Linux-Advocacy Digest #201, Volume #30 Sun, 12 Nov 00 23:13:03 EST
Contents:
test!!! ([EMAIL PROTECTED])
Re: NT/2000 true multiuser? ("Les Mikesell")
Re: NT/2000 true multiuser? ("Les Mikesell")
Re: Lets try serious advocacy/discussion. ([EMAIL PROTECTED])
Re: NT/2000 true multiuser? ("Erik Funkenbusch")
Re: Linux + KDE2 = 8{ ([EMAIL PROTECTED])
Re: OS stability (sfcybear)
Re: OS stability (sfcybear)
Re: Of course, there is a down side... ("Les Mikesell")
Re: A Microsoft exodus! ("JS/PL")
Re: The Sixth Sense ("PLZI")
Re: NT/2000 true multiuser? ("Erik Funkenbusch")
Re: OS stability ("Les Mikesell")
Re: OS stability ("Erik Funkenbusch")
----------------------------------------------------------------------------
From: [EMAIL PROTECTED]
Subject: test!!!
Date: Mon, 13 Nov 2000 02:58:04 GMT
test!!!
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
From: "Les Mikesell" <[EMAIL PROTECTED]>
Subject: Re: NT/2000 true multiuser?
Date: Mon, 13 Nov 2000 03:13:30 GMT
"Erik Funkenbusch" <[EMAIL PROTECTED]> wrote in message
news:GoGP5.7837$[EMAIL PROTECTED]...
> >
> > > No, it doesn't. Try running an X program from a telnet login, what
> > happens?
> >
> > On all my machines if the remote has permission to open a window back
> > on the display that is the parent of the telnet session, it will do it
> > automatically. Isn't that the way things are supposed to work?
>
> Really? Now how do you do that without an X server?
If you start out without X, DISPLAY won't be set at all and any
remote X programs will fail to start just as they would if you
tried to run them at the console without starting X.
> Multiuser has nothing to do with what the client is running. For
instance,
> what if I telnet to a Unix server from a Mac or Windows box without an X
> server? Does that make Unix no longer multi-user because the client can't
> support a remote GUI?
Most unix boxes run just fine without X even locally, so X obviously is
not a requirement for multiuser access. You can access just about
everything in character mode. On windows, there is often no character
mode equivalent for GUI programs so telnet access may not be useful.
Les Mikesell
[EMAIL PROTECTED]
------------------------------
From: "Les Mikesell" <[EMAIL PROTECTED]>
Subject: Re: NT/2000 true multiuser?
Date: Mon, 13 Nov 2000 03:16:39 GMT
"Erik Funkenbusch" <[EMAIL PROTECTED]> wrote in message
news:n8HP5.7842$[EMAIL PROTECTED]...
>
>
> Huh? The kind of client is irrelevant. The Win2k TS client works on NT4,
> 9x, 2k, and CE devices. You can use the Citrix client for non-Windows
> clients. Yes, that's an extra expense, but then that wasn't what you
> originally said.
Does the citrix client work against the stock Win2k-server TS server?
Les Mikesell
[EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: Lets try serious advocacy/discussion.
Reply-To: [EMAIL PROTECTED]
Date: Mon, 13 Nov 2000 03:17:03 GMT
On Sun, 12 Nov 2000 18:56:16 GMT, "Les Mikesell"
<[EMAIL PROTECTED]> wrote:
>
><[EMAIL PROTECTED]> wrote in message
>news:[EMAIL PROTECTED]...
>> >fdisk?
>>
>> What's so hard about that?
>> It's included. We were talking about what Admin tools are needed that
>> are not included.
>
>You buy a larger hard disk an want to copy your existing setup
>over to it. How do you do it? With Linux you would connect
>it up, 'cp -a' all your partitions to their new destinations, swap drives,
>boot with a floppy and run lilo to be back in business. Cloning
>an existing setup to a new machine is equally trivial.
Assuming you know how to do all of that under Linux. I don't. Under
Windows best thing would be install new disk with the software that
comes with it (partitioning program) and re-install from the image CD
included with the machine. You could always use Drive image by PPQ
though.
>> >If you add a disk, aren't you doing administration?
>>
>> fdisk is included.
>
>How does that help with moving installed programs?
It doesn't, but that is not what he asked.
Moving programs around under Windows is a pain, no doubt. It is
quicker to re-install, or use drive image. I prefer a re-install
>> >If you add hardware, ...
>> Plug and play or USB trivial under recent versions of Windows. A
>> nightmare under any version of Linux.
>
>They very often are nightmares under Windows as well. I spend a
>couple of days trying to get an ISA network card not to conflict
>with anything else in windows even though I had allocated the
>IRQ in BIOS to legacy/ISA. I've got a modem that just plain
>won't work with win98 but it works fine in DOS and Linux. And
>of course, there is now way to tell *why*.
Sure there is if you know how. I had a similar problem getting a
PCMCIA Token Ring card, Ethernet card, modem, cardbus (which takes
another irq) and the on board sound chip to install together. It was
just a matter of manually assign resources.
>> >I'm still trying it out. So far nothing particularly serious.
>>
>> I liked Mandrake 7.0 / 7.1 but after using it for several weeks I
>> found that the ordinary tasks I do under Windows are more of a pain in
>> the ass with Linux for me. As an example reading news. I had to set up
>> 4 different programs (leafnode, slrn and the editor and spell checker)
>> then I had to fire off leafnode to retrieve messages.
>
>What's wrong with Netscape or Knode?
Netscape is years behind IE in terms of look and features. Just
looking at it even with tt fonts makes my eyes hurt. It locks you out
when loading certain pages and you have to wait till it stops before
you can do anything. Nasty program, even under Windows.
Never tried knode. What is it?
>> Printing was another problem.
>> I was stuck with a monochrome printer.
>
>Odd, no one else is.
And it would print under Netscape and from a command line but nowhere
else including from kde or StarOffice.
>> The permissions thing
>> was another pain, su'ing to root all the time to do things is just
>> extra steps.
>
>Just like admin under NT/Win2k only easier because you don't
>have to kill all your other windows and logout first.
But I find I rarely have to go into Admin under Win2k. Under Linux I
was always screwing around trying to make something work.
>> My wheel mouse half worked in
>> that the wheel worked but you had to click on the window first where
>> as under Windows it hovers and knows what window you want to scroll.
>
>Hmmm, are you really complaining about the lack of mouse configurability
>in the window manger? I think you know better.
I'm saying there IS no mouse configurations other than making the
wheel work.
At least none that I could find.
>> I know there are valid reasons and work a round's for most of these
>> problems, but to me it's not worth it. I'm not the type that wants to
>> set up cron jobs to do all my fetching at night etc. I tend to jump
>> around in programs and under Windows my News reader picks up right
>> where it left off. One click and I can look at past messages in one
>> group while downloading current ones in another. it's very easy and
>> user friendly.
>
>Which news reader does that?
Agent or Gravity.
>> Finally I realized that I was spending too much time fiddling with
>> Linux to customize to work as easily as Windows does for me and that
>> is why I dumped it.
>
>Tell us - how long did you work to find something to complain about
>this time? Takes longer and longer doesn't it?
Same complaints all along, going back to version 4.1 of Redhat.
>> If you can perform all of your functions as easily under Linux as you
>> can under Windows then by all means use it. For me it wasn't even
>> close.
>
>Since your 'functions' seem to be finding something to complain about,
>I can understand why that is easier in Windows. You really
>do have to spend a lot of time to find them in the current Linux
>distributions.
The current Linux distributions suck for a desktop user. They are
getting better and maybe in time will get to to point that they are as
easy to use and as productive as Windows. I never screw around with
Windows settings, they just work. Under Linux it's play, play play and
play again. Read, read read and read some more and then ask on a
newsgroup and get 10 different answers none of which works. Then find
out that you are one of many that are having the exact same trouble.
Waste of time Linux is.
Claire
> Les Mikesell
> [EMAIL PROTECTED]
>
>
------------------------------
From: "Erik Funkenbusch" <[EMAIL PROTECTED]>
Subject: Re: NT/2000 true multiuser?
Date: Sun, 12 Nov 2000 21:19:35 -0600
"Les Mikesell" <[EMAIL PROTECTED]> wrote in message
news:u%IP5.19741$[EMAIL PROTECTED]...
> > Multiuser has nothing to do with what the client is running. For
> instance,
> > what if I telnet to a Unix server from a Mac or Windows box without an X
> > server? Does that make Unix no longer multi-user because the client
can't
> > support a remote GUI?
>
> Most unix boxes run just fine without X even locally, so X obviously is
> not a requirement for multiuser access. You can access just about
> everything in character mode. On windows, there is often no character
> mode equivalent for GUI programs so telnet access may not be useful.
And there are command line equivelants for ghostview? Netscape works in
command line mode? What about the KDE utilities?
Most of the command line programs available for Unix are also available in
Win32 ports. If you want to do work through telnet, you can most certainly
do the same work on Unix or Win2k.
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: Linux + KDE2 = 8{
Reply-To: [EMAIL PROTECTED]
Date: Mon, 13 Nov 2000 03:20:01 GMT
Turning off DNS is the easy part. Fixing all the other stuff that
broke is not.
read his entire message.
claire
On Sun, 12 Nov 2000 22:48:22 GMT, "Les Mikesell"
<[EMAIL PROTECTED]> wrote:
>Let's see: turn off DNS, things break. Too complicated?
>
> Les Mikesell
> [EMAIL PROTECTED]
>
>
><[EMAIL PROTECTED]> wrote in message
>news:[EMAIL PROTECTED]...
>> You're over my head on that one. Sounds like you have your work cut
>> out for you.
>>
>> claire
>>
>>
>> On Sun, 12 Nov 2000 09:32:03 GMT,
>> [EMAIL PROTECTED] (Pete Goodwin) wrote:
>>
>> >[EMAIL PROTECTED] wrote in
>> ><[EMAIL PROTECTED]>:
>> >
>> >>That didn't take long :)
>> >
>> >Oh it gets worse.
>> >
>> >I noticed that smb: works just fine when I was connected to the Internet
>> >via dialup. I checked and found that the DNS addresses from my dialup
>> >were somehow the system DNS (very good for a setup with no DNS and just
>> >two nodes!). I tried to configure the network with no DNS (which is how
>> >I set it up on installation) and now a whole bunch of KDE apps are
>broken:
>> >
>> >klipper (actually when KDE starts)
>> >konqueror
>> >terminal
>> >knode
>> >kbabel
>> >katalog
>> >kmail
>> >
>> >Konsole the KDE crash handler
>> >-----------------------------
>> >
>> >(no debugging symbols found)...(no debugging symbols found)...
>> >(no debugging symbols found)...(no debugging symbols found)...
>> >(no debugging symbols found)...(no debugging symbols found)...
>> >(no debugging symbols found)...(no debugging symbols found)...
>>
>> >0x40c2ee39 in wait4 () from /lib/libc.so.6
>> >#0 0x40c2ee39 in wait4 () from /lib/libc.so.6
>> >#1 0x40c8f8e0 in __check_rhosts_file () from /lib/libc.so.6
>> >#2 0x4049b4d0 in KCrash::defaultCrashHandler () from
>> >/usr/lib/libkdecore.so.3
>> >
>> >#3 0x40bcc008 in sigaction () from /lib/libc.so.6
>> >
>> >Seems like network configuration has blown away a major section of KDE
>2.0.
>> >
>> >(Yes I have logged this as a bug with bugs.kde.org).
>> >
>> >Ho hum
>> >
>> >Pete
>> >
>>
>
------------------------------
From: sfcybear <[EMAIL PROTECTED]>
Subject: Re: OS stability
Date: Mon, 13 Nov 2000 03:07:01 GMT
In article <RfGP5.7833$[EMAIL PROTECTED]>,
"Erik Funkenbusch" <[EMAIL PROTECTED]> wrote:
> "Les Mikesell" <[EMAIL PROTECTED]> wrote in message
> news:G6GP5.19694$[EMAIL PROTECTED]...
> > > Sure, but that's just one item. Routine maintenance should be
conducted
> > on
> > > all parts in the system. Memory, Disk Controllers themselves,
Power
> > > Supplies, CPU's, motherboards, etc...
> >
> > Can you share some statistics on how often your diagnostics have
> > caught any component 'about to go bad' other than a disk drive?
> > In my experience, acutually running in production is a more
> > intense test that any diagnostic test you can get.
>
> Things like power supplies tend to show eratic voltage levels before
> failure. Memory tends to have sporadic bad bits. Most of the time it
> works, sometimes it generates a bad bit. Only a diagnostic doing
millions
> of tests will find that. Motherboards are harder, but most of the
larger
> systems have diagnostic tools available (hardware and software).
>
Let's see, The example you gave was 100 computers having a componet
failure rate of 10% (a number I have not seen you prove) and you said do
test every couple of months. Lets just say that the test last one hour
from the time you start to the time you finish and the system is back up
and running. That would mean that 100 hours would be required for each
cycle of checks and checking every 2 months (since most of the w2k
systems don't seem to stay up even 2 months I say that is a fair number
to run with. That would mean that the tests would have to be run every 6
months for a total of 600 hours or 15 or nearly 1/3 of a year! weeks of
a full time employees time. Now, I know of NO tests that would find ALL
or even a majority of the problems that could cause hardware problems so
you are still going to have hardware crashes. All of this assumes you
have a failover startatagies. Now your 10% fail rate would mean that you
would expect to have 10 hardware crashes in that year and lets asume
that you prevented every one of them (something I doubt you could do)
you have had to work 60 hours (a week and a half) to prevent each of the
10 failures (600/10)! . Grand total of individual server down time 600
hours.
Now, lets say that I have a robust server farm that is built with fail
over capability so the failure of any 1 or 2 boxes could fail without
bringdown any part of the production services (a well designed network)
and we DON't spend any time doing maintenace. we could expect a failure
ever 36.5 days (by your unproven numbers. Now, we have failover so the
overall system will stay up and running. Now in the example you used
above, I would have to work for 60 HOURS one EACH server (1 and 1/2
weeks) inorder to equal the downtime and work that you have put in
preventing the crash. I could order a new system, have it shipped,
Install the OS and restore from back up in that amount of time!!!! But
many of the failures would be something like ram (1 hour total to fix)
or a hard disk that is mirrored under volume management 1 or 2 hours to
fix. If this disk is in an external hardware raid pack that allows hot
swap, I would not even need to bring down the computer!!!!!
Now the 10% number you quoted would be a component failurer rate. That
would have included things like floppys or sound cards that I would not
have in my systems (lowering the number of failures to deal with) Or
CDdrives that I would not need even for most software upgrades and might
choose to ignore untill I needed to reload the OS. At which time the
system would be going down any way....Indeed many secure shops do not
have floppies or CD drives in their systems! These secure shops have the
solaris bios password protected and booted with no cdrom or floppy
support. This means that no one could boot from either of the devices
without the BIOS password and an external CD! Over all, I have saved
MANY hours of work making the MANITENANCE part of TCO MUCH lower!
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
From: sfcybear <[EMAIL PROTECTED]>
Subject: Re: OS stability
Date: Mon, 13 Nov 2000 03:10:37 GMT
In article <hBIP5.7849$[EMAIL PROTECTED]>,
"Erik Funkenbusch" <[EMAIL PROTECTED]> wrote:
> "sfcybear" <[EMAIL PROTECTED]> wrote in message
> news:8unje1$rrq$[EMAIL PROTECTED]...
> > > And what base TCP code would not be compiled into a machine that's
on
> > the
> > > internet? We're not talking about things like IP masquerading
here.
> > We're
> > > talking about the core TCP code.
> >
> > compile all the socket crap into modules...But the you ignor my
> > statemnet "you have NOT proven that they have the configuration
> > REQUITRED to exploit the vunerability!" There are other conditions
that
> > must be met before this bug can be exploited! Can the exploit be
> > prevented another way? Stoping spoofing at a rounter? Defragmenting
at
> > the fire wall? Filtering rulls? Does the exploit require a cetain
socket
> > to be open to exploit?
>
> No, the exploits mentioned do *NOT* require a certain socket. *ANY*
socket
> is susceptable to the exploit. it's quite easy to scan a system to
see
> which ports are open and send malformed packets to any of them. The
problem
> is in the core code that effects all sockets, not in individual
services.
> The same TCP code is used by any socket implementation you use.
>
> > > If you choose to bury your head in the sand and yell "I'm secure
until
> > you
> > > prove i'm not" rather than be proactive, that's your choice.
> >
> > No, I just don't buy into the throw in every patch and assume your
safe
> > attitude of yours. I need to know what and why the patch is going
in.
> > What is the exploit am I really exposed. Information=power!
>
> No, if you apply patches you don't assume you're safe, but it's as
safe as
> you can be until another exploit is found. Instead, you leave
yourself open
> by not applying patches which fix known exploits.
>
please provide proof of this, since the page you posted did not.
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
From: "Les Mikesell" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.advocacy,comp.os.ms-windows.nt.advocacy
Subject: Re: Of course, there is a down side...
Date: Mon, 13 Nov 2000 03:23:52 GMT
"Bruce Schuck" <[EMAIL PROTECTED]> wrote in message
news:Z_GP5.125958$[EMAIL PROTECTED]...
>
> >
> > Note that what you are really saying here is how bad NT and 98 really
are.
>
> NT 4.0 was great when it came out. And with SP6 it is pretty stable. Win2K
> is a lot better.
You wouldn't say that if you ever tried to keep one running before sp3.
Note
that current is sp6a and win2k needs sp1.
> Improvment is a wonderful thing.
If the real thing would ever catch up to the promises it might be.
>
> I try not compare Win2K to Linux circa 1995, but the Penguinistas love to
> compare Linux 2000 (wherever it is) to Win3.1 or Win 95 Or NT 4.0 SP 0.
For good reason. If you bought Win3.1 and wanted something that worked
you would have been told to buy Win95, then 98, then ME, or alternatively
NT 3.5, then 4.0, and then Win2k and you would still be trying.
With Linux, you could have started in 1993 or 4 and downloaded all the
updates free all the way to the current version.
>
> What a bunch of crybabies they are when you note the improvements in Win2K
> and their arguments look even more stupid.
No, Win2k is finally is close to a match, but you have to pay for it all
over again
and if you need the advanced server it is pretty expensive.
Les Mikesell
[EMAIL PROTECTED]
------------------------------
From: "JS/PL" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy,comp.os.ms-windows.advocacy
Subject: Re: A Microsoft exodus!
Date: Sun, 12 Nov 2000 22:22:41 -0500
"Curtis" <alliem@kas*spam*net.com> wrote in message
news:[EMAIL PROTECTED]...
> JS/PL wrote...
> >
> > "Curtis" <alliem@kas*spam*net.com> wrote in message
> >
> > > CC'd mail from ignorant user:
> > >
> > > "Hey guys, I got this nice little file off the net that does so and
so.
> > > It's wonderful. Just double click and run it." He doesn't know about
the
> > > trojan/virus planted there. He doesn't even mention where he got it.
> > > Aferall that's irrelevant info in his ignorant mind. His fellow
ignorant
> > > buddies don't even ask either. This a way too common situation.
> >
> > Future Outlook.net user
> > "Hey guys, you sent me this attachment but my mail reader says it's been
> > blocked and I should see the system administrator. What are you trying
to
> > do, get me in trouble!? Knock it off!!"
> > This will be the way too common situation.
>
> The all too common response to that response would be: "Well, it ran just
> fine over here. Something must be wrong with your program." Do you think
> the home user will call ISP or listen to his buddies who happen to have
> things running. Which is easier? Afterall, the users buddies have things
> working. :-) You guys just aren't practical man.
Actually it will probably be more like:
Newsflash!
Microsoft Outlook Blocks Frog In Blender
Frog in blender maker sues Microsoft and obtains a preliminary injunction
against Microsoft Corporations plan to block most file attachments for users
of Outlook. When asked to comment John Smith, creator of frog.exe stated
that it is a clear intent in Microsofts part to put me out of business so
that they can create their own version of frog.exe. Microsoft (of course)
denies any such plan.
------------------------------
From: "PLZI" <[EMAIL PROTECTED]>
Crossposted-To:
alt.destroy.microsoft,comp.os.ms-windows.advocacy,comp.os.ms-windows.nt.advocacy
Subject: Re: The Sixth Sense
Date: Mon, 13 Nov 2000 03:29:51 GMT
"Les Mikesell" <[EMAIL PROTECTED]> wrote in message
news:QTEP5.19645$[EMAIL PROTECTED]...
>
> "PLZI" <[EMAIL PROTECTED]> wrote in message
> news:FxCP5.196$[EMAIL PROTECTED]...
> >
> > Nothing on *nix platform comes even close
> > in the ease of use and the selection of languages.
>
> How
> do you allow any file to specify it's own interpreter
> and command line flags, and if you can't, why do
> you like it?
This of course can be done with a cute little kludge (ie. using the shell
engine to launch another engine from the script itself, and reading the extra
data within the file), but that is not my point. WSH engine itself has very
little amount of parameters - round 10 or so, most of them displaying logo or
defining the engine (CLI or GUI). What I like, is the amount of components,
and how you can use them.
Let's have a few examples:
- ADSI, my best friend. Manage any user information in LDAP (or use it for
raw LDAP access, whatever the content), Netware, NDS or NT-based system. Same
exact componet works from the VB, C++, VBA App (the whole Office), Delphi ...
you call it, it's there. And the exact same code (given the basic user
properties are available) works against any of the user management systems
mentioned above. You can use it as objects or you can rummage through it with
SQL statements. Took 30 lines of code to move the user accounts from our old
netware to NT server. And I added user's email addresses, which were not
available in the netware user info - took them from our intranet web site
(yup, create a socket component, make a POST query, put the data in NT) - one
problem solved.
- All the other nice, oh so nice readily available interfaces. Index Server.
We do have gigabytes upon gigabytes of Office docs. Want to make a
command-line query against the file server? No prob. Put the same code in
ASP - yup, web page search. Want to have intelligent linking in Word docs?
(like, find me all my previous docs from the file server which have something
to do with this document) - 8 lines of code, same exact component model.
- ADO. Want to go through 100.000 text files, word docs, powerpoint
presentations and so forth in the file server and collect, say file sizes,
authors and what have you in a ...say Solid server database (any
ODBC-compliant database will do). Use Indes Server components (ah, it takes
SQL queries as well, so no need to change the mindset to RegExp or plain
english) to retreive, ADO to put the data in. So next week the Solid is gone,
and there is Oracle on the other end? No prob. Change the ODBC link in
Control Panel, no need to touch the code.
- Nice script you have there. Want to run the script as component? No
problem. Wrap it in XML header, it's a COM component. You can use it in from
another script, VB, C++, Delphi (here we go again), Rexx, Perl... And no
compilation, it stays a script, but behaves like COM object. (you do not have
to know what it is, where it is and so on. Want to distribute it, like DCOM
object? No need to change anything).
- Automation objects. Mmm. I have 500 Word documents here. I need to extract
their content, and ... mmm... collect all the headings in this text file. A
raw-text index, if you like. Create Word as document from your script, and
off you go - you can call any function in Word through the automation
interface in the script. And access all the document's data through it. Say I
want to have our logo in Powerpoint slides changed, 500 .PPT:s here? No prob.
You can tell the Powerpoint automation object to change the picture in the
lower-left corner of every slide to anything you want. Or draw new lines to
the slides while you're at it. And let's have some copyright info in the
slides, as well as the document properties as well. No sweat.
- Say, this script which changes the company logo on the fly is nice. Let's
make sure that all the powerpoint presentations, which get uploaded through
our webserver have this logo. Move the script to IIS, make it an ASP script.
Few little changes, it does that.
- But now I have these stupid word docs here, and I want acrobat. I dont want
to install acrobat distiller everywhere. Hey, I know, let's make a web
service. Upload the doc, call the word automation object on the server
through ASP, make it print the file through distiller. Save the temp, and
stream it back to the uploader. Niiiiice. And not difficult at all.
- But I do not know how to write VBScript. Well, use J(ava)Script. Or perl.
Or Rexx. Or php. Or TCL.
Can you guess by now, why I like the WSH, ASP and all those little other
three-letter acronymes that come with Windows platform?
- PLZI
------------------------------
From: "Erik Funkenbusch" <[EMAIL PROTECTED]>
Subject: Re: NT/2000 true multiuser?
Date: Sun, 12 Nov 2000 21:32:02 -0600
"Les Mikesell" <[EMAIL PROTECTED]> wrote in message
news:r2JP5.19743$[EMAIL PROTECTED]...
> "Erik Funkenbusch" <[EMAIL PROTECTED]> wrote in message
> news:n8HP5.7842$[EMAIL PROTECTED]...
> > Huh? The kind of client is irrelevant. The Win2k TS client works on
NT4,
> > 9x, 2k, and CE devices. You can use the Citrix client for non-Windows
> > clients. Yes, that's an extra expense, but then that wasn't what you
> > originally said.
>
> Does the citrix client work against the stock Win2k-server TS server?
Yes, it does.
Microsofts WTS uses a protocol called RDP, while Citrix Metaframe uses ICA.
Citrix clients understand both RDP and ICA.
------------------------------
From: "Les Mikesell" <[EMAIL PROTECTED]>
Subject: Re: OS stability
Date: Mon, 13 Nov 2000 03:31:53 GMT
"Erik Funkenbusch" <[EMAIL PROTECTED]> wrote in message
news:1DIP5.7851$[EMAIL PROTECTED]...
> "Aaron R. Kulkis" <[EMAIL PROTECTED]> wrote in message
> news:[EMAIL PROTECTED]...
> > I'm a university-educated, systems engineer....and I don't perform
> > much "regular maintenance" on any of the Unix boxes I administrate.
> >
> > Disk drives last 5 years...We do backups every night.
> > If a power supply fails, the vendor will replace it in less than 120
> minutes.
>
> Great, I'm sure your bosses like 2 hours of system unavailability.
And I'd still like to see some real statistics about the number of
components
other than disk drives ever found that have been running perfectly in
production
but fail a diagnostic test. My experience is that hardware is much more
likely
to fail from the power cycling of a reboot then if you just let it run for
years. Of course this takes an OS that doesn't need periodic reboots to
keep it from crashing.
Les Mikesell
[EMAIL PROTECTED]
------------------------------
From: "Erik Funkenbusch" <[EMAIL PROTECTED]>
Subject: Re: OS stability
Date: Sun, 12 Nov 2000 21:39:45 -0600
"sfcybear" <[EMAIL PROTECTED]> wrote in message
news:8unm39$u07$[EMAIL PROTECTED]...
> > No, the exploits mentioned do *NOT* require a certain socket. *ANY*
> socket
> > is susceptable to the exploit. it's quite easy to scan a system to
> see
> > which ports are open and send malformed packets to any of them. The
> problem
> > is in the core code that effects all sockets, not in individual
> services.
> > The same TCP code is used by any socket implementation you use.
> >
> please provide proof of this, since the page you posted did not.
Use your brain.
How could a bug effect only a particular port? Only 2 ways.
1) There is some hard coded special case code in the general IP stack
relating to only that port number. I don't know of any such code in any IP
stack.
2) The service using the stack, that opens the specific port number has the
bug. Not the general TCP code.
Additionally, since it's in the TCP code, and not the UDP code, that means
that only connection oriented protocols are effected, which further
illustrates that it's a general TCP code failure, otherwise it wouldn't be a
kernel patch since userland sockets are a library.
------------------------------
** 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
******************************
