]>
Subject: html editor
Date: Wed, 06 Jan 1999 12:11:47 -0500
Does anyone know a good html editor for XWindow and can i run webmaker
even if i have enlightenment not kde.
Thanks, Alex
--
====================================================
- Alexandre Arsenault [EMAIL PROTECTED]
- Runnin on Linux Visitez mon site: http://pages.infinit.net/linus99
------------------------------
From: [EMAIL PROTECTED] (Matthew Malthouse)
Crossposted-To: alt.destroy.microsoft,comp.os.linux.advocacy,comp.os.os2.advocacy
Subject: Re: Anti-Linux FUD
Date: 6 Jan 1999 07:14:01 GMT
Reply-To: [EMAIL PROTECTED]
On 27 Dec 1998 22:15:15 GMT John Edstrom wrote:
}
} Judge it by another set of criteria. Suppose, just hypothetically, of
} course, that you wanted to take control of all of the computers in the
} world. Learning what software was on 90% of the desktops in the world
} would be a big help in that direction.
}
} One way to do that would be to first concentrate complete and intimate
} knowledge of, and control over, all software installed on each
} computer into one easy to carry package. Something very much like the
} registry, say.
Great conspiracy theories of the world #107933.
The really worrying thing is that it's the most credible explanation
Ive yet heard for the existance of the registry. :)
Matthew
--
And on the seventh day, He exited from append mode.
http://www.calmeilles.demon.co.uk
------------------------------
From: [EMAIL PROTECTED] (jedi)
Crossposted-To: alt.linux,alt.os.linux,comp.os.linux.advocacy
Subject: Re: Linux: Fight for survival or on victory march?
Date: Wed, 6 Jan 1999 01:07:16 -0800
On 06 Jan 1999 00:07:47 -0500, Frank Sweetser <[EMAIL PROTECTED]> wrote:
>[EMAIL PROTECTED] (Bill Unruh) writes:
>
>> In <[EMAIL PROTECTED]> Frank Sweetser <[EMAIL PROTECTED]> writes:
>>
>> >> ]Linux's survival hangs on a DOJ thread. It is an interesting article and one
>> >> Where? What DOJ throead? What has the DOJ to do with Linux?
>>
>> >nothing directly. however, microsoft does indeed have an impact on linux,
>> >and the DOJ may potentially have a great impact upon microsoft, and thereby
>> >indirectly upon linux.
>>
>> And this is "hanging by a DOJ thread"? I assume it is also hanging by a
>> Penguin thread, since one of the Antarctic penguins might fart and open
>> up the Ozone hole and fry all the plantlife on earth, and Linus would
>> starve and then where would Linux be?
>
>please keep the attributions straight. my point of view is that if
>microsoft wins vs the DOJ, it may hamper the commercial success of linux,
No one has yet begun to support Linux that isn't already
marginalized or just plain balsy (like id). When the cowards
(or rats as one man put it) start crawling into the aqueduct,
THEN we can start talking about how they might run back out
again...
>but that the biggest damage it will do the core linux (and various other
>free software packages) is force everyone to deal with all of the flamewars
>and threads about how linux/gnu/*BSD is going to die now that microsoft
>won...
>
>--
>Frank Sweetser rasmusin at wpi.edu fsweetser at blee.net | PGP key available
>paramount.ind.wpi.edu RedHat 5.2 kernel 2.2.0pre3 i586 | at public servers
>echo $package has manual pages available in source form.
>echo "However, you don't have nroff, so they're probably useless to you."
> -- Larry Wall in Configure from the perl distribution
--
Herding Humans ~ Herding Cats
Neither will do a thing unless they really want to, or |||
is coerced to the point where it will scratch your eyes out / | \
as soon as your grip slips.
In search of sane PPP docs? Try http://penguin.lvcm.com
------------------------------
From: Randal <[EMAIL PROTECTED]>
Subject: Re: Is Microsoft a nasty company ? I'm asking you this question.
Crossposted-To:
alt.destroy.microsoft,comp.os.linux.advocacy,comp.os.os2.advocacy,comp.os.linux.x,gnu.misc.discuss,uk.comp.os.linux
Date: Wed, 6 Jan 1999 14:28:36 -0800
>>This question is important not to me but to most people who dedicate
themselves to fields unrelated to computers. The answer to this question is part of
the reason
that windows is so @#$%! popular...<<
Unless Windows crashes on you repeatedly. I work with a lot of people that
don't care about computer much one way or the other, but must use them. the one
complaint is not about functionality or ease of use of software, it's ease of
use of the system as a whole and frustration level. Most people want a computer
to be an appliance. You sit down, run the program you want and it works every
time. Just like the toaster. If the system is crash prone, most
(non-computer) folks will give it up unless they have no choice. I hear regular
ranting and raving about Windows, not for the programs run, or learning curve,
but for the number of crashes and the lost work. Yes, it's often from doing
something foolish (today the secretary deleted a file she had open in a word
processor from outside the word processor---good night). but it still causes
enormous frustration.
Randy
------------------------------
From: Bob Deep <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.setup,linux.redhat.misc
Subject: Re: high uptime server designs?
Date: Wed, 06 Jan 1999 16:34:43 -0600
Leslie Mikesell wrote:
>
> I am trying to set up a machine with the goal of having a minimum of
> downtime and/or data loss for any reason. Actually this will be
> a pair of machines where the one in actual production has drives
> periodically rsync'd to a second machine that is set up so it
> can do an alternate boot to those drives if necessary to come up
> as a complete replacement. The second machine will have additional
> drives so it can be used for development and testing - it's only
> 'real' job will be to run amanda backups to tape at night so it
> can be taken down during the day if necessary. Both machines will
> have spare root/user partitions so a complete OS update or reinstall
> can be done on the second machine, tested, then rsync'd to the
> spare partition of the production machine (with it still running)
> and only have the downtime of a single reboot to come up completely
> updated - and if anything fails the old system is still there as
> an alternate LILO choice.
>
> Does anyone have any other tricks to minimize the chances of having
> to hang out at the office late to fix something?
>
> Les Mikesell
> [EMAIL PROTECTED]
Use multiple SCSI Controllers, Multiple drive towers, Hot swapable
drives and mirror all the "we cannot loos" data onto those... Hey..
Consider using a RAID 5 storage array if you have the money.
Then on your "backup" box, have the same number of SCSI controllers, the
exact same root partition available that is on the first box. (and
check it often to be sure.)
Then when you have a failure (and it will happen) you just isolate
what's broke and replace the broken peice with the hot replacement...
So, when your main CPU gets that coke spilled in it, you just re-cable
the disks to the second machine and boot... Or when a drive dies, you
just remove it from the mirror set, and be on your way until morning.
You will spend MOST of your recovery time, doing data restores for data
that gets lost... So having RAID level 5 gives you some redundancy in
the data... You can loose a drive, and as long as you replace it before
another drive fails, all your data is intact... Mirrors do almost the
same thing... They just waste 1/2 of your disk space. Raid only wastes
1/9 of it. If you can arrange things so you don't have to restore all
that data from tape all the time, you will minimize the amount of late
nights spent sitting and waiting for the restore.
Do remember that a good recovery plan will require you to keep good
backups and maintain your backup systems. IT's tempting to just upgrade
your main server, and forget that the new disk you just added is not in
the recovery plan.
-= bob=-
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list (and comp.os.linux.misc) via:
Internet: [EMAIL PROTECTED]
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Misc Digest
******************************
