Linux-Advocacy Digest #12, Volume #30 Fri, 3 Nov 00 02:13:05 EST
Contents:
Re: 2.4 Kernel Delays. ("Les Mikesell")
Re: Linux growth rate explosion! (Goldhammer)
Re: Linux growth rate explosion! ("Bruce Schuck")
Ethernet efficiency (was Re: Ms employees begging for food) (Lars Poulsen)
Re: IBM to BUY MICROSOFT!!!! ("Bruce Schuck")
Re: Why Red Hat is as bad as Microsoft (.)
Re: Ethernet efficiency (was Re: Ms employees begging for food) ("Les Mikesell")
----------------------------------------------------------------------------
From: "Les Mikesell" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy,comp.os.ms-windows.advocacy
Subject: Re: 2.4 Kernel Delays.
Date: Fri, 03 Nov 2000 03:15:37 GMT
"Chad Mulligan" <[EMAIL PROTECTED]> wrote in message
news:rEiM5.14160$[EMAIL PROTECTED]...
>
> > You were right to reboot at the first hint of trouble back then. Before
> > service pack 3 NT was so unstable as to be mostly unusable. It has
> > gotten better to the point that after sp6 even I have to admit that it
> > isn't likely to crash unexpectedly any more unless triggered by hardware
> > problems. But how much trouble have people had to deal with to
> > get to that point? And it is still annoying to have to reboot just to
> > change a machine's name.
> >
>
> Back then it was 3.51.
Oh - the one that would crash if it received a bad packet on the
network...
> The amount of trouble referred to is, IMO part of
> the learning curve needed with any OS. Many of the restarts are
> precautionary rather than necessary.
If you didn't have real crashes you wouldn't have needed to
be so cautious.
> Learning somethng new isn't much
> trouble for me and I tend to experiment on testing systems then implement
> the changes that both work and are beneficial in a group on the production
> systems.
>
> Just a question, why is changing a systems name so important to you?
I often replace services with minimum downtime by bringing up,
tuning and testing a new version of something on a different machine
with a different name, then take out the old one, rename the new,
and go on. With an IP service on unix I can just ifconfig the old
address associated with the service onto the new machine as a secondary
alias and change the hostname. No need to interfere with any other
services it was running by rebooting the whole machine. Besides,
since win2k had many of the old bugs that needed a reboot fixed,
that is a silly one to have missed.
Les Mikesell
[EMAIL PROTECTED]
------------------------------
From: Goldhammer <[EMAIL PROTECTED]>
Crossposted-To:
comp.lang.java.advocacy,comp.os.ms-windows.advocacy,comp.os.ms-windows.nt.advocacy
Subject: Re: Linux growth rate explosion!
Date: Fri, 03 Nov 2000 03:14:01 GMT
In article <bPlM5.120944$[EMAIL PROTECTED]>,
"Bruce Schuck" <[EMAIL PROTECTED]> wrote:
>
> "Goldhammer" <[EMAIL PROTECTED]> wrote in message
> news:1n1M5.11388$[EMAIL PROTECTED]...
> > On Wed, 01 Nov 2000 02:35:55 GMT,
> > Chad Myers <[EMAIL PROTECTED]> wrote:
> >
> >
> > >How 'bout a real workgroup database development
> > >environment for making small business or
> > >workgroup database applications
> > >that an average user can learn
> > >and master in days or weeks.
> >
> >
> > MySQL.
> >
> >
> > >I don't believe there is ANYTHING
> > >on Linux that competes even remotely
> > >with Access and all its functionality.
> >
> >
> > Surely you can't be serious. Access
> > vs. MySQL or Postgres? Do you realize how
> > ridiculous you sound?
>
> You win in ridiculousness. MySQL is a RDBMS.
And what do you think Access pretends to be?
Perhaps you should examine:
http://support.microsoft.com/support/access/content/reldesign.asp
And note it's in the Access section.
You could also start reading "Building Applications
with Microsoft Access 97", page 8, where they begin
explaining to you what the R in "RDBMS" means.
> Access is a database tool that can store it's own data
I should hope so.
> or it can be just a front end to an RDBMS
MySQL supports ODBC.
>and it includes query
>builders, forms/forms designers,
>report writers, VBA etc etc.
All of which (aside from the MS-specific GUI shit)
can be done by simple SQL scripts, fed in to the mySQL
client.
I get the feeling that when Chad Myers blathered
on about the superior "functionality" of Access, you
took this to mean that subset of functionality
of Access which does not pertain to database
management at all, ie, windows-specific gui shit.
--
Don't think you are. Know you are.
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
From: "Bruce Schuck" <[EMAIL PROTECTED]>
Crossposted-To:
comp.lang.java.advocacy,comp.os.ms-windows.advocacy,comp.os.ms-windows.nt.advocacy
Subject: Re: Linux growth rate explosion!
Date: Thu, 2 Nov 2000 19:51:18 -0800
"Goldhammer" <[EMAIL PROTECTED]> wrote in message
news:8ttahl$c50$[EMAIL PROTECTED]...
> In article <bPlM5.120944$[EMAIL PROTECTED]>,
> "Bruce Schuck" <[EMAIL PROTECTED]> wrote:
> >
> > "Goldhammer" <[EMAIL PROTECTED]> wrote in message
> > news:1n1M5.11388$[EMAIL PROTECTED]...
> > > On Wed, 01 Nov 2000 02:35:55 GMT,
> > > Chad Myers <[EMAIL PROTECTED]> wrote:
> > >
> > >
> > > >How 'bout a real workgroup database development
> > > >environment for making small business or
> > > >workgroup database applications
> > > >that an average user can learn
> > > >and master in days or weeks.
> > >
> > >
> > > MySQL.
> > >
> > >
> > > >I don't believe there is ANYTHING
> > > >on Linux that competes even remotely
> > > >with Access and all its functionality.
> > >
> > >
> > > Surely you can't be serious. Access
> > > vs. MySQL or Postgres? Do you realize how
> > > ridiculous you sound?
> >
> > You win in ridiculousness. MySQL is a RDBMS.
>
>
> And what do you think Access pretends to be?
> Perhaps you should examine:
>
> http://support.microsoft.com/support/access/content/reldesign.asp
>
> And note it's in the Access section.
>
> You could also start reading "Building Applications
> with Microsoft Access 97", page 8, where they begin
> explaining to you what the R in "RDBMS" means.
>
Yup. But if you read all the documentation you would see it is so much more.
> > Access is a database tool that can store it's own data
>
>
> I should hope so.
Do you? Access doesn't have to store it's own data. It's flexible.
>
>
> > or it can be just a front end to an RDBMS
>
>
> MySQL supports ODBC.
So does Excel.
>
>
> >and it includes query
> >builders, forms/forms designers,
> >report writers, VBA etc etc.
>
>
> All of which (aside from the MS-specific GUI shit)
> can be done by simple SQL scripts, fed in to the mySQL
> client.
I see how Linux plans to conquer the desktop. It's going to give the users
terminal access to MySQL data and make them write their own sql scripts
instead of giving them Forms and Reports and menus.
Hilarious.
------------------------------
Date: Thu, 02 Nov 2000 19:48:47 -0800
From: Lars Poulsen <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy,comp.arch,comp.os.netware.misc
Subject: Ethernet efficiency (was Re: Ms employees begging for food)
The notion that as ethernet wire utilization goes up, the throughput
peaks at about 30% and after that it becomes an unproductive mess
of collisions is common but incorrect. It is rooted in the assumption
that collisions are bad, abnormal things, and that once utilization
goes up, the retransmission is likely to collide again. Both of those
beliefs are wrong. THAT IS NOT HOW ETHERNET WORKS.
The first point I want to make, is that when a station wants to
transmit, it does not just start to transmit immediately, and
take a collision. Instead, it first tries to sense a carrier,
and if the wire is currently active, it will try to train the new
packet to the postamble of the current packet (plus a "minimum
packet gap"). That means that the current transmission WILL get
through and be productive.
On a two-node network, when the current packet ends, it will be
the new packet's turn. Thus, the wire occupancy can be very close
to 100% with no collisions. (In a switched network, each port
on the switch is a two-node ethernet, to this is a common special
case which requires no special case handling in silicon.)
On a loaded network, there is a good chance that when the node
latches on to the postamble and starts transmitting, it will
collide with someone else who is doing the same thing. COLLISIONS
ARE NORMAL ON A LOADED NETWORK. At this point, the exponential
backoff comes into play to resolve the problem, and the random
element means that one of the (possibly many nodes) will have
a very short delay. On the first collision, the second timeout
will probably expire while the first node is successfully putting
its packet out, and the second node will then train to the end of
that packet.
The people who designed this did a marvelous job, and if the
network is built to spec, a network segment with 100+ nodes
offering more load than the available bandwidth will still
perform at 60+ percent useful throughput.
Max Devlin said that it is unproductive to make a best-case
analysis, and we should assume that the network is not built to
spec. I would argue that in a high-performance environment, we
MUST insist on functional equipment, and while it is nice that
the gear is resilient enough that even a very poorly installed
network will pass data in low-load condistions, we should not
be surprised that it fails when installed outside of the designed
specifications. Most of the ethernet configuration rules are
there to ensure that collisions are detected early. If these
rules (max run lengths, maximum repeaters in path, etc) are
violated, then collisions will be detected late, and the network
will indeed behave the way Max claims it does.
Can we just settle this by agreeing with Max: I'm willing to
believe that the ethernets with which he has been involved
have performed as he says. (I have also worked on networks
that did that, but if it became a problem, we cleaned them
up, and it was not that hard; the specification booklet
told us exactly how to make it work.)
--
/ Lars Poulsen - http://www.cmc.com/lars - [EMAIL PROTECTED]
125 South Ontare Road, Santa Barbara, CA 93105 - +1-805-569-5277
------------------------------
From: "Bruce Schuck" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.advocacy,comp.os.ms-windows.nt.advocacy
Subject: Re: IBM to BUY MICROSOFT!!!!
Date: Thu, 2 Nov 2000 19:53:41 -0800
"Les Mikesell" <[EMAIL PROTECTED]> wrote in message
news:QXpM5.12759$[EMAIL PROTECTED]...
>
> "Bruce Schuck" <[EMAIL PROTECTED]> wrote in message
> news:WimM5.120952$[EMAIL PROTECTED]...
>
> >
> > The OpenBSD people claim they are usually 6 months ahead of Linux/Unix
in
> > fixing exploits.
> >
> > Go ahead and sleep through those 6 months of "open" vulnerabilities.
>
> Why don't you ask them how many years they are ahead of anything
> from Microsoft?
What percentage of the market does the secure open source project (OpenBSD)
hold compared to the insecure one (Linux) ?
And why do so many open source programs have holes in them?
Why do script kiddies prefer Linux to break into and install DDOS tools?
Because Linux is an open door by default!
------------------------------
From: [EMAIL PROTECTED] (.)
Subject: Re: Why Red Hat is as bad as Microsoft
Date: 3 Nov 2000 04:34:50 GMT
Roberto Alsina <[EMAIL PROTECTED]> wrote:
> In article <8tsomc$419$[EMAIL PROTECTED]>,
> [EMAIL PROTECTED] (.) wrote:
>> Roberto Alsina <[EMAIL PROTECTED]> wrote:
>> > El mi=E9, 01 nov 2000, . escribi=F3:
>> >>Roberto Alsina <[EMAIL PROTECTED]> wrote:
>>
>> >>> Since I had not mentioned headers, and I was replying to something
> tha=
>> t
>> > didn't >> say headers, how the fuck is one supposed to guess you
> actuall=
>> y mean
>> > HEADERS, >> when saying KERNEL? If you are too stupid to write
> coherentl=
>> y,
>> > spare the world >> your idiocy.
>> >>
>> >>Because I said "headers" in my first post in this thread on the
> subject.=
>> I=20
>> >>suspected that context would last, but I guess I underestimated you.
>>
>> > Perhaps you should recheck the thread. The first mention to
> "mandrake is=
>> not
>> > linux" came from you and contained zero instances of the word
> "headers".
>>
>> > Read it, see what a stupid fuck you are, then come back.
>>
>> Alright, read it you stupid shit; and heres what I said:
>>
>> "Mandrake is not linux; they completely redid the kernel headers and
>> broke massive amounts of legacy software as a result. Thats why
>> there are "rpm-mdk"s, brainiac."
>>
>> No wonder you cant understand anything. You cant READ.
> Actually, jackass, you seem to be even more stupid than I thought.
> That is not "your first post on this thread on the subject".
> The first post you made on this thread on the subject of why Mandrake is
> not Linux was this:
> http://x66.deja.com/threadmsg_ct.xp?AN=687198524.1
Lets just quote it in full, shall we?
"Absolutely correct. Redhat linux is, next to Mandrake
(which isnt actually linux at all) the biggest piece of shit
unix-look-alike that there is.
This is the SECOND time they did this with gcc...Remember 5.0?
Really, there are three Linuxes worth mentioning at all:
SuSe (so much like FreeBSD now its frightening)
Debian
Slackware"
Now, I think that the important part is where I say:
"Redhat linux is, next to Mandrake
(which isnt actually linux at all) the biggest piece of shit
unix-look-alike that there is."
> Where you did not say "headers" at all.
You're absolutely right. I didnt say KERNEL EITHER, DIPSHIT.
I said it was a piece of shit. I did not qualify it until...
> Your second post was this:
> http://x66.deja.com/threadmsg_ct.xp?AN=687609167.1
> THAT is the one where you mentioned headers, but please notice that was
> before I was involved. I must say I missed it when I replied (maybe
> because it was far from my insertion point), at
> http://x66.deja.com/threadmsg_ct.xp?AN=687718339.1
Right. Thats where I mentioned headers, dipshit.
> However, you are still a dickhead.
I never denied that.
> BTW: even deleting the headers and replacing them with something else
> (which they didn't do) doesn't make it not linux,
Yes it does, and that isnt what they did. What they did was rewrite
them. And it makes it not linux.
> so the point still
> stands: saying Mandrake is not Linux is stupidity.
No, it isnt.
Didnt you killfile me or something?
=====.
------------------------------
From: "Les Mikesell" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy,comp.arch,comp.os.netware.misc
Subject: Re: Ethernet efficiency (was Re: Ms employees begging for food)
Date: Fri, 03 Nov 2000 06:09:30 GMT
"Lars Poulsen" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> The notion that as ethernet wire utilization goes up, the throughput
> peaks at about 30% and after that it becomes an unproductive mess
> of collisions is common but incorrect. It is rooted in the assumption
> that collisions are bad, abnormal things, and that once utilization
> goes up, the retransmission is likely to collide again. Both of those
> beliefs are wrong. THAT IS NOT HOW ETHERNET WORKS.
>
> The first point I want to make, is that when a station wants to
> transmit, it does not just start to transmit immediately, and
> take a collision. Instead, it first tries to sense a carrier,
> and if the wire is currently active, it will try to train the new
> packet to the postamble of the current packet (plus a "minimum
> packet gap"). That means that the current transmission WILL get
> through and be productive.
>
>
> Max Devlin said that it is unproductive to make a best-case
> analysis, and we should assume that the network is not built to
> spec.
On the contrary, it is unproductive to leave it that way.
> I would argue that in a high-performance environment, we
> MUST insist on functional equipment, and while it is nice that
> the gear is resilient enough that even a very poorly installed
> network will pass data in low-load condistions, we should not
> be surprised that it fails when installed outside of the designed
> specifications. Most of the ethernet configuration rules are
> there to ensure that collisions are detected early. If these
> rules (max run lengths, maximum repeaters in path, etc) are
> violated, then collisions will be detected late, and the network
> will indeed behave the way Max claims it does.
>
> Can we just settle this by agreeing with Max: I'm willing to
> believe that the ethernets with which he has been involved
> have performed as he says.
Of course - if you ignore the rules it will not work correctly.
One of the worst properties of ethernet is that it so robust
that even out of spec configurations appear to work some
of the time.
>(I have also worked on networks
> that did that, but if it became a problem, we cleaned them
> up, and it was not that hard; the specification booklet
> told us exactly how to make it work.)
Replacing hubs with switches will often fix things with
no one realizing why.
Les Mikesell
[EMAIL PROTECTED]
------------------------------
** 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.advocacy) 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-Advocacy Digest
******************************
