[prev in list] [next in list] [prev in thread] [next in thread] 

List:       ipfilter
Subject:    problems with solaris 2.5.1
From:       Stefan Watermann <stefan () NeRo ! Uni-Bonn ! DE>
Date:       1997-09-17 8:10:35
[Download RAW message or body]


we have problems with ipnat as follows:

	operating system: Solraris 2.5.1 Sparc Station 5
	ipf versions: 3.1.5, 3.1.11 and 3.2b5

Our address translation setup works fine for some time until the machine crashes
or hangs. If the machine hangs, the ethernet interface is not able echo ping
requests.

NAT table:

map le0 192.168.101.0/24 -> 193.168.251.1/32
map le0 192.168.102.0/24 -> 193.168.251.1/32
map le0 192.168.103.0/24 -> 193.168.251.1/32

The addresses which have to be translated are routed through virtual interfaces
like le0:1, le0:2 and le0:3.

error messages after system crash:

Sep 15 19:18:49 habakuk unix: BAD TRAP occurred in module "ipf" due to an illegal access to a user ad
dress.
Sep 15 19:18:49 habakuk unix: sched: Data fault
Sep 15 19:18:49 habakuk unix: kernel read fault at addr=0x2, pme=0x0
Sep 15 19:18:49 habakuk unix: MMU sfsr=126: Invalid Address on supv data fetch at level 1
Sep 15 19:18:49 habakuk unix: pte addr = 0xf597b000, level = 1
Sep 15 19:18:49 habakuk unix: pid=0, pc=0xf5f31f28, sp=0xfbe1c9a0, psr=0x44006c0, context=0
Sep 15 19:18:49 habakuk unix: g1-g7: f0054a78, ffd15a00, ffef0000, ffefdd52, 2, ffefefa0, fbe1cec0
Sep 15 19:18:49 habakuk unix: Begin traceback... sp = fbe1c9a0
Sep 15 19:18:49 habakuk unix: Called from f5f32808, fp=fbe1ca20, args=f5e75e40 f5d58430 0 f0ffffff f0
fffc00 3
Sep 15 19:18:49 habakuk unix: Called from f5f2b534, fp=fbe1caa8, args=f5d58430 14 fbe1cb28 13422 92 f
0058ae8
Sep 15 19:18:49 habakuk unix: Called from f5f2d780, fp=fbe1cb68, args=f5d58430 14 f5b3c408 0 fbe1cd38
 f5b35708
Sep 15 19:18:50 habakuk unix: Called from f5f2db7c, fp=fbe1ccc8, args=fbe1cde0 f5b35708 fbe1cd38 0 0 
f5b09780
Sep 15 19:18:50 habakuk unix: Called from f006e7e0, fp=fbe1cd98, args=f5b35708 f5d66ed8 ffdf0e76 ffff
ff01 f5b4482c 800
Sep 15 19:18:50 habakuk unix: Called from f5a9c570, fp=fbe1cdf8, args=f5b35708 f5d66ed8 f594ba28 1010
0 f5f2d890 f594cf90
Sep 15 19:18:50 habakuk unix: Called from f0043494, fp=fbe1ce60, args=f5ca10ea f5b44858 80 5 f5b44800
 f5ca10e8
Sep 15 19:18:50 habakuk unix: Called from ffd19dbc, fp=fbe01ab8, args=fbe01ec0 0 1 f0271ad8 fbe01ec0 
fbe3aec0
Sep 15 19:18:50 habakuk unix: End traceback...
Sep 15 19:18:50 habakuk unix: panic: Data fault
Sep 15 19:18:50 habakuk unix: syncing file systems... done

We have seen, that the problems with the hanging machine occurs more often in v3.1.11 than
in 3.1.5 (which seems the most stable version for us) and 3.2b5.

Does anybody know something about these problems?

Stefan

[prev in list] [next in list] [prev in thread] [next in thread]