[SLUG] Source IP with multihomed box

[Andy Eager]

Hi all,
I have a question regarding a multihomed box.
Does anyone know how the kernel decides what source IP address to use on 
a multihomed box?

For example, I have a box with:
eth0 
 inet addr:192.168.1.244  Bcast:192.168.1.255  Mask:255.255.255.0

eth1 
 inet addr:192.168.99.241  Bcast:192.168.99.255  Mask:255.255.255.0

eth1:1  
 inet addr:192.168.98.240  Bcast:192.168.98.255  Mask:255.255.255.0

and a default route set to 192.168.99.244 will send a packet out its 
default gateway with a source address of 192.168.98.240.

Seems a little strange to me.
I can see why eth0's IP address might not be used, but why not eth1 
instead of eth1:1's address?

If I swap eth1 and eth1:1 around, then the source address moves with it 
(ie becomes 192.168.99.241)

I'm doing stuff using ip routing rules and I need to know when and under 
what conditions a specific IP address is used.

Thanks in advance,
Andy

--
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

[SLUG] SCSI termination and 80 pin devices

[Andy Eager]

Hi all,

I've got an interesting question regarding SCSI bus termination and 
newer 80 pin devices.

I have:
1 x Adaptec 2100S RAID controller
2 x 36 GB 80 pin SCSI drives & hot-plug boxes.
I have done a lot with SCSI over the years, and it was my understanding 
that the SCSI bus must be terminated at both ends.  The controller 
(normally) handles one end and the other with either active or passive 
terminators. (on the disk or the bus)

I had some problems with the configuration and when querying Adaptec, 
they came back and said:

If you do have drive enclosures, note that you cannot use 
the Adaptec supplied cable with the card. This cable has terminating
resistor built-in and cannot be used on hard-drive enclosures. 
Basically, you do not want to use terminated cable under any 
circumstances if you have drive enclosures or back plane.

Anyone familiar with this?  Are drive enclosures 'auto terminating' now?

TIA

Andy



--
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html

Re: [SLUG] Linux Certification

[Andy Eager]

Terry Collins wrote:

I think you are confusing two things here; the courses run by Geoff
Robertson at Granville TAFE and LPI Certification.
 

No confusion here, Geoff (and I) have been teaching LPI at Granville 
Tafe for the past 4 semesters now (I'm actually taking this semester off 
but will probably go back to it next year).

I found the LPI course work to be very relevant to Linux users generally 
and sysadmins in particular.  I have used pretty much all of the stuff I 
learnt from LPI in running my own business, which is based on Linux 
support and coding: Not that LPI teaches you much about coding.
I actually found teaching it the best way to learn it!! (Yes I have LPI 
and got it first go along with Geof Robertson in June last year.

Earlier this year I did some course development work for Tafe NSW and 
spent a lot of time getting the LPI courses put into the AQF (Australian 
Qualification Framework).  It is hoped that it will become part of the 
AQF sometime next year which means it will be recognised both by people 
in the 'Linux world' as well as those employers who like to see TAFE 
certificates.

My $2 worth... since inflation and the removal of the 2 cent coin.

Andy Eager

--
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] Emacs & custom languages

[Andy Eager]

Hi all,

Anyone know how to write a syntax checker for a new language under emacs?

I want to use emacs to write code in a custom language (called 
Interactive Audio Scripting - IAS).

I have designed the language & written an interpreter for it, but would 
like emacs to do some syntax highlighting for me.

Have checked the emacs site & general google search shows up nothing.
Any clues?
Thanks

Andy

--
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] C - exec fn call & env vars

[Andy Eager]

Hi all,

Got a bit of a C code using execle() to call a shell script with a 
specific environment.  I've got no trouble seeing the environment vars 
in bash.

Is there any way of setting an environment variable in the shell script 
so that it modifies the environment of the *parent* process?
I fear not - but is there any way of returning something from a shell 
other than through an exit status?

Thanks,

Andy

--
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] ncurses & key scan codes

[Andy Eager]

Graham Smith wrote:

On Tue, 1 Apr 2003 16:18, Andy Eager wrote:

Hi all,

I am writing some code using ncurses and would like to be able to
distinguish between:
   Tab-Key
   Shift Tab-Key
I have tried using getch in both keypad and normal mode but all I seem
to get back is 0x09 (normal tab)
So far only tried it with TERM=linux and TERM=xterm, but you'd think
they would have back-tab functionality.
Thanks,

Andy

Hi Andy,

You can use 'od' to check keyscan codes. I used it to modify the termcap file.

No problem getting the scan codes, I find that Tab & Shift-Tab both give 
me 0x09 quit nicely.

What I want is something other than 0x09 for Shift-Tab.

Looks like terminfo (replacement for termcap by the look of it) allows 
me to define what I want but the syntax for it looks quite daunting!!
(sendmail for the keyboard!!!)

Andy

Thanks for your help

--
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] ncurses & key scan codes

[Andy Eager]

Hi all,

I am writing some code using ncurses and would like to be able to 
distinguish between:
   Tab-Key
   Shift Tab-Key

I have tried using getch in both keypad and normal mode but all I seem 
to get back is 0x09 (normal tab)

So far only tried it with TERM=linux and TERM=xterm, but you'd think 
they would have back-tab functionality.

Thanks,

Andy

--
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] wvdial

[Andy Eager]

DNS has nothing to do with wvdial.

Normally it would be set up when ppp negotiates the link.  (Normally in 
'stupid' mode)

To set DNS manually, enter the nameservers of your ISP into 
/etc/resolv.conf:

/etc/resolv.conf:

nameserver  
nameserver

Hope this helps,

Andy

Geoffrey Cowling wrote:

Having had my hand helo by KDE I am with Debian almost in a Catch-22
situation, not being able to upgrade until I can access the internet.
Using wvdial I can log on, get ppp (with its PID), but clearly have no
access to DNS.  (URL, unless = number, times out). The man page speaks
of a Check DNS (presumably? if auto does not work?), but there is no
syntax/format given for the DNS Test1 etc. How do I put the DNS address
in the wvdial.conf file? Does it go in defaults or a separate section?
thanks
Geoffrey


--
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Testing for a SEGV before it happens

[Andy Eager]

Dave Airlie wrote:

Is it possible to test the accessability (rw or ro) of a memory region
before writing to it?
(I want to be able to test whether or not a string is in the code or
data segment)
uggh.. do you not know? did you not put it there?

No.  I didnt put it there.  I am writing a library function as follows:
void  log_string(char *in_str)
{
.  open file, write string, close file
.  write '\0' to *in_str to clear the log msg if in data seg
}
I agree it's ugly, but it needs to be done for backward compatablity 
reasons.

Under QNX I had a function _ispermitw(void *addr) to test write access.

Yes, from user space.
From kernel space I actually know how to do!!
Regards,

Andy



--
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] Testing for a SEGV before it happens

[Andy Eager]

Hi all,

I've got a GNU-C question:

Is it possible to test the accessability (rw or ro) of a memory region 
before writing to it?

(I want to be able to test whether or not a string is in the code or 
data segment)

Thanks,

Andy

--
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] Code Information

[Andy Eager]

Thanks to everyone for this.  Nice summary for search:

objdump
readelf
size

Andy

--
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] Code information

[Andy Eager]

Hi all,

Anyone know of a program that will tell me the code & static data 
segment sizes of an executable ?

(without running it and using ps)

Thanks,

Andy


--
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Serial Tunneling

[Andy Eager]

Richard Luckhurst wrote:


Hi list

I am trying to sort out a problem with a set up for a client. I have
2 PC's with 8 port serial cards in them and they work fine under
redhat 7.2. At present I have the 8 ports connected to ports in 
/etc/services and I can telnet to these ports OK. The idea is I
want to push serial data in a port and have it come out on the
corresponding port at the other end. I have the 2 computers
connected with a dial up ppp link. What I need is some way
of connecting or tunnelling the ip ports. Has anyone got any ideas?

stunnel is probably what you want.  

www.stunnel.org

Andy




--
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] xdmcp

[Andy Eager]

Gareth Walters wrote:


- Original Message - 
From: "Howard Lowndes" <[EMAIL PROTECTED]>
To: "Mail List - SLUG" <[EMAIL PROTECTED]>
Sent: Tuesday, November 19, 2002 11:18 AM
Subject: [SLUG] xdmcp


on a RH system running kdm, I have put the following into the /etc/kderc
file:

[Xdmcp]
Enable=true


You also want the following line commented out in /etc/X11/xdm/xdm-config:

! SECURITY: do not listen for XDMCP or Chooser requests
! Comment out this line if you want to manage X terminals with xdm
DisplayManager.requestPort: 0

(Note that the comment starts with a bang

Andy


--
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] KDE sound server dies

[Andy Eager]

Andy Eager wrote:


Hi all,

I have noticed that when KDE starts up the sound system works just 
fine (When I switch desktops, my  machine makes a cute little 'quack' 
sound).

After some (long) time, it stops and the following messages are sent 
to .xession-errors:

Unable to open file /usr/share/sounds/KDE_Desktop3.wav

I can:
cat /usr/share/sounds/KDE_Desktop3.wav > /dev/dsp 

Sorry, that plays garbage, should have read:
play cat /usr/share/sounds/KDE_Desktop3.wav -d /dev/dsp




and I hear the ever familiar 'quack' sound, so the file is definately 
there and valid.

I have tried restarting the sound server from the KDE control centre, 
but still no joy.
Logging out of KDE and back in again fixes the problem.

Any ideas?

Andy





--
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] KDE sound server dies

[Andy Eager]

Hi all,

I have noticed that when KDE starts up the sound system works just fine 
(When I switch desktops, my  machine makes a cute little 'quack' sound).

After some (long) time, it stops and the following messages are sent to 
.xession-errors:

Unable to open file /usr/share/sounds/KDE_Desktop3.wav

I can:
cat /usr/share/sounds/KDE_Desktop3.wav > /dev/dsp

and I hear the ever familiar 'quack' sound, so the file is definately 
there and valid.

I have tried restarting the sound server from the KDE control centre, 
but still no joy.
Logging out of KDE and back in again fixes the problem.

Any ideas?

Andy


--
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] iptables ip-address format question...

[Andy Eager]

[EMAIL PROTECTED] wrote:


G'day all...

When creating my iptables firewall rules, if I specify an address as 
192.168.0.0 will that match all 192.168.*.* addresses?

No, you'd want 192.168.0.0/16 for that

Andy


--
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Simple Newbie Questions

[Andy Eager]

Curnow wrote:


G'Day Gang,
I have a couple of simple(to you guys) questions regarding Red Hat 7.2 
using KDE as desktop

1: When a new program is installed from the root account how do I make 
it available to the other users? 

It is automatically available to other users (assuming it was meant to be).
You only need to install it from the root account.



2: I am using Apache to test cgi scripts and need to set up a mail 
server. Does Apache have mail server capabilities and how do I add new 
users to it, or do I need a separate mail server to do this? 

Install postfix for this or struggle with the (probably) allready 
installed sendmail.

Hope this helps.

Andy



--
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] Encrypted send/recv on a socket

[Andy Eager]

Hi all,

I am writing a simple tcp socket application: One machine opens a socket 
and listens while another machine connects to the socket and sends / 
receives messages.

I need to have the data encrypted in both directions.  One thought was 
to use ssh with the port forward option (-L port:host:hostport).

Just wondering if anyone knows of a library function similar to send / 
receive that handles encryption so I dont have to use ssh. (Similar to 
SSL as used by https)

Thanks,


Andy Eager.

--
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Open Office crashes X

[Andy Eager]

Ken Foskey wrote:

  On Wed, 2002-10-02 at 11:17, Andy Eager wrote:
  
Hi all,I recently installed open office (in network mode) and haven't been able to have it run without crashing X.I have Star Office 5.2 running on the same system(s), installed in the same way and they seem to work fine.
  
  This sounds like the error with the X windows environment. Check whetheryou can upgrade your X,  there was a bug in one of the versions,OpenOffice.org just triggered that bug reliably.KenF


Yep, I worked that out: 
    Went and bought Star Office 6.0
    Installed it & same thing happened.
    
I found that there was a problem with the XFree 4 mga driver (Matrox G200)
Just before it crashed, I got a message from the kernel:

Oct  2 20:29:24 Node4 kernel: mtrr: 0xd800,0x100 overlaps existing
0xd800,0x80

Anyway, I finally gave up (after disabling dri ...  etc etc) and just went
back to the V3 driver (XF86_SVGA)

That all works fine and now I've got both.

Thanks all,

Andy

[SLUG] Open Office crashes X

[Andy Eager]

Hi all,

I recently installed open office (in network mode) and haven't been able 
to have it run without crashing X.
I have Star Office 5.2 running on the same system(s), installed in the 
same way and they seem to work fine.

After installing on my server on an nfs drive (/nfs/usr) using setup 
-net and then doing a local workstation
install as a normal user, I start up open office and the following happens:

a)The Open Office logo appears
b)The writer appears and I start typing
c)At some (random) time while I am typing, my entire X server 
crashes and gdm presents me with a new logon screen.

All I am left with is a core dump (around 50 MB) and a relatively boring 
message in .xsession-errors that looks as if you'd killed X
yourself.

I've tried this in the following scenarios:
- as root on the server
- as a normal user on the server
- as root on a 'workstation'
- as a normal user on the server

I did notice that clicking the spell checker was a sure way to bring on 
an instant crash.

Environment:
- RedHat 7.3 on all machines
- Running XFree86 (version 4)
- kernel 2.4.18-4

Any Ideas?

Andy E.



-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Redhat Logo

[Andy Eager]

Jeff Waugh wrote:

  
  
I am setting up a Redhat 7.2 box for a client and they have just asked mehow I can kill the Redhat Logo page that sits behind the graphical loginand replace it with an image of their own.  I've never even consideredthis before. Does anyone know how to do it?
  
  Run gdmsetup as root, or just edit /etc/gdm/gdm.conf.

I think thats meant to be gdmconfig.  Also, gdmphotosetup will let you add
icons for each user using the chooser.

Andy

[SLUG] name resolution

[Andy Eager]

Hi all,

Is there an easy way to tell what name resolution scheme was used in 
order to resolve a host lookup:

eg:

ping node3.mydomain   --> search dns... not found,  search files...found

ping www.blah.com.au  --> search dns... found

That is, tracing the paths/actions defined in nsswitch.conf and what 
each action returned.

Regards,

Andy


-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Why won't a time server synch

[Andy Eager]

Howard Lowndes wrote:

>I have an ntpd daemon running on one of my boxes and it is synch'ng with
>external clocks just fine:
>
>ntpdc> peers
> remote   local  st poll reach  delay   offsetdisp
>===
>=LOCAL(0)127.0.0.1   10   64  377 0.0  0.00 0.00095
>=130.155.98.1192.168.254.17   1 65536  377 0.06883 -0.017593 0.94656
>*128.250.36.2192.168.254.17   1 65536  374 0.07060  0.003403 0.92740
>=192.189.54.17   192.168.254.17   2 65536  377 0.06625 -0.156729 0.94658
>
>but if I try to synch with it, even as the localhost I get the following:
>
># ntpdate -bu localhost
>28 Aug 04:51:17 ntpdate[27869]: no server suitable for synchronization
>found
>

Have you lifted restrictions on localhost and your local network in 
/etc/ntp.conf and setup localhost as a server ?

server 127.0.0.1

restrict 127.0.0.1
restrict 192.168.1.1 mask 255.255.255.0



-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] usb 32MB memory stick

[Andy Eager]

david_perry wrote:

>I have an IBM 32MB memory stick which I can't get to work on my Redhat 
>7.2 box. I've followed Stuart's advice (slug posting 8 Aug 02, IBM 
>Linux Roadshow -USB Mem Key works on Linux).  I ran lsmod and confirmed 
>that usbcore, usb-hhci, usb_storage, scsi_mod and sd_mod were 
>installed.  I had to use insmod to add sg.o
>
>When I try and mount the memory stick I get a message saying
>" mount: /dev/sda1: unknown device "
>
>I don't have any other scsi devices present.  Any pointers what to do 
>next?  I'm fairly new at this so it's probably something dead obvious!
>
Have a look at /var/log/messages to see firstly if the device was 
recognised when you plugged it in.
Then have a look at /proc/scsi/scsi and you should see it there.  If it 
is there, then you should be able to access it using the normal sd_mod 
(upper level driver that sits on top of scsi_mod).  You shouldn't need sg.o

Andy



-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] device driver entries in /dev

[Andy Eager]

Jessica Mayo wrote:

>On Thu, 1 Aug 2002, Andy Eager wrote:
>
>>Peter Hardy wrote:
>>
>>>I don't think there's an easy way to tell, at any given point in time,
>>>which USB device is which.  About the best you can do is say that
>>>devices are enumerated in the order they're plugged in, and know in what
>>>order they're plugged in.  Of course, for devices that were there at
>>>boot, it isn't quite that simple..
>>>
>>Thats a shame.  The same problem happens when scanner.o is loaded:  The 
>>only way to find out which device to use is to run sane-find-scanner.
>>
>
>Ok. I don't have usb set up anywhere to look at this, but look under
>/proc/scsi/
>
All that shows is what scsi devices are installed:  Not the /dev entry. 
 (In this case /dev/scd0)
Attached devices:
Host: scsi0 Channel: 00 Id: 00 Lun: 00
  Vendor: CREATIVE Model: CD5233E  Rev: 1.00
  Type:   CD-ROM   ANSI SCSI revision: 02

>
>and see what's there. There should be some info about what each scsi
>device is. also see if there is a
>/proc/usb/
>directory. I'm guessing now, though.
>
Once again, only shows the devices: Not the /dev entry used by the module.

>
>I would like to know if you find something useful.
>There's lots of interesting stuff in /proc, and that's probably where
>sane-find-scanner is looking.
>
No, I think sane-find-scanner is probing well known scanner entries for 
a scanner.
(By actually opening the device and asking it)


-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] device driver entries in /dev

[Andy Eager]

Peter Hardy wrote:

>On Tue, 2002-07-30 at 21:27, Andy Eager wrote:
>
>>My question is:  Is there any way of telling which /dev file is assigned 
>>to the new device?  I had a poke around /proc for some kind of clue but 
>>couldn't see anything.
>>
>
>I don't think there's an easy way to tell, at any given point in time,
>which USB device is which.  About the best you can do is say that
>devices are enumerated in the order they're plugged in, and know in what
>order they're plugged in.  Of course, for devices that were there at
>boot, it isn't quite that simple..
>
Thats a shame.  The same problem happens when scanner.o is loaded:  The 
only way to find out which device to use is to run sane-find-scanner.

Thanks for the confirmation regarding /dev entries.

Regards,

Andy E.

-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] device driver entries in /dev

[Andy Eager]

Hi all,

A general question regarding kernel drivers (modules) and how they 
attach a device into the /dev filesystem.

I have a USB CD-ROM which when plugged in is recognised and enumerated 
by the USB sub system as expected.

My question is:  Is there any way of telling which /dev file is assigned 
to the new device?  I had a poke around /proc for some kind of clue but 
couldn't see anything.

I know I can find it by attempting a mount on all the usual entries 
(/dev/scd0, scd1 ...) but there must be a better way.

I had a similar problem at Granville TAFE where, after presenting a 
lesson on USB, a student pulled out a USB memory stick and the same 
thing happened:  It was recognised by usbcore OK, but I couldn't mount 
it.  How would I go about finding where sg.o puts it in /dev?  (I assume 
that usb-storage and sg are the only drivers needed)

Regards,

Andy

-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] Firewalls & ident service

[Andy Eager]

Hi all,

I know this question is open to debate, but is it wise or desireable to 
offer auth services through a firewall?

I gather it is only used when sending mail by the remote smtpd to 
identify the sender.  At the moment I reject incomming packets bound for 
port 113 with a tcp-reset.

On another point, what about incomming traceroutes.  Are they really 
dangerous?

Any thoughts?

Regards,

Andy E.

-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] alternative to fetchmail

[Andy Eager]

Hi all,

Anyone know of an alternative to fetchmail?

I've been having problems with fetchmail and the attglobal pop server 
(pop4.attglobal.net)

I can get fetchmail to work with other ISP's and have never had any 
trouble with it before.

When I try to receive mail greater than a certain size from the 
attglobal server I get a socket error.
(It works ok though with outlook configured for pop3 retreival from the 
same site)

Regards,

Andrew Eager

-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Question regarding sysv scripts

[Andy Eager]

Jill Rowling wrote:

>How about a background task on the FW machines that checks to see if the
>server is running. If it is running, the FW program goes back to sleep. If
>it is not running (ping or something) then the FW program checks again after
>a set time (in case you have just unplugged the ethernet for a short time)
>after which it shuts the FW down.
>
>I think HA server configurations do something like this only they use a
>serial line separately connected to do the "server up" sensing.
>
>The background task could be either a cron job (messy) or a normal program
>or a daemon.
>
Thanks to everyone for the replies.  I have some good ideas to work with.  

I guess my main problem is that both the firewall machines are POWERED 
from the server box, so I cant really poll the server because if the 
server is shutdown, then the power will be turned off at the end of the 
shutdown (bringing both the firewall boxes down with it with a crash!) - 
Maybe I need to disable the power off feature on the server box.

Andy

>
>
>Cheers,
>
>Jill.
>



-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] Question regarding sysv scripts

[Andy Eager]

Hi all,

I have an interesting question regarding SysV startup / shutdown scripts.

One machine 'the server' has two smaller machines (firewalls) physically 
installed inside it.  They are single board computers that fit into the 
5.25" drive bays on the server.  I have configured them as choke & 
bastion firewalls with iptables: - No problems there.

I want to be able to shutdown the server as well as the the two f/w 
machines at the same time, so I thought about doing this:

a)Have a script in init.d that is linked from rc0.d that uses ssh to 
shutdown each of the f/w machines.  (ssh fw1 halt,   ssh fw2 halt)

b)delay for about a minute

c)continue the shutdown process on the server.

The problem is that ssh wants a password and presumably there is no way 
of providing this from the  console during a shutdown.

Can ssh be configured not to ask for a password ? (I know this is unusual).

Any better way of doing this?

Regards,

Andy
 


-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] tulip driver & RedHat 7.3

[Andy Eager]

Triet Hoai Lai wrote:

>>
>>
>>Anyone out there using the tulip.o module with a DEC-21143 (CNET
>>100/10) NIC ?
>>
>
>Have you tried de4x5.o module?  It is an alternative driver for
>DEC-21x4x and works perfectly with my SMC DEC-21041.
>
>Cheers,
>TL
>
Thanks very much for that.  I didnt know that this driver would work 
with a 21143.  It does quit nicely.

Regards,

Andy


-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] tulip driver & RedHat 7.3

[Andy Eager]

Hi all,

Anyone out there using the tulip.o module with a DEC-21143 (CNET 100/10) 
NIC ?

I found back when I upgraded to RH7.2 that their tulip.o driver was 
broken.  Further investigation pointed me to Donald Becker's driver on 
www.scyld.com, which I was able to compile and worked OK on kernel 2.4.9-21.

Now that I have upgraded to RH7.3 with kernel 2.4.18-4, I find that the 
tulip.o module is still broken and that compiling DB's tulip.o module 
doesnt work anymore (The machine just hangs as soon as I try to insmod 
it from single user mode).

Any similar experiences?

Another question:  Can I get another tulip.o module from another distro 
(using the same kernel version) and use that driver?
(In other words, is the kernel the same accross different distro's with 
the same version?)

Regards,

Andrew E.

-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Mozilla mail & IMAP

[Andy Eager]

Peter Rundle wrote:

>>
>>
>>> Anyone know anything about a broken imap in mozilla-mail-0.9.9?
>>>
>>>   
>>
>
>  Dunno but I'm using
>
>
>Mozilla 1.0 Release Candidate 2
>  
>
>Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0rc2) Gecko/20020510
>
> and imap works fine. 

What imap server are you using?  I'm using courier-imap.  (imap-2001a-10)
I tried Mozilla 1.0 Release Candidate 3 and found no difference.

>
>
> HTH
>
> Pete
>
>
>



-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Mozilla mail & IMAP

[Andy Eager]

Matthew Hannigan wrote:

> I found at least one new imap bug on mozilla's bugzilla:
>
> http://bugzilla.mozilla.org/show_bug.cgi?id=90494
>
> Apparently depends on the imap server you're using.
>
> Do you know which it is?
>
Thanks for that.  I am using courier imap and that bug seems to be it. 
 Guess I will stick with the older version for now.




-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Mozilla mail & IMAP

[Andy Eager]

Tony Green wrote:
1022470275.1138.77.camel@cavey">
  On Mon, 2002-05-27 at 12:59, Andy wrote:
  
Hi all,Anyone know anything about a broken imap in mozilla-mail-0.9.9?I just upgraded from RH7.2 to 7.3 and now my mozilla-mail doesnt work with imap.  It certainly used to!!I use courier-imap and I know it is working since I can read mail from M$-Outlook configured for read with imap

Can you give us an idea what isn't working.  Any errors etc?

I can now that I put mozilla-0.9.2 back!!

It seems as though ver 0.9.9 doesnt even attempt to connnect to the imap
server.
My /etc/xinetd.d/imap file has logging enabled for both success and failure.
 I cant see anything when mozilla-0.9.9 attempts to read mail via the imap
server.  

I replace mozilla0.9.9 with 0.9.2 and everything works again.

I also noticed that I wasn't able to send mail either.  I use procmail on
server mail.aes and the same 'Unable to connect to mail.aes" message appears.
 I did try replacing with an IP address, but that didn't work either.

In short, I reckon 0.9.9 is just stuffed!

Any similar experiences ?

Andy

Re: [SLUG] Re: ask "core"

[Andy Eager]

Angus Lees wrote:

>At Wed, 15 May 2002 18:52:03 +0800, henry wrote:
>
>>I often see "core dumped"  or "gdb -core=/usr/file"
>> but there are lots of machine-code when I read the file(core) from editor-Joe
>>Could someone share his experience  or give some clues?
>>What on earth it is for ?
>>
>
>debugging. its a copy of what was in memory at the time the core was
>generated (usually in response to a segfault).
>
>for example, do:
>
> gdb /path/to/core /usr/bin/executable_that_generated_the_corefile
>
slight correction:
   gdb -c /path_to_core /path_to_exe_that_generated_the_corefile

>
>
>and type "where" at the gdb prompt.
>
Assumes you have compiled with the -g or -gdb option to be of much use.

Andy


-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] Netscape & font problems

[Andy Eager]

Hi all,

Does anyone know how to fix that (longstanding) problem with netscape 
rendering fonts ?

I often get the message:

Warning: Cannot convert string "-*-helvetica-.-iso8859-1" to 
type FontStruct

I had a look with google and a cursory glance at the Font-Howto but it 
didn't tell me much other than how to make the font larger (so its at 
least readable) using the .Xdefaults file with the following:

   Netscape*documentFonts.sizeIncrement: 20
   Netscape*documentFonts.xResolution*iso-8859-1: 100
   Netscape*documentFonts.yResolution*iso-8859-1: 100

Any ideas ?

Regards,

Andy Eager.




-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] xcdroast

[Andy Eager]

Hi all,

Got a strange thing happening with x-cdroast.  A system with both a USB 
cd-writer and IDE reader does the following:

When Reading from the usb device, the sector size reported is 2048

When reading from the IDE device  the sector size reported is 2352 
 (audio cd??)
(Though it sees the CD as data in the type field of the 'CD-Info' window)

On a different system with IDE reader and RW, there is no problem and 
the sector size is reported as 2048 as expected.

The log I get is as follows:  (Note the difference in sectors - same cd 
in different device)

(IDE CD-ROM using sg drivers)
Tue Apr 30 12:38:36 2002 XCDR 0.98alpha9: Read data track 
/tmp/cdroast/track-01.img
Tue Apr 30 12:38:36 2002 XCDR 0.98alpha9: Executing: 
/usr/lib/xcdroast-0.98/bin/xcdrwrap READCD dev= "0,0,0" sectors=0-95791 
-s retries=32 f= "/tmp/cdroast/track-01.img"


(USB CD-RW using usb/sg drivers)
Tue Apr 30 12:44:35 2002 XCDR 0.98alpha9: Reading CD with toc file 
/tmp/cdroast/track.toc (bulk)
Tue Apr 30 12:44:35 2002 XCDR 0.98alpha9: Read data track 
/tmp/cdroast/track-01.img
Tue Apr 30 12:44:35 2002 XCDR 0.98alpha9: Executing: 
/usr/lib/xcdroast-0.98/bin/xcdrwrap READCD dev= "1,0,0" sectors=0-95804 
-s retries=32 f= "/tmp/cdroast/track-01.img"

Since the sector size is wrong, the image it makes is far too large (for 
a data cd).

Any ideas ?

Regards,

Andy

-- 
SLUG - Sydney Linux User's Group - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Help with fdisk please

[Andy Eager]

Dont have your character set, so cant reply to you

>I am running out of space on one of my machines,  I *think* I have more
>room on this hard drive, but I havent used fdisk much and I don't want to
>lose any data.  Can someone
>please tell me if I have any room left that I can make another partition
>with.  Do I have some space from Block 633(4?) to 789?  If indeed I have some 
>room left, how do I go about makingthis partition.
>
>Thanks for the help
>
>Tony
>
>
>[root@localhost /root]# df -h
>FilesystemSize  Used Avail Use% Mounted on
>/dev/hda7 1.2G  835M  286M  75% /
>/dev/hda5  53M  4.9M   45M  10% /boot
>/dev/hda6 988M  540M  398M  58% /home
>/dev/hda1 1.6G  1.5G  157M  91% /mnt/win
>[root@localhost /root]# fdisk /dev/hda
>
>Command (m for help): p
>
>Disk /dev/hda: 255 heads, 63 sectors, 789 cylinders
>Units = cylinders of 16065 * 512 bytes
>
>   Device BootStart   EndBlocks   Id  System
>/dev/hda1   * 1   204   1638598+   b  Win95 FAT32
>/dev/hda2   205   217104422+  82  Linux swap
>/dev/hda3   218   789   45945905  Extended
>/dev/hda5   *   218   224 56196   83  Linux
>
Space here between 225 and 352

>
>/dev/hda6   353   480   1028128+  83  Linux
>/dev/hda7   481   633   1228941   83  Linux
>
Space here as well between 634 and 789

>
>
>Command (m for help): 
>
To do this:
a) Start fdisk
b) Select 'n' for new partition
c) Select logical partition
d) Enter start & end cylinders as requested
e) Select 'p' to display (as you did above) and take note of the new 
partition numbers
f) Select 'w' for write.

(You may have to reboot after this)

then type mkfs /dev/hdaX, where X is the new partition number (make sure 
you get this right)

Create your new mount point  (eg mkdir /newdisk
then you can mount the partition:  mount /dev/hda/   /newdisk

Regards,

Andy E.

>
>
>_
>Do You Yahoo!?
>Get your free @yahoo.com address at http://mail.yahoo.com
>



-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Deleting folders

[Andy Eager]

Didn't know about the -d option to rm!
I'd be using rmdir to remove directories though.

Andy

Marko wrote:

>Hi Guys,
>
>I'm trying to delete a folder in a GNU bash shell, it's goes like this
>
># rm -d filename.d
>rm: remove directory 'filename.d'?   y
>rm: cannot unlink 'filename.d': Is a directory
>
>is there another option?
>
>thanks
>Marko
>



-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] HELP: local filesystems not mounting

[Andy Eager]

Andy Eager wrote:

> Hi all, 

Just worked it out... Writing the post actually made me go back and 
reread (again) my fstab.

>
> I loaded another system with /usr on another partition just to see 
> what it does.  It works fine and the fstab entries are the same on 
> both machines.  (the mount/check options are 1 1 for / and 1 2 for 
> both /boot and /usr) 

   But there was another filesystem on the same physical drive as /usr 
with options 1 1!!   (Bugger)

Regards,

Andy


-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] HELP: local filesystems not mounting

[Andy Eager]

Hi all,

I've just tried to upgrade a system from RH6.2 to 7.2.  Done it before 
from 7.1 --> 7.2 without any probs but this time its been a real bugger.

First I had to move my /usr to another partition to make room for 7.2. 
 (running out of space on / ).  I did this OK and checked with a reboot 
of 6.2 before doing the upgrade.

After the upgrade, I found during boot that the system just hung at 
"Checking file systems".  I get out of that by typing Ctrl-C, but then 
it fails to mount my other local filesystems (namely /usr)

Log messages are:
  

rc.sysinit: Checking root succeeded...
rc.sysinit: Remounting root filesystem as read/write
rc.sysinit: Finding module dependencies
updfstab : execvp: No such file or directory
kudzu:  Updating /etc/fstab failed
kudzu:  Failed
sysctl:   net.ipv4.ip4.ip.forward = 1



I loaded another system with /usr on another partition just to see what 
it does.  It works fine and the fstab entries are the same on both 
machines.  (the mount/check options are 1 1 for / and 1 2 for both /boot 
and /usr)

Any ideas, I've been tearing my hair out for the past 6 hours trying to 
work it out!!

Also, my tulip module which used to work nicely, now seems not to (no 
network, the autonegotiation seems to be using 100MB, and I have a 10MB 
hub - the options for tulip.o seem to be ignored by RedHat.

Thanks lots for anyone that can help.

Andy E.

-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] stopping a module from autoloading

[Andy Eager]

Hi all,

Is there any way to stop a module from autoloading ?

I am trying to get xcdroast to use my ide cd drive as the reader.  When 
I put a cd into the IDE drive, it wants to mount the drive.  (and it 
does!).  When I try to rmmod ide-cd, it just goes and puts it back 
again.  Got so sick of it that I physically renamed the ide-cd.o to 
something else just so I could modprobe ide-scsi for xcdroast.  

There must be a better way!!

Thanks,

Andy E.

(Hope to get to slug tonight, I'm on baby sitting duty so I might have 
to bring my kids along for their first slug meeting - dont worry they'll 
be quiet I've got a suitable bribe lined up for afterwards)



-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] Installing Linux from USB-CDROM

[Andy Eager]

Hi all,

Has anyone ever installed Linux from a USB CDROM?

I have a machine that supports USB CDROM boot which I have duly selected.
The machine boots OK from the USB CDROM but then fails to recognise 
itself (using RedHat 7.2 install disks).  It seems to be registering the 
usb device ok and installs the scsi emulator, but then fails to 
recognise the very media from which it booted!!

Messages I get are:

Initialising USB mass storage driver
usb.c : registered new driver usb-storage
scsi0 : SCSI emulation for USB mass storage devices

   a five second delay  

scsi : device set offline - not ready or command retry failed after bus 
reset : host 0 channel 0 id 0 lun 0
WARNING: USB mass storage data integrity not assured
USB mass storage device found at 2
USB mass storage support registered

Any ideas anyone?

Thanks,

Andy

-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] ask summary of installed programs

[Andy Eager]

Just so happens I am working on a script to do a similar thing for the 
LPIC course currently being undertaken at Granville TAFE.

rpm -qai | egrep "^Name|^Size" | sed "s/  */ /g" | awk 'BEGIN {FS=" "} \
$1 == "Name" && $2 == ":" { printf ("%s: %-40s", $1, $3) } \
$1 == "Size" && $2 == ":" { printf ("%s: %s\n", $1, $3) }' \
| sort -k 4 -n  > packlist

-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Laptop + Debian + no way to install it (help?).

[Andy Eager]

Grant Parnell wrote:

>On Tue, 5 Mar 2002, Ryan Verner wrote:
>
>>Gidday...
>>
>>I'm wanting to configure my Toshiba Portege 3110CT to dual boot
>>debian/win2k.  At the moment its running XP (very, very bloddy slowly), and
>>I have two partitions - a 4.5gig fat32 (c:), and a 1.2gig fat32 (d:).
>>Basically, I want to install win2k to the 4.5gig, and erase the existing
>>1.2gig fat32 and turn this into linux native/swap partitions.
>>
>>Under normal circumstances, this would be damn easy.  However, the portege's
>>do not have internal cdrom or floppy drives.  I have no floppy drive, and no
>>way to access one, and the cdrom drive is external (pcmcia), and even though
>>I know linux does support it, the toshiba bios does not allow me to boot
>>from this.  This is the latest flashed firmware.
>>
>>The main issue is actually getting linux onto the machine.  Somehow I need
>>to invoke the debian install program.  I can't drop to DOS, as I can't think
>>of any way to do this, so loadlin isn't an option either.
>>
You should be able to get a dos prompt by running command.com.
(Run --> c:\command.com)  (I just hate that backslash!!!)

>
>I haven't had to use Windows for years but I'd be surprised/amazed if they 
>finally got rid of the dos command prompt. 
>
Yep, I think they have.

Andy


-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] How this for a simple recommendation for IP-Tables firewalling and ICS

[Andy Eager]

Jeff Waugh wrote:

>
>
>>OK, I'm prepared to be knocked down in flames here (only if your gentle)
>>
>
>You want to be flamed for the consultant's suggestion? :)
>
No, but I backed it up with my own research, therefore it makes it mine 
as well!!!

>>but recently I paid a reasonably well known Linux consultant to advise me
>>on a job I'm doing for a paying customer to install a firewall (that has
>>to do masquerading of all well known services).  His/her sugestion was to
>>stick with ipchains for the time being (preferably under 2.2 kernel
>>because the helper modules were already written and known to work)
>>
>
>His/her suggestion was absolutely correct (if you need to do masq on all
>well known services). Generally, for everyday business firewalls, you don't
>have to worry... How often do you find that h323 is an absolute requirement
>for a small business?
>
>- Jeff
>
Jeff, I absolutley respect your opinion, after all you pointed me in the 
direction of postfix in all its glory and I owe a great many beeers 
for that and a great many emails on other things as well.  However, I 
dont know that H323 (as esoteric as it may be) is what we were talking 
about here.  Masquerading of _WELL KNOWN SERVICES_ like telnet, 
realaudio, ssh etc is the key point.  

As far as I can see, NAT (a particular case of masquerading) has a 
limitation with these well known services when compared with ipchains. 
 The old helper tasks handle connection tracking (which is a real 
bastard) as well as NAT... Particularly if you need to masquerade 
through both bastion and choke firewalls.

If I'm missing something someone please let me know.

Yours enquiringly,

Andy E.



-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] How this for a simple recommendation for IP-Tables firewalling and ICS

[Andy Eager]

DaZZa wrote:
[EMAIL PROTECTED]">
  On Tue, 5 Mar 2002, Andy Eager wrote:
  
Certainly pretty good as far as a basic explanation goes, problem isthat masquerading is not yet up to the level of ipchains and thats whatmost people want. (One IP address, masqueraded to many machines for usewith ftp, realaudio etc).  I still reckon that ipchains with a 2.2kernel is still the simplest and most generally accepted way to dofirewalling if you want particular services masqueraded.

I'm interested to know your reasoning here.

OK, I'm prepared to be knocked down in flames here (only if your gentle),
but recently I paid a reasonably well known Linux consultant to advise me
on a job I'm doing for a paying customer to install a firewall (that has
to do masquerading of all well known services).  His/her sugestion was to
stick with ipchains for the time being (preferably under 2.2 kernel because
the helper modules were already written and known to work)  After some of
my own research, I came up with the following from the man (rusty) himself:

 494: Altering the Destination of Locally-Generated
Connections
 495: 
1.15  rusty  496: The NAT code allows you to insert
DNAT rules in the OUTPUT chain,
 497: but this is not fully supported in 2.4 (it
can be, but it requires a
 498: new configuration option, some testing, and
a fair bit of coding, so
 499: unless someone contracts Rusty to write it,
I wouldn't expect it
 500: soon).
 501: 
 502: The current limitation is that you
can only change the destination
1.16  rusty  503: to the local machine (e.g. `j DNAT --to 127.0.0.1'),
not to any other
1.15  rusty  504: machine, otherwise the replies won't be translated
correctly.
1.1   rusty  505: 
 506: Special Protocols
 507: 
 508: Some protocols do not like being NAT'ed. 
For each of these
 509: protocols, two extensions must be written;
one for the connection
 510: tracking of the protocol, and one for the actual
NAT.
 511: 
 512: Inside the netfilter distribution,
there are currently modules for
 513: ftp: ip_conntrack_ftp.o and ip_nat_ftp.o. 
If you insmod these into
 514: your kernel (or you compile them in permanently),
then doing any kind
 515: of NAT on ftp connections should work.  If
you don't, then you can
 516: only use passive ftp, and even that might not
work reliably if you're
 517: doing more than simple Source NAT.
[EMAIL PROTECTED]">
  What, exactly, doesn't work under iptables?I have a 2.4 kernel running iptables, and it seems to do everything fine -telnet, ssh, ftp, ICQ, irc, real audio, http, https - I haven't foundanything yet that _doesn't_ work.
  
Are you masquerading all of these (both passive & active ftp?).  I would
be keen to hear that there _are_ conntrack modules for the other protocols
you mentioned.  
  
I am more than happy to learn, so please don't hammer me with lots of hurtful
flames if I've missunderstood something here.  
(I'm basically a very sensitive person on the inside!! make love not war!!)
  [EMAIL PROTECTED]">
So, what am I missing?DaZZa

Re: [SLUG] How this for a simple recommendation for IP-Tables firewalling and ICS

[Andy Eager]

Certainly pretty good as far as a basic explanation goes, problem is 
that masquerading is not yet up to the level of ipchains and thats what 
most people want. (One IP address, masqueraded to many machines for use 
with ftp, realaudio etc).  I still reckon that ipchains with a 2.2 
kernel is still the simplest and most generally accepted way to do 
firewalling if you want particular services masqueraded.



Christopher Booth wrote:

>http://www.linuxnewbie.org/nhf/intel/security/iptables_basics.html
>
>Any major flaws in this, anything important missed ?
>
>Is it worthwhile for recommending to somebody ?
>
>Chris
>



-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] Omni & ghostscript

[Andy Eager]

Hi all,

I've just upgraded a few of my boxes to RH7.2 and applied the updates 
provided on EverythingLinux's errata CD's.

I found that the rpm freshen wanted Omni as a required rpm in order to 
upgrade ghostscript.  Omni wasnt on the errata CD or on the RH7.2 
install CD's, so I downloaded it and everything went OK.

Does anyone know if Omni is truly needed ? If so is it just something 
that was left of the errata CD?  Or have I overlooked something obvious 
and loaded (yet another) package on my allready burdened machine.

Regards,

Andrew Eager

-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] mime types with gnome

[Andy Eager]

Indeed I do.  But do I have to use nautilus?  It should work with gmc 
to! (where does gmc get its mime types from:- its so much quicker)

Andy


Jeff Waugh wrote:

>
>
>>When I double click a pdf file (I use gmc) I get an error message:
>>
>
>  ^ you may find this easier with Nautilus.
>
>- Jeff
>



-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] mime types with gnome

[Andy Eager]

Hi all,

Has anyone got any experience with the gnome mime types?  I've never 
been able to get the application editor to work and wonder if its just 
me or a long standing problem with gnome.

I have installed Acrobat for viewing pdf files, and it works fine if I 
open Acrobat and then open the file.

When I double click a pdf file (I use gmc) I get an error message:
   
"Unable to open . with the command:
" xpdf %f"
" To fix this bring up the mime properties editor in the gnome control 
center and edit the default open action for application/pdf"

Well I have done this (a hundred times) and set it up to point to 
Acrobat, but it still tries to use xpdf.
(even when I reset to defaults, ggv is replaced as the default 
application but it still tries to use xpdf)

Any ideas?

Regards,

Andy

-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Toshiba laptop & X

[Andy Eager]

Doing my usual trick of replying to myself so some other poor bastard 
doesn't spend hours trying to fix the same problem.
The problem below **can** be fixed by using the frame buffer device (or 
Option "noaccel") but the result is a very slow X server.

Do the following on a Toshiba 2520 CDS:
Xconfigurator   --preferxf4   --expert
After autoprobing, select S3 ViRGE/MX from the list of cards
Choose "Generic Laptop Display Panel 800x600"  (1024x768 does not work)
   
The rest is straight forward.

Andy

>>
>> Andy Eager wrote:
>>
>>> Hi all,
>>>
>>> I have been tearing my hair out trying to get X working correctly on 
>>> my Toshiba 2520 laptop.
>>>
>>> I have tried a multitude of different video timings and have come to 
>>> the conclusion that the problem does not lie within video timings.  
>>> The generic laptop (800 x 600) setting behaves in exactly the same 
>>> way as the standard VGA (640 x 480) setting.
>>>
>>> The problem is as follows;
>>>
>>> X starts up and seems to work OK.  Whenever a window is moved around 
>>> the screen (or text scrolls) a whole bunch of horizontal lines 
>>> appear on the screen.  There seems to be some kind of pattern when 
>>> this happens:  I can see the standard icons appear at regular spaces 
>>> accross the screen (not clearly but they are there).  For example 
>>> the RedHat icon appears only once when the screen is still, but when 
>>> a window is moved, it appears 'ghost like' another 2 times at equal 
>>> intervals across the screen.
>>>
>>> I had a look at google and found only one reference to a similar 
>>> problem:  Here the author states that the frame buffer device seemed 
>>> to fix the problem.
>>> Anybody got any ideas why this would work (haven't tried it yet) and 
>>> whats going on in the first place?
>>>
>>> Thanks
>>>
>>> Andy.
>>>
>>>
>>
>>
>
>
>



-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Toshiba laptop & X

[Andy Eager]

Henry T Wijaya wrote:

> What's the chipset of your video card in your laptop? 

Its an S3 Inc. | Virge/MX (autoprobed by Xconfigurator) and Win-98 
reports the same
The X server used is XF86_SVGA and XFree4 driver reports: (default).

I tried running Xconfigurator --preferxf3, but it made no difference.

>
>
> Andy Eager wrote:
>
>> Hi all,
>>
>> I have been tearing my hair out trying to get X working correctly on 
>> my Toshiba 2520 laptop.
>>
>> I have tried a multitude of different video timings and have come to 
>> the conclusion that the problem does not lie within video timings.  
>> The generic laptop (800 x 600) setting behaves in exactly the same 
>> way as the standard VGA (640 x 480) setting.
>>
>> The problem is as follows;
>>
>> X starts up and seems to work OK.  Whenever a window is moved around 
>> the screen (or text scrolls) a whole bunch of horizontal lines appear 
>> on the screen.  There seems to be some kind of pattern when this 
>> happens:  I can see the standard icons appear at regular spaces 
>> accross the screen (not clearly but they are there).  For example the 
>> RedHat icon appears only once when the screen is still, but when a 
>> window is moved, it appears 'ghost like' another 2 times at equal 
>> intervals across the screen.
>>
>> I had a look at google and found only one reference to a similar 
>> problem:  Here the author states that the frame buffer device seemed 
>> to fix the problem.
>> Anybody got any ideas why this would work (haven't tried it yet) and 
>> whats going on in the first place?
>>
>> Thanks
>>
>> Andy.
>>
>>
>
>



-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] Toshiba laptop & X

[Andy Eager]

Hi all,

I have been tearing my hair out trying to get X working correctly on my 
Toshiba 2520 laptop.

I have tried a multitude of different video timings and have come to the 
conclusion that the problem does not lie within video timings.  The 
generic laptop (800 x 600) setting behaves in exactly the same way as 
the standard VGA (640 x 480) setting.

The problem is as follows;

X starts up and seems to work OK.  Whenever a window is moved around the 
screen (or text scrolls) a whole bunch of horizontal lines appear on the 
screen.  There seems to be some kind of pattern when this happens:  I 
can see the standard icons appear at regular spaces accross the screen 
(not clearly but they are there).  For example the RedHat icon appears 
only once when the screen is still, but when a window is moved, it 
appears 'ghost like' another 2 times at equal intervals across the screen.

I had a look at google and found only one reference to a similar 
problem:  Here the author states that the frame buffer device seemed to 
fix the problem.
Anybody got any ideas why this would work (haven't tried it yet) and 
whats going on in the first place?

Thanks

Andy.


-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] X setup

[Andy Eager]

Can anyone remember the name of that damn X utility that allows you to 
setup the modeline parameters?
(Not xf86config).

A nice little gui that lets me move the screen up/down, left/right, 
wider/narrrower, shorter/taller and then prints the video timings?


(I hate it when I've used a utility before but cant remember its name.)



Thanks,

Andy.

-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] help for script

[Andy Eager]

Bernhard Lüder wrote:

>Hi,
>
>I need to convert 2 lines of data to one.
>
>This is the data I can extract from the data base:
>FIXEDdata1,Flexibledata1,Neededdata1
>FIXEDdata1,Flexibledata2,Flexibledata3,Flexibledata4,Flexibledata5,Flexibled
>ata6,Flexibledata7,Neededdata2,Neededdata3,Flexibledata8
>
>This is the data I want to use in a comma separated from:
>Neededdata1,Neededdata2,Neededdata3
>
>How can I strip off the FIXEDdata and the Flexibledata so I only keep the
>Neededdata?
>
>There is always multiple double lines of data with some single lines, like:
>FIXEDdata1,Flexibledata1,Neededdata1
>FIXEDdata1,Flexibledata1,Neededdata1
>FIXEDdata1,Flexibledata1,Neededdata1
>FIXEDdata1,Flexibledata2,Flexibledata3,Flexibledata4,Flexibledata5,Flexibled
>ata6,Flexibledata7,Neededdata2,Neededdata3,Flexibledata8
>FIXEDdata1,Flexibledata1,Neededdata1
>FIXEDdata1,Flexibledata2,Flexibledata3,Flexibledata4,Flexibledata5,Flexibled
>ata6,Flexibledata7,Neededdata2,Neededdata3,Flexibledata8
>FIXEDdata1,Flexibledata1,Neededdata1
>FIXEDdata1,Flexibledata2,Flexibledata3,Flexibledata4,Flexibledata5,Flexibled
>ata6,Flexibledata7,Neededdata2,Neededdata3,Flexibledata8
>.
>

Use awk:

awk 'BEGIN {FS=","} \
NF == 3 { printf $3 ","; while (NF == 3) getline } \
NF == 10 { print $8 "," $9 }' inputfile.txt




-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] help for script

[Andy Eager]

Bernhard Lüder wrote:

>Hi,
>
>I need to convert 2 lines of data to one.
>
>This is the data I can extract from the data base:
>FIXEDdata1,Flexibledata1,Neededdata1
>FIXEDdata1,Flexibledata2,Flexibledata3,Flexibledata4,Flexibledata5,Flexibled
>ata6,Flexibledata7,Neededdata2,Neededdata3,Flexibledata8
>
>This is the data I want to use in a comma separated from:
>Neededdata1,Neededdata2,Neededdata3
>
>How can I strip off the FIXEDdata and the Flexibledata so I only keep the
>Neededdata?
>
>There is always multiple double lines of data with some single lines, like:
>FIXEDdata1,Flexibledata1,Neededdata1
>FIXEDdata1,Flexibledata1,Neededdata1
>FIXEDdata1,Flexibledata1,Neededdata1
>FIXEDdata1,Flexibledata2,Flexibledata3,Flexibledata4,Flexibledata5,Flexibled
>ata6,Flexibledata7,Neededdata2,Neededdata3,Flexibledata8
>FIXEDdata1,Flexibledata1,Neededdata1
>FIXEDdata1,Flexibledata2,Flexibledata3,Flexibledata4,Flexibledata5,Flexibled
>ata6,Flexibledata7,Neededdata2,Neededdata3,Flexibledata8
>FIXEDdata1,Flexibledata1,Neededdata1
>FIXEDdata1,Flexibledata2,Flexibledata3,Flexibledata4,Flexibledata5,Flexibled
>ata6,Flexibledata7,Neededdata2,Neededdata3,Flexibledata8
>.
>

Step 1: - Make sure the short lines are unique.

Step 2:  Use the following substitution patterns with sed or egrep:

For the 'short' lines with only 3 bits of data:
s/\(,[a-zA-Z]*$\)/\1/

For the longer lines:
s/\(,[a-zA-Z]*\)\(,[a-zA-Z]*\),[a-zA-Z]*$/\1\2/

(Untested but should work OK)

Regards,

Andrew E.

-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] insmod a module permanently

[Andy Eager]

Hi all,

Anyone know how to insmod a module without the 'autoclean'.
(I just want to load it permanently)

I had a look at insmod, but the only option I see a -k option which 
turns on autoclean but not one to disable it.

Regards,

Andrew E.


-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] Terminal emulation

[Andy Eager]

Hi all,

Does anyone know of any terminal emulation packages (other than xterm) ?

I've got a QNX 2 network and can only use telnet to get into it. 
 Naturally enough, when I log into QNX, xterm will run OK in VT102 mode 
but I have to tell qnx to talk to me in VT102 mode and in doing that I 
lose a lot of the keyboard functionality (arrow keys for a start).

I guess what I am looking for is some kind of 'generic' terminal 
emulator that will let me plug in escape sequences for the screen as 
well as for the keyboard.

Any help greatly appreciated.

Thanks,

Andrew E.


-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] Tracing a shutdown

[Andy Eager]

Hi all,

Is it possible to trace a shutdown ?

I'm finding that when shutting down (or rebooting) my system hangs for 
about 20-30 seconds while doing each of the following:
 
  shutting down interface eth0
  shutting down kernel logger
  shutting down system logger

I tried putting a "set -v" in each of the /etc/rc.d/init.d/network and 
syslog scripts, but that doesn't seem to tell me much.  (They each 
execute to completion and then hang).  This happened after doing an 
upgrading rh6.2 from the ximian site using the Redhat channel

Any help much appreciated.

Regards,

Andrew E.


-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Thanks & triple boot question

[Andy Eager]

Adam Vaughan wrote:

>
>Redhat 7.1 is on a partition when I installed it I gave it it's own
>/boot and /swap.  the /boot and /swap I made big enough to support 4
>linux version.
>
>Redhat 7.0 is on it's own partition, sharing the /swap & /boot .  I
>didn't think it would do this but because there was already a linux
>native /swap and /boot it didn't ask me to create one in the install.
>
>By the line I mean the:
>image=/boot/vmlinuz-2.2.16-22
>root=/dev/hda7  (7 is the partition with Redhat 7.0)
>What what I put in there instead?
>
>Thanks,
>Adam.
>
Your lilo should look something like this:

...  Common stuff up top

image = /boot/vmlinuz-2.2.16-22
  label = linux7.0
  root = /dev/hda7

image = /boot/vmlinuz-2.4.2-2(assuming you are running the kernel 
that came with the disk)
  label = linux7.1
  root = /dev/hdaX

The X will be whatever partition you installed RH7.1 on.  If it was the 
same partition as RH7.0 then that line will be /dev/hda7.  To check, you 
can boot up into RH7.0 and run fdisk to see
(fdisk /dev/hda, then press p to print the partitions.  The linux native 
types are 83.  If you have two type 83 partitions then presumably one of 
them will be hda7.  You want the other one (probably hda8 if it was 
installed after 7.0)

Regards,

Andrew E.


-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] [OT] Someone trying to hack me

[Andy Eager]

DJ! wrote:

>- Original Message -
>From: Erik de Castro Lopo <[EMAIL PROTECTED]>
>Date: Wednesday, September 19, 2001 7:59 am
>Subject: Re: [SLUG] [OT] Someone trying to hack me
>
>>Anybody who connects a windows box directly to the net is asking for
>>trouble, even if its just on a dial up modem.
>>
>
>What a stupid fscking remark. I'm not even going to dignify your 
>stupidity with further comment.
>
>DJ!
>
I don't think the remark was that stupid, though it probably should have 
been expanded to any box rather than just a windows box.

 I used to use Win-NT directly connected to the net and sure enough it 
got hit.  Similarly I used Linux connected directly to the net and again 
it got hit.  Now I use a combination  and an ipchains firewall.  I 
haven't been hit since.

Like the cat says... Its only a matter of time...

Andrew E.


>


-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Samba...

[Andy Eager]

Phillipus Gunawan wrote:

> I'm really sorry because was late to read these email(s)
>
>  
>
> Without harming everyone else, yes, most of the time, for us -newbies- 
> is a bit difficult to find a good doc, even sometime can't think the 
> proper question to post it here (I did it a few times...)
>
>  
>
> I was trying to set up samba with NIS and create local user in the box 
> as the authentication to the window$ (for a start before I really have 
> to deal with pam)
>
>  
>
> This is how I create the user:
>
>  
>
> [1] useradd -g 100 -d /dev/null -c mywidow -m -s /bin/false widowbox
>
>  
>
> [2] smbpasswd -a -m widowbox
>
>  
>
> On the widow machine, I try to join the domain, It give me an error 
> saying that no user name or bad password.
>
> Anyone have a -bit- clue?
>
Do you have a trust account setup on your linux box ?  Win-nt needs this 
for it to do any kind of remote (network) authentication.  In your 
passwd file you will need the following: (note the '$' sign after the 
name of the host.

   NODE2$:*:1000:900:Trust Account:/dev/null:/dev/null

You will also need to use the -m option (as you did above) with 
smbpasswd when importing the trust account.

You probably allready know, but the smb-howto has some info on this.

Regards,

Andrew E.


-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Gnome 1.4 - Mozilla / Abiword bugs

[Andy Eager]

Jeff Waugh wrote:

> 
> 
>> Has anyone downloaded the new version of Gnome 1.4 (the new 1.4 not the 
>> old 1.4 released a while back). 
> 
> 
> Heh, there's (so far) only one GNOME 1.4, but Ximian did upgrade their
> packages recently...

Yeh, I know and thats the problem.  They upgrade their packages  (quite 
considerably ) and still called it 1.4.  I know that's like having 
RedHat 6.2 with a 2.2.19 kernel, but at least in RedHat's case, the 
distro release does give some indication of backward compatability - Not 
the case here I can assure you.

> 
>> I got the latest release from the ximian site last night, and loaded it up.
>> Now Mozilla mail (the funky new bluey / grey look) crashes when you try 
>> to compose an outgoing email (or forward an existing message).
>> 
>> Also Abiword seems to be even less stable and they have removed the 
>> 'classic' gnome look as an option on startup.
> 
> 
> I only use Ximian GNOME at work, and at the moment we're stuck with a
> dial-up connection so haven't upgraded yet. If no one else has tried, we can
> always find out from the Ximian monkeys (try their MonkeyTalk app, or #gnome
> on irc.au.gnome.org).

If you keep a copy of the redcarpet package directory  
(/var/cache/redcarpet) somewhere on your network and then copy it back 
to /var/cache/redcarpet when you do your upgrade, it only takes an hour 
or two (closer to two but not 7 - 8 hours as the full 115 M from scratch 
would take).




-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] Gnome 1.4 - Mozilla / Abiword bugs

[Andy Eager]

Has anyone downloaded the new version of Gnome 1.4 (the new 1.4 not the 
old 1.4 released a while back). 

I thought the first release was pretty good, Mozilla not too slow 
(though slower than ye old Netscape 4.7x), and Abiword only crashed 
every now & then.

I got the latest release from the ximian site last night, and loaded it up.
Now Mozilla mail (the funky new bluey / grey look) crashes when you try 
to compose an outgoing email (or forward an existing message).

Also Abiword seems to be even less stable and they have removed the 
'classic' gnome look as an option on startup.


Anyone having similar problems ?

Regards,

Andrew E.


-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Help I got hacked!!

[Andy Eager]

Thanks for that lovely piece of information.  I guess the key to the 
whole thing is to have a reasonably quick and easy way of installing 
from scratch without having to go through all the headaches associated 
with a 'clean install'.

Fortunately I keep all data backed up, so thats not a problem but having 
built the machine from scratch as a humble newbie only 6 months ago 
( I couldn't even spell Linux...) I've got stuff like nfs, ntp, 
named, samba etc, etc which all just 'evolved' over time. 

Wouldn't a reasonable compromise be to do the following:

verify each installed package:rpm -V -a
(Now we know each package is OK)

for each file in all directories except home, do:   rpm -qif 
(If it doesn't belong to any package then warn user)

The problem I have is that it will probably take me days to rebuild (and 
re-remember) everything I did over the past six months in terms of 
administration.  Then having done that, the same thing happens 
again. (because I still don't know how they got in)


Andrew E.
( A desperate optimist who untill now, believed that all people (even 
hackers) were good )





Jeff Waugh wrote:

> 
> You *need* to take the machine offline, and rebuild or replace it. This is
> the only safe way to deal with it (and not having it come back to bite your
> arse next rainy day).
> 
> - Jeff
> 



-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] Help I got hacked!!

[Andy Eager]

Hi all,

Help...

I've just managed to get myself back online after a few days of 
repairing damage caused by my first hack attack.

I got hit by 'luckroot' which basically changes a whole heap of stuff in 
/etc, /bin and all the usuals.  It seems to be an address / port scanner 
which collects stuff and emails it back to someone somewhere 


I fixed everything (I think except for /dev/hdc, /dev/console and the 
/dev/tty's) 

My question is:  Can I safely force rpm to reinstall the dev package ? 
(without screwing my hard disk settings in the process ?

Also, does anyone know of a service on the net that can attempt to find 
these holes in my firewall / sysconfig without doing damage ?

Thanks,

Andrew E.

PS:   Haven't these hacking  pr??ks got anything better to do with their 
time !!


-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] [OT] Webmail as an ISP

[Andy Eager]

Hi all,

As yet another casualty of one.net,  I was wondering if anyone had any 
experience with webmail as both an ISP and site hoster.

I joined up with optus just to get connected, but couldn't get the site 
redelegated without paying a bomb.

It seems to me that one.net were reasonably unique in that 'personal' 
webspace provided as part of most unlimited plans could be used to hold 
a registered domain name.  (ie www.provider.com.au/~yourname  could be 
used as www.yourdomain.com.au)

Now you need to pay $20-$30 per month to host the site and another 
$20-$30 per month for 'unlimited' access.

Were we just on a good thing ?
Is there someone else out there doing what one.net did ?


Andrew


-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Using fget() with shell line editing

[Andy Eager]

Andrew,

Thanks for your reply.

Nothing wrong with ncurses, but want it to be able to run with normal 
unix i/o redirection  (or piping) as well as from stdin.  Don't want 
ncurses sending anything other than boring old text (maybe even the odd 
line-feed or form-feed).


Andrew Bennetts wrote:

> 
> You'd basically have to look at the termcap/terminfo stuff to handle
> up-arrow properly, I think.
> 
> Just curious, what's so bad about using ncurses?
> 
> -Andrew.
> 



-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] Using fget() with shell line editing

[Andy Eager]

Hi all,

I've got some code which I want to be able to run from a console without 
using ncurses.

Does anyone know if it's possible to have the up/down arrow keys return 
previous line entries (like it does in the shell) using fgets() or the 
like ?

At the moment, I can do all the normal line editing stuff with backspace 
/ del etc but an up arrow gives me ^[[A.  (presumably scan code for 
up-arrow).

Regards,

Andrew E


-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] Dumping a file in hex

[Andy Eager]

Hi all,

Does anybody know how to dump/edit a file in raw (hex) mode ?
   - From within X ?
   - From the command shell ?

I thought there was something like 'hexedit' in Gnome 1.0 but since 
upgrading to 1.4 my 'development' folder seems to have mysteriously 
emptied itself.

Hope this isn't one of those dumb questions.  I get the feeling though 
that for Linux this must be a fairly common thing to do and so it 
probably is..

Thanks

Andrew E


-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Good Editors for C development

[Andy Eager]

So does this mean that there is a linux version of Visual Studio ?

As the man says... Where do ya get it  ??

Andrew E.


MD wrote:

> Yep...
> 
> Visual Studio is a start...
> 
> MRD
> 
> -Original Message-
> 
> Hi all,
> 
> I'm currently using KDE's kwrite as an editor (running on Gnome 1.4 
> which is pretty schmick).
> 
> Does anyone know of any good development environments / editors for 
> writting c-apps with multiple source files.
> 
> kwrite is good (it has various language highlighting) but something with 
> at least multiple windows would be good and perhaps with some other 
> features found in 'other' development environments.
> 
> 
> Regards,
> 
> Andrew E.
> 
> 



-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] Good Editors for C development

[Andy Eager]

Hi all,

I'm currently using KDE's kwrite as an editor (running on Gnome 1.4 
which is pretty schmick).

Does anyone know of any good development environments / editors for 
writting c-apps with multiple source files.

kwrite is good (it has various language highlighting) but something with 
at least multiple windows would be good and perhaps with some other 
features found in 'other' development environments.


Regards,

Andrew E.


-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] FTP / Masquerading

[Andy Eager]

Hi all,

I've been using ipchains both as a firewall and masquerader for other 
machines on my network.

It all works fine except for FTP, which I must run in passive mode 
(pftp) so that the client is doing all port opens. 

Are there any problems using a normal active ftp over a masquerading 
connection ?
(Normal active ftp works only on the machine running ipchains which is 
the machine that the physical link is connected to)

Thanks,

Andrew E.


-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] Where to get a 2nd hand laptop

[Andy Eager]

Hi all,

I saw someone mention a weekend computer market at North Ryde (North 
Rocks ??)

Where abouts is it and when is it on.

I have a friend who saw a cockroach run into his toshiba.   He then 
placed the said laptop into the freezer in an attempt to kill it.

He killed the wrong animal - doh!

Regards,

Andrew E


-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

Re: [SLUG] Linux & viruses

[Andy Eager]

Jamie Wilkinson wrote:

> This one time, at band camp, Andy Eager said:
> 
>>   a)   Periodically, examine all running tasks.
>>   b)   For each task, do an 'rpm --verify' for the package that this 
>> process belongs to.
> 
> 
> How about a virus that renames itself to 'ls', you check the process list
> and verify that ls is in the package database.
> 
If a virus did copy something over ls, it would be caught as follows:

   rpm -qif `which ls` would return the package name fileutils
   rpm --verify fileutils  would show that ls had been modified.

I'm sure there are reasons why my idea is oversimplistic, but this would 
not be one of them.

In addition, it seems to me that since the rpm --verify option (or its 
alternative) is free and not a bad thing to do on executables every now 
and again anyway.


Regards,

Andrew Eager.


-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] Re: Linux & viruses

[Andy Eager]

Marc Lawrence wrote

> 
> Cheers
> Marc (yes I *know* I'm sending this from Outlook - get off my
> case, okay?  I'm trying, that's why I'm on SLUG)

Thanks for the info,  I was also using Outlook until a couple of weeks 
ago.  Set yourself up postfix and imap then you can use pretty much any 
mail reader you want (Mozilla is pretty good) as well as having 
centralisedmail storage.

(Courtesy of Jeff Waugh et al..)

Regards,

Andrew Eager.



-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] Linux & viruses

[Andy Eager]

Hi all,

I've been looking around for any information on virus scanners for 
Linux.  Apart from McAfee, I couldn't see much. 

While I was doing this it occurred to me that it might be possible to 
scan for viruses by doing the following:

   a)   Periodically, examine all running tasks.
   b)   For each task, do an 'rpm --verify' for the package that this 
process belongs to.

Assuming that it is not possible to create a task in Linux without 
creating a process ID entry (and therefore visible in the /proc directory.).

Am I being to simplistic here?
I'm not a virus aware person, so I don't spend a lot of time trying to 
work out the various ways of screwing up a linux box. 

Regards,

Andrew Eager.


-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] IMAP Question

[Andy Eager]

Given imap's apparent lack of configuration (for anything), how does it 
know which directory to get my mail from [ //var/spool/mail//].

Is this one of those programs that needs to be configured in the source 
code ?

Regards,

Andrew E.



-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug

[SLUG] Netscape - Profile managers

[Andy Eager]

Hi all,,

First, Thanks to Jeff and others for info on setting up imap / postfix &
fetchmail.

After an interesting couple of days setting it all up I now have the
following questions regarding multiple users:

a)   Does anyone know if there is a Linux profile manager for Netscape
similar to  version (I can't say the name)

I'v worked out how to map an ISP user name to a local user name (user ..
here is... ) but what if I wan to work with email using different ISP
id's ( ie local user andy wants to send / receive mail with to/from
headers marked as andy1@domain1 as well as email to / from
andy2@domain2).  I know I can do this if I have a NIS user (I use NIS
over a network with samba et al...)  mapped to each ISP user, but that
means logging  off each time or running 2 X servers / gnome sessions.

Any Ideas?

Andrew E
(PS: I'm going to write a nice easy howto about this if someone hasn't
already.  It all looked much harder than it was mainly due to so  much
daunting info about sendmail and so little about postfix)




-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://slug.org.au/lists/listinfo/slug

[SLUG] What pkg contains smtpd

[Andy Eager]

OK,  I give up- where the hell is SMTP daemon kept ?  (ie what rpm
package)

I found imap OK (got RH 6.2) and did a search on Redhat site only to
find 4 matches for smtp (2 of which were fetchmail and two perl script
thingies?

Am I missing something here?  All of the other inet daemons were really
easy to find!!

Andrew E.



-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://slug.org.au/lists/listinfo/slug

Re: [SLUG] Fetchmail setup

[Andy Eager]

Jeff Waugh wrote:

> 
>
> > So I run fetchmail on my server and have clients (netscape and that other one
> > that I'm not allowed to name) read mail from the server using IMAP also ?
>
> Yep; and fetchmail is nice for collecting multiple accounts and delivering
> to specific users, etc.

I had a look at fetchmail and set up a simple .fetchmailrc looking like:

   poll   proto auto user "myname" pass "my_pass"

I used imap in place of 'auto' above but that didn't work.  Pressumably my ISP
doesn't support IMAP.  I noticed that with auto it tried imap first anyway then
POP3 which did work.

When it download a test email, it bombed out after the read saying:

  " SMPT connect to localhost failed"

Does this mean I need to have sendmail running as well ?
Can't it just download the messages on its own and leave them somewhere for me to
read them later ?

> [ The answer is to this again is IMAP, your message store is always on the
> server in that case. Thus you can use any client, and all your mail is in
> one place, easy to backup, etc. ]

When you say server, I assume thats my server not the ISP's .
 I cant see anything in fetchmail that tells me how / where the mail is stored.



-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://slug.org.au/lists/listinfo/slug

Re: [SLUG] Common mail storage Outlook/Netscape

[Andy Eager]

Jeff Waugh wrote:

> > Does anyone know of a way to have Outlook Express & Netscape mail share
> > messages ?  (IE: Have a message read from the POP server by Outlook and
> > readable in both Outlook & Netscape, whithout doing the obvious and
> > forwarding / rereading)
>
> To solve this problem on a *nix platform, have your server run an IMAP
> daemon (one particularly good one is Courier IMAP), and use it instead of
> POP. POP should be left for dead. Faster than uucp.

So I run fetchmail on my server and have clients (netscape and that other one
that I'm not allowed to name) read mail from the server using IMAP also ?

> > Also is there any way of having Outlook Express store its messages on a
> > network filesystem.  I tried setting the location of the message store
> > to my 'Z:' drive but it would have any of it.  Upon checking with the
> > MS site, I found that OE had to use 'local media'.
>
> This is not a Linux question. In fact, these don't really sound like Linux
> questions at all, but as we've come to find out, Free Software can solve
> problems you'd never expect it to.

In fact these were linux questions..  My 'main' linux server uses pppd,
ipchains (firewall & proxy) and a caching name server to serve internet
connections to my 'other' linux boxes (most of which are dual boot), on a full
time basis.

The Z: drive I refered to is served by samba and the netscape I refer to is
being run on linux as I type.

How much more Linux do you want

My main reasons for asking these questions are migration related.  As a newbie
(& recently joined SLUG member) I can't just drop what I know (QNX in great
detail, WINNT just a bit).   In fact having been serious about linux for just
under 6 months,  I reckon I've done OK.  (I've even written a simple kernel
module!!)

So, as your rules say... DONT FLAME THE NEWBIE...

Thanks for your help,


Andrew E.




-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://slug.org.au/lists/listinfo/slug

[SLUG] Linux & Windows Integration

[Andy Eager]

Simon Bryan wrote:Now the problem:

> I want to put all our user directories onto the new 'NetDrive'
> I need to share them for use by Windows
> I want to make them the same home directories as appear in AUC -
currently
> '/home/popusers//username, where  is the year they started in
Year 7
>
> So, am I going to be able to achieve this?
> Should I setup the 'NetDrive' as an NT or Linux drive (it will do
either)?
> When setup how do I get both systems to see the home directories
(pretty OK
> with Windows - map a drive letter and use that in user creation
scripts).
> Pretty sure I would do the same from Linux using Samba, but how to set
the
> correct permissions?
> At the moment on the Windows computer we have
/Users/Students//Username
> as home directories and only have the one share at /User/Students then
use
> permissions on the directories to ensure privacy below that.
> Is this a job for Winbind?
>
> Any suggestions, resources, readings etc appreciated as I have a
couple of
> weeks to sort this all out in my head before I have to implement it.

Samba documentation is huge.
check man pages for :
smb.conf, smbd, nmbd

Have a look at the SAMBA HOWTO.

There is also a browser based configuration tool (swat) - Samba Web
admin tool,
that helps you to create the smb.conf file and provides quick access to
the
various bits of documentation.

Also a book by  McCune called "Integrating Linux & Windows" gives a good
start.

In short, the answer to your question is definately yes.

Hope this helps,

Andrew Eager





-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://slug.org.au/lists/listinfo/slug

[SLUG] SRR message passing for Linux / Linux limitations

[Andy Eager]

For anyone interested,  S/R/R message passing has been implemented in
the kernel and seems to run nearlly as fast under Linux as it does under
QNX.

In fact the only limitation is summarised below by the author and is
exactly the question I was asking in last Friday's SLUG Q&A session (re:
copying data between two non-related processes - quickly!)

Seems to me that if this is a limitation of the Linux kernel, it should
be fixed!  Not an unreasonable thing to ask of a kernal and must be one
of Linux's few limitations.

1) I don't know how, in Linux, to access the address space
 of
 anything but the currently running process. In fact, it is
 not
 generally possible: the process's memory may be swapped to
 disk.
 This forces me to buffer all messages in kernel space
 before
 causing a context switch to the process for which the
 message is
 destined. Unfortunately this means that all data must be
 copied
 from sender to kernel to receiver, whereas under QNX
 message
 passing results in a single copy, directly from the
 senders
 address space to the receivers. Oh well.


The package is available from www.cogent.ca

Regards,


Andrew E.






anything but the currently running process. In fact, it is not
generally possible: the process's memory may be swapped to disk.
This forces me to buffer all messages in kernel space before
causing a context switch to the process for which the message is
destined. Unfortunately this means that all data must be copied
from sender to kernel to receiver, whereas under QNX message
passing results in a single copy, directly from the senders
address space to the receivers. Oh well.



-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://slug.org.au/lists/listinfo/slug

[SLUG] Common mail storage Outlook/Netscape

[Andy Eager]

Hi all,

Since its so quiet, I thought I'd ask...

Does anyone know of a way to have Outlook Express & Netscape mail share
messages ?
(IE: Have a message read from the POP server by Outlook and readable in
both Outlook & Netscape,
whithout doing the obvious and forwarding / rereading)

Also is there any way of having Outlook Express store its messages on a
network filesystem.  I tried setting the location of the message store
to my 'Z:' drive but it would have any of it.  Upon checking with the
MS site, I found that OE had to use 'local media'.

This is surely crazy in this world of shared data...

Thanks,

Andrew E.


-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://slug.org.au/lists/listinfo/slug

Re: [SLUG] Silly ISP type question [ OT ]

[Andy Eager]

Jim Boing wrote:

> Hi all,
>
> Yes it's another one of those annoying "What's the
> best ISP for a Linux user" type questions. Let me
> start by stating that I'm just off the boat so to
> speak and hence have little/no knowledge of the ISP
> situation here.
>
> As I'm a cheap b*d I want to sign up with One Net
> and their unlimited download. Rang their support chaps
> the other day and was told a) they don't support Linux
> and b) they couldn't give me any info (i.e DNS etc)
> coz I wasn't using their CD installer (run away .. run
> away!). Now all the basic info is on the site ... I
> just need to know if anyone here dials in to them, if
> they have any weird and wonderful login requirements
> and if they're any good.

I use One-net and find them particularly good.  My 56K modem with compression
gets me a good 6-8K when connected to local sites.  There's no real trick, just
get  ppp to do the work for you (HOWTO-PPP)

In my system,  I have a 'server' with an  IP-chains firewall and proxy.   The
server is set up so that ppp persistantly dials the one-net server whenever it
kicks me off (every 4-6 hours).  I just use the power switch on the modem to
connect to the ISP every morning and turn it off at night.  (No safer firewall
than turning the modem off).

Also, with  local name-caching you'll get pretty good results:

 So for cheap full time access to the net:

a)Install a second (3rd, 4th...) telephone line
b)Pay Telstra $17 per month for line rental plus 20c per call (at most 6
a day)
c)Pay One-Net  $21.95 a month for unlimited plan

Apart from the Telstra install costs, about $1-$2 a day will get you full time
access.

Regards,

Andrew E.



-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://slug.org.au/lists/listinfo/slug

Re: [SLUG] Name of program for process causing disk activity

[Andy Eager]

> According to the man page sar(1)
>
> " Collect,  report, or save system activity information "
>
> I'm not sure what you needed reminding of since you have the
> names of the three programs I mentioned.
>
> Regards,
> Anand

Regarding sar:

I've got RH6.2 and I don't have it.  After a look at the RH site, I found
it came in the sysstat pkg which looks like only came into existance in
RH7.0 and above.

Is this right ?

Andrew E


-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://slug.org.au/lists/listinfo/slug

Re: [SLUG] Copying memory in kernel

[Andy Eager]

>
> > I know its a bit of a pain and some would say this old dog should learn new
> > IPC tricks (and I have) but after 10 years of code writing using the SRR
> > (send/receive/reply) paradigm, and having never been able to write an application
> > that has brought the system down I need a lot of convincing.  In addition the
> > code written has been designed in a modular fashion very quickly and almost
> > always works 'first go'.
>
> heh, yep I did some work with the L4 microkernel for my eng. thesis, which
> has similar blocking IPC. It certainly ensured clean code with well defined
> interfaces. I was tempted to do a similar IPC module for Linux then too.
> Maybe someone has already done one? Could be interesting ...

There is a package done by FC Software called 'SIMPL' (synchronous interprocess
message passing)
They did their stuff in user-space and based it around fifo's.  It also handled
message passing accross the network which was pretty nice, but it did suffer
performance problems.  (having to copy messages to/from the fifo's while having to use
other code to synchronise the whole thing)

Regards,

Andrew E.



-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://slug.org.au/lists/listinfo/slug

[SLUG] Next SLUG meeting

[Andy Eager]

Hi all,


Can anyone confirm whether there is a SLUG meeting tonight at UTS ?
( I'd like to go but I have an insistant wife who wants to know what
we're doing tonight!!!


Andrew E



-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://slug.org.au/lists/listinfo/slug

Re: [SLUG] startup without login

[Andy Eager]

Mike Brooks wrote:

> Hi, We're installing some SBCs with a stripped Linux installation to
> operate as a realtime monitor. This is interfacedinto an embedded
> system mounted on some industrial machinery. What can be done to avoid
> the login? I'vescratched through some FAQs but they seem to focus on
> the runlevel which doesn't appear to do what I want. Basically, I just
> want to load X and run the application.

If all you want to do is start x and load an application (and you don't
care if the other 6 screens are still available (ALT-CTL-) then
just have inittab (/etc/inittab) start up X (or gdm or xdm).  If you
want to remove the other 6 tty's get rid of the mingetty cmds in
/etc/inittab.  I've never done this and I reckon unless you have a very
good reason to do otherwise, you should leave at least one console tty
for 'emergency-access'.  If anything does go wrong with X, you have no
way into your system.

Regards,

Andrew Eager


-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://slug.org.au/lists/listinfo/slug

Re: [SLUG] Copying memory in kernel

[Andy Eager]

Dear Harry,

Thanks for your ideas, I am an old hand at QNX which used Send/Receive/Reply
message passing.
For anyone familiar with QNX they will need no convincing that this is a very
simple and extremely robust means of IPC that will:

   a)  Force you to write code in a highly modular fashion.
   b)  Is intuitive in that there is no need for the coder to worry about
synchronisation between tasks
   c)  Very good for fault tolerance & easy to debug.
   d)  Very difficult to write code that will deadlock.


> >The problem here is that there are typically many clients (Task A's)  are
> >sending requests
> >to their corresponding server (Task B).
>
> So, do you have a one-to-one correspondence between clients and servers,
> ie for every Task A there's a corresponding Task B, or are there many Task As
> for a given Task B?

No there is no 1:1 correspondance.  There may be many client tasks (Task A) for
each server task (Task B).  In addition, a client task may send to more than one
server task. For example, you might have  3 server tasks (svrA, svrB , svrC) and
say 100 client tasks (clnt001 to clnt100).  Clnt001 may send a message to svrA,
get a reply and some time later send a message to svrB and wait for a different
reply.

> What form do the requests take?  It sounds like maybe message queues
> might be a better choice.  They are specifically designed to allow the passing
> of requests from one process to another.

Yes, I have tried 2 different user space implementations using both message
queues (msgget et al) and shared memory (shmget /shmat etc).  I used the msg Q to
handle synchronisation and pass shared memory handles to the server and shared
memory for the actual transfer. This got over the problem of exceeding the msg
Q's 16K memory limit.   It was quite slow (compared  to QNX on same hardware) and
was not linear with respect to number of clients running.  (Idealy, if one client
ask running on the system and it transfers N bytes per second then 2 simultaneous
client tasks running should each transfer N/2 bytes per second).

There is no restriction on message size, though under QNX2 there was an upper
limit of 64K due to segmentation restrictions.  Most of the stuff I am porting
will live with a 64K limit but it would be nice to remove that restriction if
possible.

> My guess would be that it's not that easy, because the sharing of data like
> this is precisely why shared memory exists.  As I say, though, threads are
> good for this sort of thing, if there's a one-to-one correspondence between
> clients and servers.

I reckon your right!  My first thought was that at least one of the tasks (either
client or server) will not be running at the time of the request and may
therefore have been paged out into swap.  Perhaps I'm being naive, but how about
this for an idea:

a)   Server registers itself with the kernel module (KM) where the KM then:
- does a virt_to_phys() on the receive buffer passed.
- Places the server task in a blocked state waiting to be woken by a
client

b)A client comes along and does a send request to the server.  The KM
then:
- does a virt_to_phys() on the send buffer passed.
- Copies the memory between the two physical address determined.
- wakes up the server task indicating its receive buffer is now
ready.
- Places the sending task in a blocked state waiting for a reply.

The reply follows the same sort of process as above.

The one VERY BIG caveat here is that the 'non-current' task is still in physical
memory at the time of the transfer.  If not then you either get a segmentation
fault or (even worse) you trash some other poor innocent task's data segment.
Thanks to David's previous post, mlock() will get around that.

**

I know its a bit of a pain and some would say this old dog should learn new
IPC tricks (and I have) but after 10 years of code writing using the SRR
(send/receive/reply) paradigm, and having never been able to write an application
that has brought the system down I need a lot of convincing.  In addition the
code written has been designed in a modular fashion very quickly and almost
always works 'first go'.

I think this sort of IPC is worth implementing in Linux (and not just for me!!).
I think if more people knew about it, it would be far more popular than it is,
particularly in the areas of process control / synchronous server apps etc.

Hope I haven't dragged on too much about this and trust it has some interest to
others.

Thanks for your help and ideas.

Andrew E.

(PS: should I be sending only one reply to the slug site, or one to the author
and cc the slug site ? - I am a newbie to this site and SLUG rules say you can't
flame me for asking dumb questions but I don't want to anoy anyone by sending
mail directly)




-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://s