On 03.05.04, 10:31:40, Martin Husemann wrote:
> On Mon, May 03, 2004 at 10:17:40AM +0200, Bernd Sieker wrote:
> > So it's not me or my hardware, then. I've had similar experiences
> 
> I have this machine:
> 
> NetBSD 2.0_BETA (VOGT) #0: Fri Apr 23 00:25:55 CEST 2004
>         martin@emmas.aprisoft.de:/usr/src/sys/arch/sparc/compile/VOGT
> total memory = 65416 KB
> avail memory = 61428 KB
> bootpath: /sbus0/qec0/be0
> mainbus0 (root): SUNW,Sun 4/75: hostid 5541d471
> cpu0 at mainbus0: cache chip bug; trap page uncached: CY7C601 @ 40 MHz, TMS390C602A FPU
> cpu0: 64K byte write-through, 32 bytes/line, hw flush: cache enabled
> 
> running as a DSL router/NAT/firewall at work. It worked flawlessly for me
> since I installed it (and fixed the ipfilter configuration with Darren's
> help):

What are those ipfilter issues? I didn't realize anything unusual.

> 
> 10:27AM  up 10 days,  9:53, 1 user, load averages: 0.09, 0.12, 0.08
> 
> But then it has not realy been stressed:
> 
> Name  Mtu   Network       Address               Ibytes     Obytes
> pppoe 1492  <Link>                          4045956721  777443220

4GB in, 770MB out in 10 days? Yes, mine is slightly more busy,
500MB in, 250MB out in 6 hours, but not _that_ much more.

> 
> No IPv6 here.

Here neither.

The CPU is a v7? Not that I really think it's related, but just to
make sure.

My IP-NAT is quite strained at times, having to translate a rather
active p2p client (overnet). Sometimes I have 2500+ active MAP/RDR
entries in the table, but it usually calms down after a while to
about 200.

At those times I have ping replies in the order of 5000ms to any
outside address, even if directly from the router.

I'm thinging of getting a SPARCstation 5 to do the DSL routing
stuff, and reserve the SS20 for fairs and trade shows.

NetBSD 2.0_BETA (PORTIER.MP) #7: Sun May  2 00:30:04 CEST 2004
        bernd@boa.home.loc:/usr/source/2.0/src/sys/arch/sparc/compile/obj/PORTIER.MP
total memory = 64816 KB
avail memory = 60400 KB
bootpath: /iommu@f,e0000000/sbus@f,e0001000/espdma@f,400000/esp@f,800000/sd@1,0
mainbus0 (root): SUNW,SPARCstation-20: hostid 7271bfaf
cpu0 at mainbus0: mid 8: TMS390Z50 v0 or TMS390Z55 @ 75 MHz, on-chip FPU
cpu0: physical 20K instruction (64 b/l), 16K data (32 b/l), 1024K external (32 b/l): cache enabled
cpu1 at mainbus0: mid 10: TMS390Z50 v0 or TMS390Z55 @ 75 MHz, on-chip FPU
cpu1: physical 20K instruction (64 b/l), 16K data (32 b/l), 1024K external (32 b/l): cache enabled

I initially thought the issue might be ipnat/ipfilter with an SMP
machine, but after ripping out one CPU module and running a UP
kernel, I still got the same problems, to so I put it back in and
resumed MP operation.

I also had some severe problems (ultimately leading to a panic or
freeze) when running squid with squidGuard with an inconsistent
cache directory (left over after a crash). Individual squidGuard
processes were killed with signal 6 (ABRT), and probably restarted
by the master, until in the end I got a panic. I can't remember
which one, and I have now disabled squidGaurd for the time being.

> 
> Martin

-- 
Bernd Sieker

NetBSD - the power to swerve (penguins, worse than cane toads).
		-- Julian Assange