🚀 go-pugleaf

(replying to my own posting:)

Since I haven't heard from anyone regarding my problems with dp-2.2,
and I have made considerable headway in getting the thing to work, I
thought I'd followup on these problems, just so others of you who are
just starting to install dp-2.2, won't have the same problems that I
had...

> I have recently installed dp-2.2-beta on my Sun ELC (SunOS4.1.2) and
> have read through the DialupPPP.doc several times.  I reconfigured the
> kernel and added the appropriate /etc/dp files.  I am slightly at a
> loss as to what to add to /etc/hostname.dp0, so I didn't use it.  Instead,
> I added the line:
> 
>   ifconfig dp0 129.212.49.100 129.212.49.101 netmask + -trailers up 
> 
> to /etc/rc.boot.
> 
> My machine (hostname = echidna) is a standalone machine (at home) and
> as such, doesn't have an ethernet connection.  In /etc, I only have
> one hostname.??0 file, and that is hostname.xx0 with the contents:
> echidna.
> 
> In my /etc/hosts file, I have the following lines:
> 127.0.0.1	echidna localhost loghost
> 129.212.49.100	echidna-dp0
> 129.212.49.101	pecan-dp0
> 
> "pecan" is a Solbourne machine that I administer at work, with the
> 129.212.49.* IP addresses assigned to its subdomain.  This is the
> machine that I intend to connect to, but first things first.
> 
> After adding the above "ifconfig..." line to /etc/rc.boot, and then
> rebooting, I see the following error messages (interspersed with normal
> messages):
> 
> keyserv: unable to register service
> rpc.statd: unable to register (SM_PROG, SM_VERS, udp)
> Oct 10 01:53:31 echidna inetd [117]: Cannot register service: RPC: Unable to send; errno = Host is unreachable
> 
> I have no idea what I'm doing wrong... Does anybody out there have a clue?

What I ended up doing to avoid the above boot-time messages is to move
the "ifconfig", "/usr/etc/dpd", and the "/usr/etc/route" statement to
the very end of the /etc/rc file.  That way all of the daemons that
complain above have already started and don't complain when I now
add the new statements.  Does this seem like a reasonable thing to do?
Am I missing anything by doing this?

> Ignoring the above errors, I manually invoke dpd using the following
> command:
> 
> /usr/etc/dpd -d 3
> 
> "dpd" just core dumps (and does not run).  There is nothing reported in
> /var/adm/dp/.  Any hints on this one??  I have heard that it (dpd) core
> dumps (sometimes) when the dp-conf files aren't setup properly...
> Is this true??  If so, where do I start looking for errors?

I was right about dpd core dumping.  I recompiled it with -g and used dbx
to discover that it was getting a SIGSEGV in the readconfig portion of
the code, while trying to print out an "unknown MODEM" message.  I played
around with my dp-conf file until I got it working.  What I did to fix it
is get rid of ALL blank lines in the file.  Not a real friendly parser, but
I'm very happy that it works now.

Now, on to some NEW problems:

Regarding routing:

I have assigned my home machine's dp0 interface the address of 129.212.49.100,
and my work machine's dp0 interface the address of 129.212.49.101 .  The
work machine (named pecan) has an IP address of 129.212.49.1 . The ethernet
at work is also connected to the rest of the site thru a router
(129.212.49.200).

My ifconfig and routing commands on my home machine are:

ifconfig dp0 129.212.49.100 129.212.49.101 netmask 0xffff0000 -trailers up
and
/usr/etc/route add default 129.212.49.101 1

My ifconfig and routing commands on my work machine (pecan) are:

ifconfig dp0 129.212.49.101 129.212.49.100 netmask 0xffff0000 -trailers up
and
/usr/etc/route add 129.212.49.101 129.212.49.100 1

Now, it's easy to get from my home machine to my work machine and
vice versa.  I have another machine on the network at work (129.212.49.2)
to which I've added "/usr/etc/route add 129.212.49.100 129.212.49.1 1".
So, now I can get to both of these machines.  My question is: Do I have
to add a "route" statement to *every* host on my network in order to be
able to get to them?  And how do I get to machines on the "other side"
of the router?  The way I'm doing it now is to rlogin to pecan, and then
rlogin to the "remote" machines, but that's not really elegant...


Regarding FTP/rcp thoughput:

I'm using WorldBlazer modems at both ends and have the following
configuration in use:

WorldBlazer - SA - Version LA5.01 - Active Configuration
 B1  E1  L1  M1  Q2  T   V1  X0  Y0
&C1 &D3 &G0 &J0 &L0 &Q0 &R3 &S1 &T4 &X0
S000:1   S001=0   S002:128 S003=13  S004=10  S005=8   S006=2   S007=40
S008=2   S009=6   S010=14  S011=70  S012=50  S018=0   S025=5   S026=1
S038=0   S041=0   S045=0   S046=0   S047=4   S048=0   S050:254 S051:253
S056=17  S057=19  S058:2   S059=0   S060=0   S061:0   S062=15  S063=0
S064:1   S068=255 S069=0   S090=0   S092:1   S093=8   S094=1   S100=0
S104=0   S105=1   S111=255 S112=1   S151=4   S155=0   S180=2   S181=1
S183=25  S190=1   S191:6   S253=10  S254=1   S255=255

Does this seem like a reasonable configuration?

[To Bob Sutterfield:]

Bob, a couple of weeks ago you wrote to comp.protocols.ppp claiming FTP
throughputs of 2.4-2.7 KB/s for /vmunix and 2.8-3.5 KB/s for /usr/dict/words.
I'm getting apprx. 2.4-2.6 KB/s for /vmunix, but only 2.0 KB/s for
large (~500K) text files.  Incidentally, when I'm uploading to work
from home, I often see the CTS (and thusly SD) LEDs go off.  Should
data be being held back this frequently?

Bob, could I possilbly get a copy of your WorldBlazer configuration so
that I can see what we're doing differently?

 
Any help with these problems would be greatly appreciated,

Larry Ader


-------------------------------------------------------------------------
 Larry Ader                      || 
 Amdahl Corporation              ||  ames   \
 1250 E. Arques Ave.  M/S 139    ||  apple   > !amdahl!lra00
 P.O. Box 3470                   ||  att    /
 Sunnyvale, CA   94088-3470      ||  uunet /
 ph. 408/992-2722                ||             or lra00@amdahl.com
 or 1-800-538-8460  x22722       ||
-------------------------------------------------------------------------