Re: Problems with linux and TIS fwtk

Juhani Rautiainen (jrauti@sasu1.carelian.fi)
Tue, 20 Oct 1998 09:05:13 +0300 (EET DST)


On Mon, 19 Oct 1998, Juhani Rautiainen wrote:

>
> Hi!
>
> We are using linux machine as a firewall in our firm. We upgraded
> our machine recently and in the same time switched to newer kernels
> from 1.2 series which served well. Problem is that TIS fwtk v2.1
> has problems with linux kernel (at least it seems so). If I use
> 2.0.35 or 2.0.36-pre13 plug-gw sometimes freezes the whole machine.
> With 2.1.125 kernel this doesn't happen but then http-gw doesn't
> work correctly. When web page has form which use POST-method result page
> doesn't come back correctly. I always get following requester from
> Netscape: "A network error occured while Netscape was receiving data.
> (Network Error: Connection reset by peer). Try connecting again."
> Machine which serves as firewall is K6-2 w/64 MB. Network cards are
> 3com 3c509s. So does anyone have a idea how I could debug this?
> I can use tcpdump if you tell which options should be used.
>

To continue my on thread I've analyzed the problem further and i have
an idea about possible problem but I no really nedd help. Here are
dumps from succesful and failed connection in the part theim seem
to differ. Results were taken with tcpdump -vv

Succesful is from 2.0.36-pre13 and failed from 2.1.126-pre2.
Ovi is firewall and sasu15 is my NT workstation (yuch). In dump2
there is remarks in the part where the possible problem can be seen.

dump1:
7:38:08.650933 ovi.carelian.fi.www > sasu15.carelian.fi.1277: P
25981:27441(1460) ack 654 win 32120 (DF) (ttl 64, id 103)
17:38:08.660933 sasu15.carelian.fi.1277 > ovi.carelian.fi.www: . ack 27441
win 8760 (DF) (ttl 128, id 2066)
17:38:08.660933 ovi.carelian.fi.www > sasu15.carelian.fi.1277: P
27441:27525(84) ack 654 win 32120 (DF) (ttl 64, id 104)
17:38:08.840933 sasu15.carelian.fi.1277 > ovi.carelian.fi.www: . ack 27525
win 8676 (DF) (ttl 128, id 2322)
17:38:08.840933 ovi.carelian.fi.www > sasu15.carelian.fi.1277: P
27525:28708(1183) ack 654 win 32120 (DF) (ttl 64, id 106)
17:38:08.870933 ovi.carelian.fi.www > sasu15.carelian.fi.1277: P
28708:29608(900) ack 654 win 32120 (DF) (ttl 64, id 108)
17:38:08.870933 ovi.carelian.fi.www > sasu15.carelian.fi.1277: F
29608:29608(0) ack 654 win 32120 (ttl 64, id 109)
17:38:08.870933 sasu15.carelian.fi.1277 > ovi.carelian.fi.www: . ack 29609
win 8760 (DF) (ttl 128, id 2578)
17:38:08.950933 sasu15.carelian.fi.1277 > ovi.carelian.fi.www: F
654:654(0) ack 29609 win 8760 (DF) (ttl 128, id 2834)
17:38:08.950933 ovi.carelian.fi.www > sasu15.carelian.fi.1277: . ack 655
win 32120 (DF) (ttl 64, id 111)

dump2:
17:35:56.945017 ovi.carelian.fi.www > sasu15.carelian.fi.1274: P
27474:28206(732) ack 654 win 32120 (DF) (ttl 64, id 3695)
17:35:56.946374 sasu15.carelian.fi.1274 > ovi.carelian.fi.www: . ack 28206
win 8007 (DF) (ttl 128, id 58385)
17:35:56.946408 ovi.carelian.fi.www > sasu15.carelian.fi.1274: P
28206:28258(52) ack 654 win 32120 (DF) (ttl 64, id 3696)
17:35:56.958871 ovi.carelian.fi.www > sasu15.carelian.fi.1274: P
28258:28988(730) ack 654 win 32120 (DF) (ttl 64, id 3698)
17:35:56.960212 sasu15.carelian.fi.1274 > ovi.carelian.fi.www: . ack 28988
win 8760 (DF) (ttl 128, id 58641)
17:35:56.960242 ovi.carelian.fi.www > sasu15.carelian.fi.1274: P
28988:29048(60) ack 654 win 32120 (DF) (ttl 64, id 3699)
To me problem seems to be here. Where did bytes between 29048
and 29560 go?
17:35:57.043119 ovi.carelian.fi.www > sasu15.carelian.fi.1274: R
29560:29560(0) ack 654 win 32120 (DF) (ttl 64, id 3702)
17:35:57.159539 ovi.carelian.fi.www > sasu15.carelian.fi.1274: P
28988:29048(60) ack 654 win 32120 (DF) (ttl 64, id 3704)
17:35:57.159996 sasu15.carelian.fi.1274 > ovi.carelian.fi.www: R
264155987:264155987(0) win 0 (ttl 128, id 58897)
17:35:57.224684 sasu15.carelian.fi.1275 > ovi.carelian.fi.www: F
607:607(0) ack 9707 win 8508 (DF) (ttl 128, id 59153)
17:35:57.224712 ovi.carelian.fi.www > sasu15.carelian.fi.1275: . ack 608
win 32120 (DF) (ttl 64, id 3705)

As you can see it seems that something goes wrong towards the end of the
transmission. It may be that there is something wrong with the NT (which
I can believe) but it doesn't explain why this works with the 2.0.* -
series.

---
Juhani Rautiainen 		jrauti@kati.fi

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/