@peter.riche:

Damn…
Okay, thanks to you ;-)

But isn't there any solution? (patch, add-on or else??)

If there is, it would be up to the FreeBSD team. Check on their forum/website. :)

Hi, i am change 3Com card 3C905X to Planet with Realtek chipset (ENW-9503A) and disable acpi in to pfSense v. 1.0.1 (hint.acpi.0.disabled=1) and works perfectly!!!!!

We temporarily connected a cdrom to the unit itslef and just installed it.

After I posted, I noticed that there was a 12/12 build as well.  I used pfsense.img.gz from ~sulrich, which I believe is the embedded image.

I get the same result as with 12/11.  Hitting ? at the mountroot prompt displays zero boot devices.

Entering nothing causes the following:

panic: Root mount failed, startup aborted.

Post to freebsd-mobile.  This is a kernel issue.  Be prepared to provide tons of information to them, they are quite picky and require:

dmesg -a
uname -a

Just to even open the door on the mailing list.

I have both of them, but I like the b's better, too :-)

Any luck so far? So far my experiences has been to load Gentoo on the Qube. One thing to take note of is that the Qube 3 need special partition sizes and that the kernel cannot exceed a1800K and has to be bz compressed. This site might be helpful if you have not seen it yet. http://gentoo.404ster.com/texts.php?action=view&id=10

You will only get something used for that price and the specs. Just make sure the device you are looking for has at least 128mb ram, is x86 based and the nics are freebsd supported.

A nokia IP330 is pretty loud from what I have heared from others (search the forum for ip330, there are some discussions about that device).

What version of pfSense did you test with? If it's not already one of the latest snapshots that run freebsd 6.2 please retest with that ( http://pfsense.com/~sullrich/1.0.1-SNAPSHOT-12-06-2006/ ).

I don't know how much of this applies to pfSense, but with m0n0wall the included opensource driver has problems with this card, so you had to use the one provided my ZNYX. Nics are so cheap that we just bought a new one.

I don't think there is a workaround for this. The situation can even become worse if a user has a typo during login as some radius servers then usually delay the answer to prevent brute force attacks. You should try to discuss this at the m0n0 list as the CP we use is nearly a 100% port of the m0n0 feature and as it is still under developement there we don't plan to touch it to stay syncable with their code regarding to this.

@hoba:

Guess it should be able to handle that load as you "only" have 8 mbit/s. I tested a wrap here at a 16 mbit/s adsl2+ and gave it a hard time (p2p, multiple ftp downloads, etc) and it didn't break though it was not able to deliver the full speed and the cpu was heavily loaded. Just give it a try and monitor cpu load at status>system. If you get the latest snapshot there are some rrd cpu graphs now too  ;)

I have a WRAP running 1.01 release here at work running on a 45meg link.  I can only get throughput of about  16mbit/s on it which is right in line with what Hoba is experiencing.  I have tried all the tweaks in the config (polling, etc etc) and 16meg/s is about the most I can get out of the WRAP box.  Polling made response time jump about 10x (~1ms to ~10ms) but I didnt really see any bandwidth improvement…

I can assign console redirection, but swapping com1/2 isn't possible. Too bad you have to recompile the kernel for this…

Thanks GBelanger.
I've tried to change the ram stick for one 512 Mb stick and now I don't get each time i reboot the box.

I've noticed something and I don't know if it can help.
I've noticed that when I reboot the box just at the time it's under heavy load (cpu usage about 90 to 100%), the reboot is quite longer : it's quite long at the time it should display :
starting /usr/local/etc/rc.d/*.sh… done (twice)

And often, I get messages about out of swap space just instead of the normal display.

Hope this can help.

Thanks

Vincent

ok all loaded today and working  8) thank you for your help.

Ryan

Strangely enough, these problems have now gone away…

I did change the location of the router, but little else changed.

Regarding this, just some precisions I would like (if you have time, Scott):

What do the FreeBSD bug(s) refer to? Recognizing the RNG/ACE on the CPU or actually using it? Is this a major bug or something that can be worked out? Would it be worth it so start a bounty, for example?

I would participate in a bounty if you think it's doable. If its purely dependent on the FreeBSD core, then we'll just have to wait it out…

Thanks-

Guillaume

My guess would be Linux uses the HLT instruction to idle the CPU when it's not needed, and pfSense does not. Maybe you should try IPCop. It's based on Linux.

Thanks! That was a lot easier than I had expected.

-Alex

It could as well be some kind of weird attack where someone tries to fill up your switch with crap to do some manipulation or whatever (similiar to mac flooding: http://en.wikipedia.org/wiki/MAC_flooding ). It should not occur in a sane environment.