Ed,

Thanks for the reply (and also to the others). I have solved the problem,
though I'll be damned if I understand why...

I have written two shell scripts that I have been using to start and stop
the RedHat PPP setup, created by netcfg. I run Linux on a server that does
not have a screen so I telnet or run an Xterm using Hummingbird from an NT
workstation, where the majority of my work is done. So if I can't be
bothered starting the X Server, I just telnet. Hence the scripts.

These scripts work fine so I thought I would just call these instead of
diddling the /etc/diald/connect script that comes with diald. However, if I
call these scripts, the two minute delay occurs. I'll be darned if I know
why. I haven't got to the bottom of that. I have attached the start script
to this note, just in case someone is curious enough to figure this out.

Given this frustration, I then decided to use the /etc/diald/connect script
and the delay disappeared, but now I have another problem. This, I think, is
a problem with the chat script. When the modem is initialised (which is
simply ATZ), I get a 127 return code from chat and it doesn't proceed any
further. I don't understand why yet. The modem works just fine, as I have
already implied.

My next question: Is there a CHAT reference or guide apart from the man
page? The man page refers to some information in the UUCP docs, but I can't
see exactly where it is. Maybe somebody knows of a web page that explains
this arcane piece of software.

Regards,

Mark Harrison
mailto:[EMAIL PROTECTED]


> -----Original Message-----
> From: Ed Grimm [mailto:[EMAIL PROTECTED]]
> Sent: Tuesday, 8 December 1998 4:29
> To: Mark Harrison
> Cc: [EMAIL PROTECTED]
> Subject: Re: FW: Delay in start of dialling...
>
>
> Mark Harrison wrote:
> >
> > Actually, I take that back. It takes about two minutes. Which,
> I'm sure you
> > would agree, is next to useless.
>
> 45 seconds before it starts dialing is already enough that many types of
> connection attempts will time out...  I have problems enough from the
> 40-50 seconds that my system takes to actually dial.  (Actually, a large
> portion of that time is waiting for the ISP to respond.  It dials in
> about 5 seconds, gets the login prompt from the ISP at about 20 seconds,
> sends the login, gets the passwd prompt at about 25 seconds, sends the
> passwd, gets authorized at about 35 seconds, pppd starts, and then waits
> 5-15 seconds to get an IP number.
>
> > -----Original Message-----
> > From: Mark Harrison [mailto:[EMAIL PROTECTED]]
> > Sent: Tuesday, 8 December 1998 9:25
> > To: [EMAIL PROTECTED]
> > Subject: Delay in start of dialling...
> >
> > I have tried to configure diald. Here is my setup:
> >
> > o RedHat 5.0 with 2.0.32 kernel, no updates applied
>
> I have a RH 5.1 install on my system.  Should be close enough.  Kernel
> version shouldn't be a problem, in this case.  (diald doesn't care about
> the kernel from about 2.0.0 until somewhere in the 2.1.x chain.
> Depending on your kernel compile options, even 2.1.131 works just fine
> with diald.)
>
> > o IPMasq is working correctly
> > o My internal network is 192.168.0.x
> > o I use a dynamic IP address from the ISP.
> > o BIND is configured and running correctly. It has one zone for
> my internal
> > network (overkill. I know:) and caches for the rest of the world.
>
> Depends - do you have more than one machine on your internal network?
> If so, that very well might not be overkill, especially if things might
> change on the network.  (It really helps out if you might have some
> mobile DHCP clients using the network, too.)
>
> > o Latest diald RPMs from an official RedHat mirror site (0.16.5a)
> >
> > At the moment, I have shell scripts that start and stop the PPP
> link. These
> > scripts, use the configuration from the network control-panel applet
> > provided with RedHat.
> >
> > When I have had diald configured and running, it seems to take over 45
> > seconds to start dialling. Is this normal? The Linux machine is only a
> > 486-100 with 32Mb of RAM but it is hardly doing anything.
>
> I have a 486-66 with 16M of RAM I could try it on.  And, I just
> remembered that it has RedHat 5.0 on it currently.  However, a friend of
> mine assures me he has absolutely no performance problems with his
> 486-33 with 8M RAM, which provides his network and his father's network
> with demand dial access, SMTP, and DNS.  Not to mention firewall.
>
> I have seen a delayed start like that caused by a buggy version of
> diald-top.  I used to run diald-top all the time, but lately, it doesn't
> run very often...  Other diald monitoring programs also have the
> potential to slow it down like that.  I don't know if any of them do
> slow it down, however.
>
> > Does someone here have a similar configuration that works who
> would be we
> > willing to share their diald configration files?
>
> Well, I don't have the similar configuration right now, but I could make
> one.  However, I think I'd prefer to work from your configuration, and
> fix the problem, so that the next time this problem comes along, we know
> have an idea of what they're doing wrong, and can patch their setup,
> rather than doing a blind replace.
>
> Ed
>

start-ppp

Reply via email to