Traffic Shaping wizard error
-
It's in the Traffic Shaping section.
-
Aha! so I should the speed and run the wizard again?
-
Well, I was kind of wrong, every time you run the wizard that setting will disappear. Where I was wrong is that you don't want to set the interface speed, but your actual internet speed there. Under "qInternet" set that 97% of your interface speed.
Example:
25Mb/s (LAN) - 97% = 24.25Mb/s (qInternet)EDIT: I wish I knew the magic behind how this thing does its math. If I change my LAN to anything other than 100Mbit/s or 1Gbit/s it throws an error now saying the child is higher than the parent.
altq on em1 hfsc bandwidth 26214.4Kb queue { qACK, qDefault, qP2P, qGames, qOthersHigh, qOthersLow } queue qACK on em1 bandwidth 20% hfsc ( ecn , linkshare 20% ) queue qDefault on em1 bandwidth 10% hfsc ( ecn , default ) queue qP2P on em1 bandwidth 5% hfsc ( ecn , linkshare 5% , upperlimit 5% ) queue qGames on em1 bandwidth 20% hfsc ( ecn , linkshare 20% ) queue qOthersHigh on em1 bandwidth 10% hfsc ( ecn , linkshare 10% ) queue qOthersLow on em1 bandwidth 5% hfsc ( ecn , linkshare 5% ) altq on em2 hfsc bandwidth 26214.4Kb queue { qLink, qInternet } queue qLink on em2 bandwidth 20% qlimit 500 hfsc ( ecn , default ) queue qInternet on em2 bandwidth 24606.72Kb hfsc ( ecn , linkshare 224606.72Kb , upperlimit 224606.72Kb ) { qACK, qP2P, qGames, qOthersHigh, qOthersLow } queue qACK on em2 bandwidth 20% hfsc ( ecn , linkshare 20% ) queue qP2P on em2 bandwidth 5% hfsc ( ecn , linkshare 5% , upperlimit 5% ) queue qGames on em2 bandwidth 20% hfsc ( ecn , linkshare 20% ) queue qOthersHigh on em2 bandwidth 10% hfsc ( ecn , linkshare 10% ) queue qOthersLow on em2 bandwidth 5% hfsc ( ecn , linkshare 5% ) -
hey…!?! I think i just realized this is a bug. Look at the linkshare and upperlimit numbers, they appear to add an extra "2" in front. Now it makes sense why the child is bigger than the parent.
-
I have the same problem and I'm running the AMD64 bit version. I need the traffic shaping for my VOIP phones. This thread seems to have ended without a resolution other than identifying the issue. Is there a way to fix it or is it a matter of waiting until it gets fixed by others?
Whats odd is my test PC, an older AMD64 based system with old components in a box the size of a refrigerator works great. This is happening on the system I want to use in my production system based on an Intel atom with gigabit NICS.
Can the wizard be run and then manually edited?
-
I am pretty sure I found my problem and am posting this here in case others have the same problem. I had the error when I was running in a test environment without a WAN connection. I plugged proper network connections in the system and deleted my traffic shaping rules, both under Traffic Shaping and Floating Rules. I re-ran the single-wan multilane wizard and everything is happy - including me.
-
Possible bugs can be submitted via redmine:
http://redmine.pfsense.org/projects/pfsense/issues
Can the wizard be run and then manually edited?
Yes.
-
I am pretty sure I found my problem and am posting this here in case others have the same problem. I had the error when I was running in a test environment without a WAN connection. I plugged proper network connections in the system and deleted my traffic shaping rules, both under Traffic Shaping and Floating Rules. I re-ran the single-wan multilane wizard and everything is happy - including me.
Nope that wasn't my problem. When I reboot I get the same error. So, is this a known problem? Can I do anything to fix it?
-
To explain the bug that I see…
queue qInternet on em2 bandwidth 24606.72Kb hfsc ( ecn , linkshare 224606.72Kb , upperlimit 224606.72Kb ) { qACK, qP2P, qGames, qOthersHigh, qOthersLow }Notice on the linkshare and upperlimit numbers there is an extra "2" before the rest of the nubmers. This is why if I set 25 or 30 it still fails. I can set 225 and it will work, but that's still way above my bandwidth.
I think I saw one post saying that its now old code and it has been updated, so you should update to the latest SVN release. But for me that's just not-possible. I wish i knew which line of code that came from I would take a look and see where its injecting that extra digit.
-
Whoo.. finally I figured it out. Don't use the wizards, they are broken and create broken rules.
Create your rules from scratch for both the floating rules, and for the queue's themselves. Follow this guide to help you understand how to create everything and what everything means. https://calomel.org/pf_hfsc.html
Here are my queue's, don't forget to create the floating rules in the firewall.
[2.0.1-RELEASE][]/root(1): grep -A2 -e '^ *queue' -e 'altq' /tmp/rules.debug altq on em1 hfsc bandwidth 25Mb queue { ack, dns, bulk, web, torrent, games } queue ack on em1 bandwidth 30% qlimit 500 hfsc ( realtime 20% ) queue dns on em1 bandwidth 5% qlimit 500 hfsc ( realtime 5% ) queue bulk on em1 bandwidth 20% qlimit 1000 hfsc ( ecn , default , realtime 20% ) queue web on em1 bandwidth 5% qlimit 500 hfsc ( realtime (10%, 10000, 5%) ) queue torrent on em1 bandwidth 1% qlimit 500 hfsc ( upperlimit 95% ) queue games on em1 bandwidth 20% qlimit 750 hfsc ( realtime 20% ) altq on em2 hfsc bandwidth 30Mb queue { web, bulk, games, torrent, dns, ack } queue web on em2 bandwidth 5% qlimit 500 hfsc ( realtime (10%, 10000, 5%) ) queue bulk on em2 bandwidth 20% qlimit 1000 hfsc ( ecn , default , realtime 20% ) queue games on em2 bandwidth 20% qlimit 750 hfsc ( realtime 20% ) queue torrent on em2 bandwidth 1% qlimit 1500 hfsc ( upperlimit 95% ) queue dns on em2 bandwidth 5% qlimit 500 hfsc ( realtime 5% ) queue ack on em2 bandwidth 30% qlimit 500 hfsc ( realtime 20% ) altq on em0 hfsc bandwidth 30Mb queue { bulk, dns, ack, web } queue bulk on em0 bandwidth 20% qlimit 1000 hfsc ( ecn , default , realtime 20% ) queue dns on em0 bandwidth 5% qlimit 500 hfsc ( realtime 5% ) queue ack on em0 bandwidth 30% qlimit 500 hfsc ( realtime 20% ) queue web on em0 bandwidth 5% qlimit 500 hfsc ( realtime (10%, 10000, 5%) ) -
@djroketboy:
queue bulk on em1 bandwidth 20% qlimit 1000 hfsc ( ecn , default , realtime 20% )
With regard to ECN you might need to tune net.inet.tcp.ecn.enable check http://forum.pfsense.org/index.php?topic=46960.0
-
With regard to ECN you might need to tune net.inet.tcp.ecn.enable check http://forum.pfsense.org/index.php?topic=46960.0
Cool, thanks. I just enabled it.
[2.0.1-RELEASE][]/root(3): sysctl net.inet.tcp.ecn.enable=1 net.inet.tcp.ecn.enable: 0 -> 1 -
I had the same issue.
My setup is dual WAN DSL, loadbalanced w/failover pfsense 2.0.1. Each DSL line has at least 6.25Mbps bandwidth.
I resolved it by using the wizard then correcting its mistakes:
- Go to "Firewall > Traffic shaper > By Interface > LAN > qInternet": Check that your total bandwidth is your measured speed 12.58Mbps (approx. 2x your individual line speed)
- Go to "Firewall > Traffic shaper > By Interface > WAN": This is where the wizard screws up. Replace the 1Mb under "bandwidth" with 6.3Mbps (Your measured speed for that line)
- Go to "Firewall > Traffic shaper > By Interface > WAN2": This is where the wizard screws up. Replace the 1Mb under "bandwidth" with 6.3Mbps (Your measured speed for that line)
Basically make sure the bandwidth under LAN fits into the sum of each line's bandwidth.
Then Save, reboot.
So wizard is ok, just fix the couple little mistakes after. Or do it manually, but once you do it with the wiz you can see what needs to be filled out and do it yourself.
-
Some questions:
1.) is the qLink (default) queue necessary for the LAN interface? It's auto setup by the traffic shaping wizard.
2.) are "drops" in a queue something that should be expected? should they be ignored? or have you found there rarely to be "drops" listed beneath your status > queues? On large file downloads at high speeds I see 5000, 7000+ although the resulting file is fine.
