For the record:
Planet FGSW-2620VM (24 100Mb + 2 1Gb ports)
connected to Alix 2D13.
Got it setup with 2 VLANs for ISPs going to a single WAN port. Various VLANs for local subnets going to OPT1. A VLAN for the main local subnet connected directly without tagging to LAN.
Took a while to understand the web GUI of the Planet device and work out what some of the "English" meant. Don't buy one of these if you want a user-friendly setup GUI!
Don't use VLAN1 for an ordinary client VLAN - it seems to insist on sending stuff from VLAN1 ports to/from all over. I gave up trying to find more knobs to tweak to keep VLAN1 in its place.
In the end it does work - 2 ISPs are funneling through it to WAN, a bunch of internal subnets are on LAN and OPT1. And it was a reasonable price and available in Nepal.
I have an all-Gb port fancier model coming soon, will give an assessment of that next week…

Oh, nice. Good looking! :P
Thx

I can't say I have, and with the price tag on that nic, considering you can get server grade nics for a 10th of the price, it may be hard to find someone who've also used that specific one.  With luck though someone may chime in.  or hell, maybe even someone using a realtek RTL8111E onboard.  I wouldn't see any difference in performance between the one you listed and the onboard. Seems like a real common onboard nic.  Only realtek nics i have are 10/100 Pci cards.  I wouldn't beable to tell you the performance of the 8111e, but the 10/100 did work and pegged at 98~100 solid.

Yes those should easily handle it.

Steve

Without Snort or Squid that board/CPU will handle >1Gbps so, yes, I'd say it will be fine.

Steve

Tried this card (501-6738-10/qgexpci) and system would hang at probing phy failed, only the led's for the first nic will light up.

I just wanted to post to let everyone know this card is a gamble as the user above seems to have no problem but others have had the same issue as me. I wished to also post in the following threads in case it helps anyone from making the same mistake but they appear to be locked.

http://forum.pfsense.org/index.php?topic=46799.0;prev_next=prev
http://forum.pfsense.org/index.php/topic,33960.0.html

I also previously got stung by the quad port mt (and posted about this).

Sorry, I see your question only now.
I'd like to tell you that I'm a Linux guru and the revised code is mine… but actually I was desperate and when I thought it was impossible to make work the e372 key, I started to try every MessageContent related to a huawei HSUPA Keys. I was very lucky, cause after some experiment, the intenet key started to work.
The key give me a connection speed of 4Mb in DW and 4Mb in UP while the same internet key on a Windows client give me 14 Mb in DW and 4 Mb in UP.
I don't know if this poor DW bandwidth is due to a improper MessageContent I found or it is due to a slow USB 1.1 port, but it works suffice to provide a decent vpn service.

Let us all know after you try.  I'd guess yes.  PFsense's graphics requirements are very low and very compatible.

Nice.  :)
Feel free to suggest any improvements to the firebox wiki page in this thread:
http://forum.pfsense.org/index.php/topic,59821.0.html

Steve

Agree. When BIOS didn't see the disk I knew I was hosed.

Can Plop see it ? :http://www.plop.at/en/bootmanager/index.html

I figured that out when i had it connected to the net and restored the backup, very good to know

If you want to try the 2.5.8 version of the ixgbe driver:

fetch -o /boot/modules/ http://files.nyi.pfsense.org/jimp/ixgbe-2.5.8/amd64/if_ixgbe.ko
or
fetch -o /boot/modules/ http://files.nyi.pfsense.org/jimp/ixgbe-2.5.8/i386/if_ixgbe.ko
-or-
fetch -o /boot/modules/if_ixgbe.ko http://files.nyi.pfsense.org/jimp/ixgbe-2.5.8/amd64/if_ixgbe-no_tso.ko
or
fetch -o /boot/modules/if_ixgbe.ko http://files.nyi.pfsense.org/jimp/ixgbe-2.5.8/i386/if_ixgbe-no_tso.ko

And then toss this in /boot/loader.conf.local

A PCI 3G modem? I'm not familiar with one. Do you have a link?

There are many USB 3G modems which work: http://doc.pfsense.org/index.php/Known_Working_3G-4G_Modems

Steve

I see you have Squid3 beta. My recommendation is to uninstall Squid3, remove all related files do a clean reboot with no traces of Squid3 and then do clean install of stable Squid 2.7.9 pkg v.4.3.3

For hard disk cache system use aufs and for memory replacement use LFUDA

Yup, 2.1 saw everything.  FYI to folks:  The chipset isn't listed on the vendor docs, but pfSense 2.1 see the RealTek chipset as 8168/8111.

So it works great with 2.1 so far.

Thanks, for some reason I was thinking they were access ports. That explains it.

An Atom will do approx. 50Mbps OpenVPN in one direction.
http://forum.pfsense.org/index.php/topic,27780.0.html
I have never thought about it too deeply but I can't see why you would be limited to your upload speed. A VPN connection may well require more bidirectional data than an unencrypted connection but you will still get more data in one direction.
If you are using the pfSense box as a VPN server, for securing a public connection for example, then you are limited by the slowest part of the connection.

Steve

@adam65535:

How stable are the i350 Intel nics?  The only reason I didn't choose them is they are not on the 8.3 hardware compatibility list.  I figured even if they did show up soon they would be unstable since they would be new.

Mine work fine.  They're actually a fairly old part (in hardware terms), the i350 was launched in 2011-Q2.

I don't see them on the 8.3R HCL but it's explicitly listed in the release notes.

http://www.freebsd.org/releases/8.3R/relnotes-detailed.html

@ebo2002fr:

Does it worth increasing vm.kmem_size and vm.kmem_size_max ?

Not likely.

@ebo2002fr:

Do you think it is a hardware issue ?

Yes, your "hard drive" shouldn't be causing:
@ebo2002fr:

2013-07-18  13:27:05  crit  routeur  kern  kernel  ata2: hardware reset timeout
2013-07-18  13:27:04  crit  routeur  kern  kernel  ata2: port is not ready (timeout 10000ms)
2013-07-18 13:26:56 crit routeur kern kernel g_vfs_done():ad4s1a[WRITE(offset=192856064, length=16384)]error = 6
2013-07-18 13:26:56 crit routeur kern kernel g_vfs_done():ad4s1a[WRITE(offset=192856064, length=16384)]error = 6
2013-07-18 13:26:42 crit routeur kern kernel g_vfs_dkone():ad4s1a[WRITE(offset=6144000, length=2048)]error = 6
2013-07-18 13:26:42 crit routeur kern kernel ng_vfs_done():ad4s1a[WRITE(offset=65536, length=2048)]error = 6
2013-07-18 13:26:42 crit routeur kern kernel g_vfs_done():ad4s1a[WRITE(offset=199290880, length=2048)]uerror = 6
2013-07-18 13:26:42 crit routeur kern kernel ata2: hardware reset timeout
2013-07-18 13:26:42 crit routeur kern kernel ata2: port is not ready (timeout 10000ms) tfd = 00000080

This is running a magnetic 5400 rpm HD.  I did notice I put it on a quad core 8gb ram machine with a HD and I got about 400k.  Then I tried it on my junky neoware ca22 running nanobsd with 512ram and got 4MB.  Can anyone explain?  Is it because the nanobsd version is reading instructions from ram and not disk?