This would also be useful for me.. But one of the lines can be shaped still right, which is better than none, if VoiP could be forced down the shaped line, UNLESS the shaped line is down, then that'd be fine.

@dvserg: May be possable adding more (2-3) specific wizards to pfSence? existing wizard per users wizard etc.. Give it a shot. –Bill

@mwalsh: I am getting the following error (BETA3-PLUS-11548-PLUS-11567): Acknowledge All    .:.    04-21-06 07:34:04 - [filter_load]There were error(s) loading the rules: pfctl: the sum of the child bandwidth higher than parent qwanRootpfctl: linkshare sc exceeds parents sc/tmp/rules.debug:31: errors in queue definitionpfctl: the sum of the child bandwidth higher than parent qlanRootpfctl: linkshare sc exceeds parents sc/tmp/rules.debug:32: Syntax error in config file: pf rules not loaded The line in question reads [ the sum of the child bandwidth higher than parent qwanRootpfctl]:      .:. I have tried re-running the traffic shaper and have reduced the Voip settings to almost nothing but I am still getting it. Any help would be greatly appreciated Make sure you put the correct bandwidth in for up/down speeds on the second screen in the wizard. –Bill PS. Beta 4 is out, older releases are unsupported

@nima.m: In Traffic Shaper section, When you press the button "Remove Wizard" or pressing the tab "EZ Shaper wizard" it will remove your existing shaper rules and queue completely. Can you please put a messagebox asking the user if they really want to proceed before doing that ? During the last 3 days, I pressed those button/tab by accident many times and was forced to begin from the beginning. EZ Shaper wizard does have this warning (don't click next before reading the text on the screen!).  The other item we can put a javascript warning on I 'spose. –Bill

@hadi57: hi, one of my clients is able to use limewire after i setup the trafic shaper, how to stop limewire please? thank you Unsurprising.  It's a shaper, not a blocker. –Bill

Can't do that, get lots of error when I do that, To make it easier for understand, I changed back the name of my queues to the orginal name and sending the picture and my config file agian in this message. What do you think I should do now ? [image: Untitled-1.gif] [image: Untitled-1.gif_thumb] [image: Untitled-1.gif_thumb] TrafficShaper.xml.txt

That connection belongs to the same state and will be treated the same way  like the outgoing request.

@basset: Oh … Sorry about that ... I'm getting the latest from CVS now. Blaiming newbie status for this  :-[ Basset [/quote] HEAD is not meant to be consumed by non-developers.  You should be running code in the RELENG_1 branch.  Better yet, unless you find a bug that has been fixed post release, you should really run a released binary.  We can't and won't support anything else.  I'm marking this thread solved.  Thanks –Bill

Search the forum, this has been asked and answered before.

It's up to you to set your bandwidth limit. 20% might have been a bit conservative. However, I recommend running some bandwidth test with the shaper turned off first. Test upstream and downstream. I have not yet seen a line that had exact the advertised speed. Some are a bit slower, some even a bit faster. Don't trust your provider  ;) In case your line is slower than advertised and you use the advertised values shaping won't be successful when reaching your lines limit.

cool. thanks

I think what he's looking for is a way to monitor the total throughput for the month….possibly auto-throttle at a certain usage amount so that he doesn't go over on his usage.....ie you get 100gb/month transfer limit then pay large amounts per gig after that.

@billm: If I work it that way, you OK with bandwidth disappearing if HFSC is selected as the scheduler? This convo might be better off on the dev list, I don't check the forums that often :) –Bill Ok.. I'll move this to the dev list later today when I wake up  :).

@nima.m: Thanks for the response. Then, Will this scenario be possible : I have 5 computer connected to my Lintex router trough LAN. Can I configure pfsense to have low prioritizing  computer 1 based on the "IpNumber/Mac Address" and standard prioritizing computer 2 based on the "IpNumber/Mac Address" and high prioritizing computer 3-6 based on their "IpNumber/Mac Address" ? Yes this can be done but not so easily with the wizard. You will need to run the wizard then create custom queues and rules. I use this setup on my lan where I give piority and 59% of the bandwidth to a set of computers (me only ;)) and 39% to a few others. However you can't use mac addresses so I had to make use of static arp to assign the ipaddress.

Just wanted everyone to know everything is working great. This blows m0n0wall away. Clients are pissed off as they were getting a 10Mbs feed but they were only paying for 3Mb. NTOP Great addition really helps with figuring out where my client traffic is coming from and who to point the finger at.

tcp/udp port forward from external 22105 to internal 192.168.105.105:22105 some other forwards to 192.168.1.2 internal network is 192.168.0.0/16 queues: upload: name: up105105 bw = 1% priority = 1 m2= 25kB RED,ECN, qWanroot other filelds are empty download: the same as above but name: down105105 m2 is empty qLanroot rule: upload queue - up105105 download queue - down105105 (I found that specifying here qP2Pdown doesn't work) LAN->WAN protocol: any source: 192.168.105.105 (single host) rest of fields are empty/default squid is running in transparent mode, version: snapshot 25/03/2006 config-siroccofw1.local-20060405004147.xml.tar.gz.txt

:-[ OK we wait up to version 1.1 Danke / thx Stefan

There's a limit of 256 queues in ALTQ, fyi.  We can change that if/when it becomes an issue (at this time pfsense rule generation performance will be a bottleneck long before you get to 256 queues though). –Bill

@sullrich: This is to be expected.    CARP/pfSync does not sync the ALTQ assignments.  When a new connection (state) is created then the queue is reassigned. This will change in the future with our Layer 7 shaper detection work that is ongoing. I think this is also "fixed" in the kernel patches that I made not too long ago (not in 1.0). –Bill