net-2 woes (was upgrade woes)

net-2 woes (was upgrade woes)

Post by John Paul Morris » Fri, 13 Aug 1993 13:49:25



Thanks to the people who e-mailed me.
A short summary: ignore HJ's docs. They explicitly say that lib4.4 needs
pl10 or higher; but you dont need pl10 or higher to use lib4.4.1.
I installed lib 4.4.1 under pl9, then installed gcc 2.4.5 and recompiled
the kernel.

Now moving along:

I thought NET-2 was supposed to *fix* Linux networking?
I got all the net-2 binaries, etc; read the FAQ, and installed it.
I was pleased..briefly..and then the networking began to crumble.
(I am hardly a newbie at setting up TCP/IP stuff).

BTW, dont E-mail if you want me to see this. Net-2 broke our e-mail
(works randomly now, if at all).

I can ifconfig ok.
routing: net-2 seems to drop routes, or something. I have to add
route add default gw 137.82.52.254, otherwise nothing gets out of our
subnet. I do that in the rc files, but periodically routing just
disapears. ie, I can no longer ping different sites.

I am very suspicious of net-2 routing. Now what does this mean:
(output of netstat -r)
fs0.ee.ubc.ca           router          UGHDM           0       2 eth0
etc.

now netstat -nr
137.82.52.1             137.82.52.254  UGHDM            0       2 eth0

Why is this following the default route when the host (fs0) is on the
same subnet as I am (137.82.52.117, netmask 255.255.255.0) Is the output
of netstat wrong?

Mail:

the net-2 package didnt have mail in it, so I tried to change inetd.conf
like what I had under net-1. I use smail to handle mail; invoked as
smtpd.
from inetd.conf

smtp  stream tcp nowait root /usr/etc/tcpd /usr/bin/smtpd

When I first start up, mail seems to work. LAter, it dies. I cant
even do
telnet 127.0.0.1 25 and get the smtp daemon.
After a while (we get mail all the time) I have a dozen smptpd process
running, and netstat shows all the hosts trying to send mail. here
is the netstat output, showing weird mail. Is this an smtpd problem?)

Active Internet connections
Proto Recv-Q Send-Q Local Address          Foreign Address        (State)
tcp        1      0 localhost:1178         rflab.ee.ubc.ca:finger FIN_WAIT1  
tcp        1      0 rflab.ee.ubc.ca:smtp   ucsd.edu:3882          SYN_RECV  
tcp        1      0 rflab.ee.ubc.ca:smtp   mother.ludd.luth.:1479 SYN_RECV  
tcp        1      0 rflab.ee.ubc.ca:smtp   ucsd.edu:3941          SYN_RECV  
tcp        1      0 rflab.ee.ubc.ca:smtp   joker.cs.hut.fi:4849   SYN_RECV  
tcp        8      0 rflab.ee.ubc.ca:smtp   localhost:1158         CLOSE_WAIT  
tcp        2      0 rflab.ee.ubc.ca:smtp   mother.ludd.luth.:1723 CLOSE_WAIT  
tcp        2      0 rflab.ee.ubc.ca:smtp   mother.ludd.luth.:1897 CLOSE_WAIT  
tcp        2      0 rflab.ee.ubc.ca:smtp   ucsd.edu:4391          CLOSE_WAIT  
tcp        0      0 rflab.ee.ubc.ca:smtp   mother.ludd.luth.:2010 ESTABLISHED  
tcp        0      0 rflab.ee.ubc.ca:smtp   mother.ludd.luth.:2027 ESTABLISHED  
tcp        0      0 rflab.ee.ubc.ca:smtp   ucsd.edu:4533          ESTABLISHED  
tcp        2      0 rflab.ee.ubc.ca:smtp   srlns1.srl.ford.c:4950 CLOSE_WAIT  
tcp        0      0 rflab.ee.ubc.ca:1176   iskut.ucs.ubc.ca:nntp  ESTABLISHED  
tcp        1      0 rflab.ee.ubc.ca:telnet annex2.net.ubc.c:14900 SYN_RECV  
tcp      218      0 rflab.ee.ubc.ca:telnet annex2.net.ubc.c:10000 CLOSE_WAIT  
tcp       39      0 rflab.ee.ubc.ca:telnet localhost:1156         CLOSE_WAIT  
tcp       32      0 rflab.ee.ubc.ca:telnet localhost:1157         CLOSE_WAIT  
tcp       46      0 localhost:telnet       localhost:1184         CLOSE_WAIT  
tcp        4      0 rflab.ee.ubc.ca:finger localhost:1178         CLOSE_WAIT  
tcp        1      0 localhost:1156         rflab.ee.ubc.ca:telnet FIN_WAIT1  

Also, incoming finger, telnet, shell, finger etc. all seem to work randomly.
(I cant believe net-1 was actually more reliable here)

Outgoing telnet, rlogin, ftp and NNTP etc. seem to work (unless the routing
disappears; then I have to re-add the route). On a positive note,
ftp seems much faster. Is there something wrong with inetd???

It's not a matter of it not working at all; but incoming telnets
finger, ftp etc. work one moment and not at the next.

I followed the FAQ, and used the provided net-010 config files to
try to get the network working. But the FAQ is wrong in places:
eg
route add 137.82.52.0 ( ie $NETWORK ) (the same if i do loopback)

and I get
SIOCADDRT: Network unreachable.

This is probably just a mistake in the docs.

thanks in advance. I have a few weeks of the NET channel activists
mailings...maybe there is something in there...

--
______________________________________________________________________________
 John Paul Morrison                     |
 University of British Columbia, Canada | Hey hey!! Ho ho!!
 Electrical Engineering                 | Tax & spend liberals

________________________________________|____________________________________

 
 
 

1. CD-RW mounting woes in Mandrake 7.0 woe oh woe oh woe!

I hate to ask the often answered, but I have a TDK VeloCD CD-RW in my PII
233 machine installed as a master on the secondary IDE. I'm running
Mandrake-linux 7.0, and of course I get the bad link problem. I have run the
fix suggested on Mandrake's Web-page, but I can't figure a way to make the
fix either permenant or automatic. The fix (in case you haven't seen it is
to run "cdrecord -scanbus" to check for the presence of a working CD-RW,
then to type "cd /dev && rm cdrom && ln -s scd0 cdrom" in a terminal) This
fix works but there's got to be a better way...any thoughts.

2. Dataless clients mounting /usr through a gateway?

3. XF86 Config set-up woes & woes & woes

4. Multiple SCSI controllers

5. Kernel woes: Yes, I am a moron :)

6. BSD signal routines under SYSV

7. REDHAT 5 woes.... wondering if I am cursed..

8. Term developments and uploads

9. EISA Net-card woes (free HP27248 no supported?)

10. Aggggh, net modules woes

11. Net bsd woes

12. Autofs setup woes....... anyone help us mount /net?

13. Net Card Woes