@jim1000:

Wait … I see the problem. The hotel initiated request goes to a.b.c.d.:5001 to start the slingbox and to send commands back and forth. pfsense need to know about it and direct it to the inside router while it's in use. This means I will need to get off the inside router asap and just use it for internet surfing only. Normal internet stuff ok. Outside - in requests will need configuration.

So both answers are right. Cascading routers work in only surfing. Anything complex means double nat is bad.

Back to the original question about atom processors?

Yes, you got it!  If the incoming port is fixed, yes, you can map it all the way back.  If it is dynamic, you're hosed.

Yeah, the atom should be fine.  I refer back to the link I posted comparing your CPU versus the one I have.  You should have no issues.

@robi:

You wrote in your first post that the target budget is about $200. For $400, I think you can find stuff that fits your needs.

Considering that reality of that unit would be closer to $250-$300 built, I would say $400 is within reason but at the very high end. Do you have any suggestions?

@thezfunk:

I have that box.  There is a number of USB 3.0 ports on the back…I don't stay up on the supported USB NIC options with pfSense but if a supported gig USB 3.0 NIC is supported would that be a possibility?

While not ideal, I would consider it for use as a second WAN as well as the option to have a second LAN interface in the future. I've ordered one of those boxes so when it arrives I'll give it a shot. Still interested in learning more about a device in the $400 price range mentioned above.

Do you have any feedback on USB >> RJ45 adapters that are reliable? I see some decent name brands in the $40-$60 price range. Do I need to go that expensive or can I stay in the $14-$30 range?

Thanks.

Pretty much exactly that^.  ;)
The factory installs use a different update URL to ensure they are getting the correctly customised updates and a special USB memstick image is available via the support portal for re-installing.
That said I don't beleive that we apply a tweak such as this for the C2758. I have dealt with a few cases where people had accidentally upgraded to the vanilla image and the console was unaffected.

Steve

It's not such a bad choice, it was a cheap choice.

Just to report - the PCI version of the same network card (RP-3200R) is running flawlessly with pfSense 2.2.2. The problem seems to be related to the PCI-E in some way.

It's either a loop, or RSTP blowing up after a topology change. (just bit me!)
I'll bet the MAC address on the 3452 is lower than the 3216, and trying to take over as RSTP root bridge.  Blammo!  Nothing works.

The fix is to set the STP root bridge pirority lower on the 3216, preferably 0 instead of the default 32768.

thanks for the replies…, will definitely take note on this.

would like t hear from others about the 32Gb ssd as its OS, should this be enough? considering its only mostly for readonly stuff as I will be having another regular hdd for squid and other write intensive stuff.

just for reference are you using the +CFUN command or is it working out of the box?

Nope. And as far as I can tell, not modeswitch either. At least the cua's look the same as before I installed modeswitch, and the script itself said there are nothing to change. It just suddenly started to work and I have no freaking idea why. That bothers me.

Sometimes life is to easy  :-\

Bios to default settings seemed to fixed the issue..

Thanks for the response. Really, it seems to be related to the VM adapter. I looked in another VM running Suse SLES 11 SP3, and ethtool gave me no media results.

Check out some power management settings in BIOS and pfSense, maybe your FreeBSD power management instructs your NIC to switch to low-power mode because it detects the same from the switch (aka "Green Ethernet"). While Windows doesn't care about it. Also check the swicth, can you disable green ethernet features on it.

Munin Plugin updated for 8-core cpu temperature (tested on Supermicro A1SRi-2758F):

#!/bin/sh # - copy this plugin script to /usr/local/share/munin/plugins/, make it executable # - symlink it to /usr/local/etc/munin/plugins/ # - restart munin-service case $1 in config) cat <<'EOM' graph_title CPU Temperatures graph_vlabel Degrees C graph_category sensors cpu_0.label Core 0 cpu_1.label Core 1 cpu_2.label Core 2 cpu_3.label Core 3 cpu_4.label Core 4 cpu_5.label Core 5 cpu_6.label Core 6 cpu_7.label Core 7 EOM exit 0;; esac temps=$(/sbin/sysctl dev.cpu | grep temperature) echo -n 'cpu_0.value '; printf "$temps" | awk 'NR==1{print $2; exit}' | awk '{gsub(/C/,"")};{print}' echo -n 'cpu_1.value '; printf "$temps" | awk 'NR==2{print $2; exit}' | awk '{gsub(/C/,"")};{print}' echo -n 'cpu_2.value '; printf "$temps" | awk 'NR==3{print $2; exit}' | awk '{gsub(/C/,"")};{print}' echo -n 'cpu_3.value '; printf "$temps" | awk 'NR==4{print $2; exit}' | awk '{gsub(/C/,"")};{print}' echo -n 'cpu_4.value '; printf "$temps" | awk 'NR==5{print $2; exit}' | awk '{gsub(/C/,"")};{print}' echo -n 'cpu_5.value '; printf "$temps" | awk 'NR==6{print $2; exit}' | awk '{gsub(/C/,"")};{print}' echo -n 'cpu_6.value '; printf "$temps" | awk 'NR==7{print $2; exit}' | awk '{gsub(/C/,"")};{print}' echo -n 'cpu_7.value '; printf "$temps" | awk 'NR==8{print $2; exit}' | awk '{gsub(/C/,"")};{print}'

Read this: https://forum.pfsense.org/index.php?topic=93999.0

@dotdash:

Per their site, it looks like it can be populated with four modules that could be either 4x GB or 2x10GB SFF. The two 'management' NICs are built-in.
Does the unit have a module with 2 SFF ports? If so, try re-seating. IIRC, ixgbe is in the kernel and doesn't need to be manually loaded.

well after reading your post i realized that maybe it didnt even have the 10gbe ports since if they were it must have in pair of 2, not 4ports installed via 2 pcie cards :/

Thanks. I did see that in the other thread and forgot to mention that I had already enabled it.

Alright, thanks for the information.

Thanks for the tip almabes, I'll keep that in mind  :P

The thread may be locked, unless someone else wants to share their ideas for shelves… :)

For that high VPN throughput, the 4860 would be good, though I wouldn't use the ports as a switch as others have mentioned.

A good case can be made for keeping that RT-N16 around as your AP. I had an RT-N16 running Tomato but it recently died. It was only acting as an access point though. Replaced it by a Ubiquity UAP-LR and a Netgear GS108T, which was still cheaper than an equivalent current model Asus wireless box.

Follow Stephen's advice and test it under stable. If the nic works fine under FreeBSD-stable, report this on the open redmine ticket.

@Jason:

@bhawk6901:

@Jason:

It's not the number of tunnels that's the problem, it's the throughput.  Those boxes aren't very fast and will top out pretty low.  I've long since stopped using my 6501 so I can't speak to improvements made in 2.2.x, but I doubt you'd see more than 70-80Mbit/s.

yes throughput would be affected i know
we have 8 branches so as per your estimate, throughtput will be hardly 10-20 mbit/s :/

8 IPSec tunnels is nothing so you're fine there.

As to throughput, yes, if they're all running full out, expect no more than 10-20Mbit/s per location.  If you expect them to be running in that range you may want to consider limiters or traffic shaping as well, just to make sure that one location doesn't monopolize your entire capacity.

Thanks for your insight :)

STARRED THIS - It wasn't directly the solution, but it got me on the right track.

BIOS had the Gen1/2 option (default was AUTO), but forcing a specific setting had no effect.

What DID have an effect was the BIOS option to enable PCIe - Default was AUTO, and I had expected it would default to active - but apparently…. not so much.  When I forced it to enabled, it showed right up.

So, THANK YOU!!!!!!

Well I installed the OS today, and came up with this realtek chip 8168/8111 B/C/CP/D/DP/E/F/G

See my post https://forum.pfsense.org/index.php?topic=93797.0 At first i thought something was maybe bad drivers.

Anywho after installing i noticed i had 3% cpu idle with NO traffic at all… makes me wonder if it will handle a full gig