How long do you leave rate running to have this behavior?
Probably even a libpcap bug as well!

IIRC the order of preference from the driver goes something like:

MSIX -> MSI -> Interrupts. If you disable one, it falls back to the next in line. On FreeBSD 8.x several driver+chip combinations have issues with MSI-X and MSI, but others work fine. So it's preferred but not required.

Polling wouldn't be selected automatically it would have to be manually enabled. (and not recommended in almost every case)

I have several clients and servers running on a test VM and things get beaten up a lot there, lots of loss/latency (artificial), clients and servers restarting, and so on… I have never seen this happen.

Is there anything weird/special about the client that gets lost? Anything different in its advanced options? Is it an assigned OpenVPN interface? Used in any NAT rules or similar?

I would prefer to have a better maintenance support in a CARP situation also. I have a high availability cluster to avoid interruptions in our network. We have about 70 users in an apartment house and I can't inform all users and we have no maintenance window because there is always somebody online.

I support the idea of traxanos. I would appreciate the following additionals functions:

during boot process an option to disable CARP that means the slave server keeps online on virtual interfaces. The disabling button on page "Status –> CARP" should be persistant for next boot.

The possibility to disable during boot already is necessary if the master pfsense will not work properly but it takes virtual addresses. With such an option an unwanted interruption could be avoided.

Doh!! My bad. It is working just fine. My Extended Query syntax was incorrect.

@jimp:

What browser/OS combination are you using?

Windows 7 and Google Chrome.

And have you tried clearing your cache?

Just tried this, it seems to have fixed the problem (at least for now). I'll keep an eye on it and see if it comes back. Thanks for your help Jim!

Try a current RC1 snapshot. Several problematic patches related to FTP were removed recently.

Hi,
you should use IPv6 ULA addresses for your purpose.
You can generate them here:
http://www.simpledns.com/private-ipv6.aspx

Already a thread on this:

http://forum.pfsense.org/index.php/topic,65231.0.html

On a few of the cheaper switches I have used, VLAN with an ID of 1 can only be setup as a native vlan and cannot be tagged on any port that it is configured for. In your picture "VLAN 1"  is setup as  a tagged VLAN for pfSense. You did not mention what the VLAN id of VLAN 1 is.

I have had consistent results when creating trunks between pfSense and layer 2 devices when all the networks that pfSense needs to talk to are tagged and there is no traffic on the untagged VLAN.

Some switches that I have come across can only bind their IP stack to the untagged only VLAN with id=1 particularly some of the cheaper Dell switches. So I normally keep VLAN 1 exclusively dedicated for switch management traffic.

If you want to communicate on a trunk between pfSense and a switch with both tagged and untagged traffic then for the untagged traffic you will have to add the raw interface "re0" (just like you have re1). Then for the tagged traffic add a VLAN from the VLAN tab and make re0 as the parent interface for that VLAN.

Can you check if everything is okay with media where you're installing your update ? It must be CF card I suppose.

Ok, I refer to a scenario where I have two servers in a high performance cluster, the two have a captive portal configured with a freeradius server when the primary goes down, the clients have to re-authenticate in the secondary, would be excellent if also replicated to the secondary server authentication so that clients do not waste time.

Just figured out that turning on the option to resolve DNS v4 first seems to solve this issue… (on the squid general settings tab).

I still stick to my statement that if you just must write to these frequently that SLC is the only way to go.

Get yourself a few of these.  If you can break them, you are good at breaking stuff.

http://www.comx-computers.co.za/TS16GCF100I-P-specifications-70564.htm

I've got a full install from ISO and updated to latest snapshot (previous one now). Also duplicated last night at home on virtualbox. Don't have to even enter any settings. Just check the box, hit save. Reload page to make sure the setting is retained. Then hit TEST button then page reloads and box is no longer checked.

If you are having trouble replicating this I can pm you RDP details to my virtualbox so you can have a look and see what I am seeing.

Not quite that simple. It isn't just a matter of flipping a sysctl bit (which is what that page is for)

I think I used Ctrl+F5 - which should disregard the cache.

OTOH, I was using Chrome at that time…no idea what the latest bugs are which Google included in the latest update...apart from 95% of all tabs now not rendering properly any more. However, the 503 "rendered" correctly :-)

@c0urier:

Mine sounds like it's related to this.

Cannot see how.