Linux-Misc Digest #30, Volume #26 Sat, 14 Oct 00 20:13:02 EDT
Contents:
Re: Different download types??? (Robert Heller)
Re: How can one app crashing bring down whole system? (Robert Heller)
Re: How can one app crashing bring down whole system? (Robert Heller)
Re: How can one app crashing bring down whole system? (Robert Heller)
IPChains and Cable Modems - Fequently loosing connectivity to the Internet ("Vinson
Armstead")
Re: VMware on Linux -- help making it work! (jeff)
Re: IPChains and Cable Modems - Fequently loosing connectivity to the (Jem Berkes)
Re: VMware on Linux -- help making it work! ([EMAIL PROTECTED])
Re: How can one app crashing bring down whole system? (Hartmann Schaffer)
Re: How can one app crashing bring down whole system? (Hartmann Schaffer)
Re: Installation of PHP 4.0.3 (Bob Hauck)
Kppp question ([EMAIL PROTECTED])
compiling 2.2.16 ([EMAIL PROTECTED])
----------------------------------------------------------------------------
From: Robert Heller <[EMAIL PROTECTED]>
Subject: Re: Different download types???
Date: Sat, 14 Oct 2000 22:18:50 -0000
"Micer" <[EMAIL PROTECTED]>,
In a message on Thu, 12 Oct 2000 06:41:28 -0700, wrote :
"> I am downloading ipmasqadm and I use Redhat 6.0. So far I have the
"> following:
">
"> ipmasqadm-0.4.2.tar.gz
"> ipmasqadm-0.4.2-2.glibc2.tar.gz
"> ipmasqadm-0.4.2-2.i386.rpm
"> ipmasqadm-0.4.2-3.src.rpm
">
"> I am sure there are about 10 other suffixes floating around out there too! I
"> have used ".tar.gz" files before. I have to do a configure and a build. But
"> what is "glibc2"? And why is there an "i386.rpm" as well as a ".src.rpm"?
"> Can anyone elaborate on the many suffixes?
The ipmasqadm-0.4.2.tar.gz file is the sources in compressed tar form.
The ipmasqadm-0.4.2-2.glibc2.tar.gz is probably *binaries* built for a
system with glibc2, in a compressed tar file.
ipmasqadm-0.4.2-2.i386.rpm is a RPM containins binaries built for the
i386 and ipmasqadm-0.4.2-3.src.rpm is a source code RPM.
Basically, ipmasqadm-0.4.2.tar.gz and ipmasqadm-0.4.2-3.src.rpm contain
much the same thing (ipmasqadm-0.4.2-3.src.rpm probably *contains*
ipmasqadm-0.4.2.tar.gz, plus a spec file and possibly some patches).
ipmasqadm-0.4.2-2.glibc2.tar.gz and ipmasqadm-0.4.2-2.i386.rpm are also
probably more or less interchangeable.
Since you are using Redhat, there is little need for you to bother with
either ipmasqadm-0.4.2.tar.gz or ipmasqadm-0.4.2-2.glibc2.tar.gz. You
can either just do a rpm -hiv (or -Uhv) of ipmasqadm-0.4.2-2.i386.rpm or
do a rpm --rebuild ipmasqadm-0.4.2-3.src.rpm followed by a rpm -hiv (or
-Uhv) of /usr/src/redhat/RPMS/i386/ipmasqadm-0.4.2-2.i386.rpm.
">
"> Thanks,
"> Micer
">
">
">
--
\/
Robert Heller ||InterNet: [EMAIL PROTECTED]
http://vis-www.cs.umass.edu/~heller || [EMAIL PROTECTED]
http://www.deepsoft.com /\FidoNet: 1:321/153
------------------------------
From: Robert Heller <[EMAIL PROTECTED]>
Subject: Re: How can one app crashing bring down whole system?
Date: Sat, 14 Oct 2000 22:18:51 -0000
"D. D. Brierton" <[EMAIL PROTECTED]>,
In a message on Sat, 14 Oct 2000 13:24:47 +0100, wrote :
"DB> One of Linux's (and Unix's) strengths is supposed to be the way in which
"DB> each process runs in its own protected mode or space (excuse my
"DB> ignorance of the precise techical terminology). It is often touted that
"DB> unlike in other well-known operating systems one application crashing
"DB> cannot bring the whole system down with it.
"DB>
"DB> On the whole my experince has borne this out. I have many experimental,
"DB> alpha, and beta pieces of software installed, which often crash, or
"DB> hang, or just vanish inexplicably, whilst everything else carries on as
"DB> if nothing has happened.
This is how things should be.
"DB>
"DB> However, both Netscape and Mozilla have managed to do what supposedly
"DB> can't be done. Admittedly, Netscape hasn't done it in some time, but
"DB> when I was running an earlier version of Mandrake and an earlier
"DB> incarnation of Netscape 4.7x the latter would sometimes lock up the
"DB> whole X session so tightly that I couldn't even ctrl-alt-F1 to the
"DB> console, or ctrl-alt-backspace to kill the X session, requiring a hard
"DB> reboot.
I suspect that X11 was wedged or crashed. The kernel almost certainly
was quite alive and well.
"DB>
"DB> Yesterday I installed Mozilla M18 and PSM, the personal security
"DB> manager. Now if I click on Tasks -> Privacy and Security -> Security
"DB> Manager the whole X session locks up. I can ctrl-alt-F1 to the console,
"DB> and from there if I "killall mozilla-bin" then the whole X session
"DB> crashes, and the console becomes unusable while gdm constantly tries to
"DB> restart and fails. Eventually a hard reboot is again required.
Again this is probably an X11 problem.
What do you have for a video card? Which X server are you using?
"DB>
"DB> My question is not about Netscape or Mozilla---I'll post something about
"DB> Mozilla's above behaviour to one of the mozilla groups---but about linux
"DB> itself. I thought that this simply wasn't meant to be possible, given
"DB> the way applications were handled by the operating system.
This is generally true. There are a couple of possibilities: kernel bugs
and hardware problems. It is also possible for some I/O drivers (which
live in kernel space) could be having troubles.
Your problems sound very much like a problem with your X server and
some interaction with your video card. Oh, one other question: are
you using the experimental USB code and using a USB mouse? The USB
code for the 2.2 and 2.3 kernels is not real stable. Note: an alpha or
beta or 'unstable' kernel (or kernel patch or driver module) can crash.
You should *not* be using an alpha or beta or 'unstable' kernel on a
*production* system. It is one thing if you are actively kernel
hacking or part of the kernel alpha or beta test crew (in which case
you presumably know what you are doing).
"DB>
"DB> So, are we guilty of hyping linux's stability into the realms of
"DB> fiction? How can an app effectively bring down the system?
"DB>
"DB> Best,
"DB>
"DB> Darren
"DB>
"DB> --
"DB> ======================================================================
"DB> D. D. Brierton Department of Philosophy, University of Edinburgh
"DB> [EMAIL PROTECTED] http://www.cogsci.ed.ac.uk/~ddb
"DB> ======================================================================
"DB>
--
\/
Robert Heller ||InterNet: [EMAIL PROTECTED]
http://vis-www.cs.umass.edu/~heller || [EMAIL PROTECTED]
http://www.deepsoft.com /\FidoNet: 1:321/153
------------------------------
From: Robert Heller <[EMAIL PROTECTED]>
Subject: Re: How can one app crashing bring down whole system?
Date: Sat, 14 Oct 2000 22:18:52 -0000
"D. D. Brierton" <[EMAIL PROTECTED]>,
In a message on Sat, 14 Oct 2000 13:54:21 +0100, wrote :
"DB> In article <[EMAIL PROTECTED]>, "Tony Lawrence"
"DB> <[EMAIL PROTECTED]> wrote:
"DB> > X isn't Linux.
"DB>
"DB> True. But then if one app is not meant to be able to bring down
"DB> another, why is it that a badly behave X app can bring X itslef to its
"DB> knees? If generally one app crashing/hanging whatever doesn't affect
"DB> the others, and X is itself just more apps I don't understand how
"DB> Netscpae and Mozilla manage to so effectively hose a work session.
X itself is just another application -- unlike the kernel, it does not
always have a good fallback mode. It is not uncommon for one
application to take another with it, especially if the two applications
(processes) are 'connected' in some way (pipes, sockets, ICP stuff,
etc.). What sometimes happens is one process does something that the
other cannot deal with (the bad application has a bug and did a nono and
the other could not deal with the problem and crashed). The X server is
sometimes affected this way. Note: the X server generally diddles the
video hardware (and also the mouse and keyboard as well) and if it
crashes, it might leave the video (and keyboard and mouse) in a 'bad'
state. The kernel is still quite fine, it has cleaned up memory, etc.
(possibly not IPC stuff and might leave sockets and pipes hanging, but
these will eventually timeout).
"DB>
"DB> > And even though you think you couldn't CTRL-ALT-F1, you probably
"DB> > could:
"DB> >
"DB> > You need to have run "savetextmode" BEFORE you have a problem.
"DB> > Therefore, the obvious thing to do is to put it in one of the startup
"DB> > scripts. Then, if you are locked in X, do the CTRL-ALT-F1 and then:
"DB> >
"DB> > press CTRL-U and then CTRL-J.
"DB> >
"DB> > Login, taking it on faith that the "Password" prompt will be displayed
"DB> > even though you can't see it. It wouldn't hurt to count to three after
"DB> > typing your "root" just to give it time to come up. Press enter a few
"DB> > times, and then carefully type "textmode"
"DB> > (again, surrounded by CTRL-J's) to restore your screen.
"DB> >
"DB> > (If you were already logged in on F1, skip the login- duh!)
"DB>
"DB> Thanks for the tip. When you say I should put "savetextmode" in one of
"DB> the startup scripts do you mean one of mine (me qua user) or in one of
"DB> the systen startup scripts?
I guess somewhere in the xinit / startx logic.
"DB>
"DB> Best,
"DB>
"DB> Darren
"DB>
"DB> --
"DB> ======================================================================
"DB> D. D. Brierton Department of Philosophy, University of Edinburgh
"DB> [EMAIL PROTECTED] http://www.cogsci.ed.ac.uk/~ddb
"DB> ======================================================================
"DB>
--
\/
Robert Heller ||InterNet: [EMAIL PROTECTED]
http://vis-www.cs.umass.edu/~heller || [EMAIL PROTECTED]
http://www.deepsoft.com /\FidoNet: 1:321/153
------------------------------
From: Robert Heller <[EMAIL PROTECTED]>
Subject: Re: How can one app crashing bring down whole system?
Date: Sat, 14 Oct 2000 22:18:52 -0000
"D. D. Brierton" <[EMAIL PROTECTED]>,
In a message on Sat, 14 Oct 2000 14:08:23 +0100, wrote :
"DB> In article <[EMAIL PROTECTED]>, "Christopher
"DB> Browne" <[EMAIL PROTECTED]> wrote:
"DB>
"DB> > 1. As far as we can tell from what you have said, there is no
"DB> > evidence that Linux had actually crashed in any of the situations
"DB> > you describe.
"DB>
"DB> Yes. You are quite right. Linux itself had not crashed. The console had
"DB> however become effectively unusable. This still seems to me like a
"DB> linux problem, in that one linux app, say netscape, had effectively
"DB> hosed another, X. And I thought that linux itself was supposed to make
"DB> sure that kind of thing didn't happen.
"DB>
"DB> > 2. There is a fundamental potential problem with software that
"DB> > interacts with the flakey video hardware that is typically sold
"DB> > with PCs.
"DB>
"DB> Right. I hadn't really taken into consideration that part of the cause
"DB> might be hardware related.
"DB>
"DB> > 5. It's still not evident from what you described that Linux itself
"DB> > had any problem at all.
"DB>
"DB> As I said above, it hadn't. I perhaps slightly misdescribed the
"DB> situation. However, with a standalone workstation if the console
"DB> becomes unusable you have soething which is tantamount to the operating
"DB> system crashing.
Not really. There is nothing stopping you from putting:
con:2345:respawn:/sbin/mingetty ttyS0
(or
con:2345:respawn:/sbin/mingetty ttyS1)
in /etc/inittab,
and including ttyS0 or ttyS1 in /etc/securetty.
Once you have done this, you can connect a terminal (i.e. a VT100 or a
LapTop running minicom, etc.) to the specicifed serial port and get in
and fix whatever mess happened to the console. Maybe you can find or
write a svgalib program that forceably sets the video card into a 'sane'
state, thus restoring the console to life.
"DB>
"DB> Thanks for the interesting response.
"DB>
"DB> Darren
"DB>
"DB> --
"DB> ======================================================================
"DB> D. D. Brierton Department of Philosophy, University of Edinburgh
"DB> [EMAIL PROTECTED] http://www.cogsci.ed.ac.uk/~ddb
"DB> ======================================================================
"DB>
--
\/
Robert Heller ||InterNet: [EMAIL PROTECTED]
http://vis-www.cs.umass.edu/~heller || [EMAIL PROTECTED]
http://www.deepsoft.com /\FidoNet: 1:321/153
------------------------------
From: "Vinson Armstead" <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.linux.help,comp.os.linux.networking,comp.os.linux.questions,comp.os.linux.redhat,comp.os.linux.setup,linux.redhat.install,linux.redhat.misc
Subject: IPChains and Cable Modems - Fequently loosing connectivity to the Internet
Date: Sat, 14 Oct 2000 22:28:58 GMT
This is a strange problem...Any suggestions or advice would be greatly
appreciated
I have been having a problem with keeping my cable modem connection active
when using my RH (first 6.2 now 7.0) server as a Firewall (running IPChains
1.3.9-17)
**Background**
I have a home LAN running a few different workstation/server (for FUN). All
the workstation/server are connected to a 10/100 Ethernet switch. The RH
server had two interfaces one Internal (192.168.x.x) and one external
(y.y.y.y).
With the original configuration I had the cable modem directly connected to
the same switch as all the other workstation/server interfaces (including
both RH server interfaces). I know technically I have two Layer 2 domains
connected to one Layer 2 device but I didn't have an extra switch ;)
====================================
| Ethernet Switch |
====================================
| | | | | |
| | | | | |
RH ws1 ws2 ws3 CM
This worked fine until I added another system to the home LAN. I ran out of
ports and decided to connect the cable modem directly to the external
interface of the RH server.
====================================
| Ethernet Switch |
====================================
| | | |
| | | |
RH ws1 ws2 ws3
|
|
CM
I know this is a "more correct" way of connecting my little network and
probably more secure, but I have been having problems maintaining
connectivity to the Internet ever since.
**Problem**
It seems that (frequently) my cable modem looses connectivity to the
Internet but after running several tests it seems that the cable modem may
not be the actual problem.
When I noticed that no one on the LAN could get to the Internet I would
(from the workstation) ping the LAN side of the RH server (this was always
successful), then the external side of the RH server (this was always
successful), then the IP address of the actual cable modem (10.x.x.x) (this
always failed) , then the external network default gateway (this always
failed).
I then checked and re-seated all the cables (all OK but did not restore
service). I ran the same test from my RH server and got the same results as
above. I would stop/start the IPChains FW (still service was down). I
rebooted the RH server (still service was down). The only way to restore
services was to power off/on the Cable modem.
I had the cable company check all my coax cables (they fixed a few problems
and my signals to the modem are "crystal clear) and replace the cable modem
and am still having the same problem. Strange but true!!!!
I them removed some of my workstation from the LAN and connected the cable
modem and the internal and external interfaces of my RH server (running
IPChains) to the Ethernet switch and all worked!! I was able to download all
five RH 7.0 ISO files (over 2.5 GB) from the Internet without intervention.
Prior to this last change I would have to power off the cable modem every 15
or 20 minutes because I would loose connectivity to the Internet.
At this point I thought I had it licked...but no such luck.
I bought a 5 port 10MB Ethernet Hub and tried to connect the cable modem and
the external interface of the RH server to the hub while all the other
devices were connected to the 10/100 Ethernet switch. This setup also began
to fail!!! I tried a number of different configuration and better luck.
It seems that when I have the cable modem connected to the Ethernet switch
everything works fine, when I connect it to the HUB things fail.
I have no idea why, the only difference I can see is that the Hub will have
more Ethernet collisions than the switch, but that is normal everyday
Ethernet networking. If there are collisions then everyone pauses for a bit
then begins again.. IEEE 802.3 CSMA/CD.
I am almost embarrassed to say I am a SR Net ENG for a large International
organization. I have been working with Ethernet since 90'..in this case I
know what works I just don't understand why???
------------------------------
From: [EMAIL PROTECTED] (jeff)
Subject: Re: VMware on Linux -- help making it work!
Date: 14 Oct 2000 22:44:10 GMT
Reply-To: [EMAIL PROTECTED]
On Sat, 14 Oct 2000 16:29:41 -0400, Carlos Moreno <[EMAIL PROTECTED]> wrote:
>
> I have no idea what could be the appropriate newsgroup to
> seek for help on this, but I'm hoping that I'll be able to
> find some info or pointers with you guys.
>
> I just downloaded the trial version of VMWare for Linux, to
> emulate Win on my Linux box (I'm so sick of this dual boot! :-))
>
> When I run it (after installing and configuring everything),
> I try to emulate Win2K, and it goes through the boot process
> (including counting the memory, etc.), but it ends up telling
> me "Operating System not found"
>
> (pretty much as an actual machine would tell you if you
> didn't connect the hard drive, or if there wasn't a hard
> drive at all).
>
> I went to the vmware site, but didn't find any FAQ or
> troubleshooting section. Could anyone give me a hand, or
> give me some pointers about where to find the info or where
> to ask?
Are you using a virtual disk or raw disk? If raw, is the disk set up with a
properly configured multi-boot program? If virtual, are the .cfg, .dsk,
etc. files properly built, residing in the proper directory, and with proper
permissions? (Mine are set to 744 for .dsk and 644 for everything else
(.dsk, .log, and .nvram, in my case - also win2k, BTW.))
To ensure that VMware itself is sound, you can try booting from a diskette.
Also, there is help available starting at http://www.vmware.com/support/.
-jeff
------------------------------
From: Jem Berkes <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.linux.help,comp.os.linux.networking,comp.os.linux.questions,comp.os.linux.redhat,comp.os.linux.setup,linux.redhat.install,linux.redhat.misc
Subject: Re: IPChains and Cable Modems - Fequently loosing connectivity to the
Date: Sat, 14 Oct 2000 18:24:49 -0500
> I then checked and re-seated all the cables (all OK but did not restore
> service). I ran the same test from my RH server and got the same results as
> above. I would stop/start the IPChains FW (still service was down). I
> rebooted the RH server (still service was down). The only way to restore
> services was to power off/on the Cable modem.
A friend of mine who has a cable modem tells me that occasionally due to
cable modem buffering errors (hardware) he has to disconnect the cable
modem and switch it off/on. Have you ever used this particular cable
modem on (e.g.) a Windows machine without the same problem?
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: VMware on Linux -- help making it work!
Date: Sat, 14 Oct 2000 23:20:53 GMT
In article <[EMAIL PROTECTED]>,
Carlos Moreno <[EMAIL PROTECTED]> wrote:
> I have no idea what could be the appropriate newsgroup to
> seek for help on this, but I'm hoping that I'll be able to
[...]
> When I run it (after installing and configuring everything),
> I try to emulate Win2K, and it goes through the boot process
> (including counting the memory, etc.), but it ends up telling
> me "Operating System not found"
Have you tried contacting vmware? They were quite helpful when I
contacted them with some questions before I bought their product.
I've been quite pleased with it, and find that W98 runs better (slower,
but more stable) under vmware than it does directly on the hardware. I
have not tried w2k or wme on it yet.
--
Jim Buchanan [EMAIL PROTECTED] [EMAIL PROTECTED]
=================== http://www.buchanan1.net/ ==========================
"It's a good thing the USS Hopper wasn't going to the Seattle area, or I
might have tried to talk the skipper into bombarding Microsoft"
-Jerry Pournelle
========================================================================
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
From: [EMAIL PROTECTED] (Hartmann Schaffer)
Subject: Re: How can one app crashing bring down whole system?
Date: 14 Oct 2000 15:34:23 -0400
In article <[EMAIL PROTECTED]>,
Tony Lawrence <[EMAIL PROTECTED]> wrote:
> ...
>You need to have run "savetextmode" BEFORE you have a problem.
is that part of some package? if not where can i find it?
> ...
hs
------------------------------
From: [EMAIL PROTECTED] (Hartmann Schaffer)
Subject: Re: How can one app crashing bring down whole system?
Date: 14 Oct 2000 15:39:52 -0400
In article <8s9ls8$jrp$[EMAIL PROTECTED]>,
D. D. Brierton <[EMAIL PROTECTED]> wrote:
> ...
>> 5. It's still not evident from what you described that Linux itself
>> had any problem at all.
>
>As I said above, it hadn't. I perhaps slightly misdescribed the
>situation. However, with a standalone workstation if the console
>becomes unusable you have soething which is tantamount to the operating
>system crashing.
it might be worthwhile to get a cheap second machine so that you can
telnet into the machine with the locked up console
hs
------------------------------
From: [EMAIL PROTECTED] (Bob Hauck)
Subject: Re: Installation of PHP 4.0.3
Reply-To: bobh{at}haucks{dot}org
Date: Sat, 14 Oct 2000 23:43:42 GMT
On Sat, 14 Oct 2000 16:14:13 +0100, Jean-St�phane Lebrun
<[EMAIL PROTECTED]> wrote:
>5. In the file "/usr/local/apache/conf/httpd.conf"
>
>I add :
>AddType application/x-httpd-php .php
>AddType application/x-httpd-php-source .phps
Shouldn't that be:
AddType application/x-httpd-php4 .php
AddType application/x-httpd-php4-source .phps
--
-| Bob Hauck
-| To Whom You Are Speaking
-| http://www.haucks.org/
------------------------------
From: [EMAIL PROTECTED]
Subject: Kppp question
Date: 14 Oct 2000 16:54:41 PDT
I just installed Red Hat 6.2 and have set KDE as my default desktop.
Kppp works fine running as root but when I try to run it as a user
it prompts for a root password.
I tried what the docs said, made a kppp.allow and tried a group as well.
Still only root can run kppp. Any suggestions? Thanks.
------------------------------
From: [EMAIL PROTECTED]
Subject: compiling 2.2.16
Date: Sat, 14 Oct 2000 23:58:58 GMT
I just installed RH7.0. and I am have 2 problems,
1) When I try to compile the Kernel, I get the following error
messages (this is after I have done make dep and make clean)
/usr/src/linux/include/linux/modules/i386_ksyms.ver:142:1: warning:
this is the location fo the previous definition
make [2]: *** [ksyms.o] error1
make [2]: leaving directory 'usr/src/linux-2.2.16/kernel
make[1]: ***'first_rule] Error2
make[1]: leaving directory '/usr/src/linux-2.2.16/kernel'
make: *** [_dir_kernel] Error 2
the second problem has to do with a USB port that I just added,
when I boot with the USB card in the computer the system hangs after
it mounts the USB filesystem and while it is checking the filesystem.
When I take the card out it boots fine starting the driver and
mounting the USB filesystem. The computer uses USB-uhci drive. I was
going to try the other driver but I am also having the problem above
and con't compile a kernel.
Any help is greatly appreciated
Thank you
John Miller
------------------------------
** 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
******************************
