Forums

×

Warning

JUser: :_load: Unable to load user with ID: 26383
Medium
Offline
Resolved
0 votes
My friend and I have the following setup at each of our houses:

-ClarkConnect 5.2
-Cable Internet 30mbps down/3mbps up
-QoS/Bandwidth module setup with the exact same rules (only diff is my network is 192.168.0.0/24 while his is 192.168.1.0/24)

A picture of his settings (basically identical to mine except as stated above):

http://dl.dropbox.com/u/5544258/QoS.jpg

The thing that we can't seem to figure out is, the settings work perfectly on my network, where I get my full 30 or so mbps down on a speedtest and my VoIP ATA works great.

Yet on his machine when he turns on the QoS service it drops his download speedtest to 12-13mbps even when the network is idle then back to normal when off. Also if he deletes all rules but leaves service on, he gets normal speeds.

We tried reinstalling app-bandwidth but no luck.
Wednesday, January 08 2014, 07:07 AM
Share this post:
Responses (7)
  • Accepted Answer

    Friday, January 17 2014, 07:28 PM - #Permalink
    Resolved
    0 votes
    Well odd thing is we are still using the exact same NIC drivers for both cards.
    The reply is currently minimized Show
  • Accepted Answer

    Friday, January 17 2014, 03:17 PM - #Permalink
    Resolved
    0 votes
    Thanks for the feedback! glad you figured it out.. NIC drivers maybe as Nick suggested?
    The reply is currently minimized Show
  • Accepted Answer

    Thursday, January 16 2014, 12:21 PM - #Permalink
    Resolved
    0 votes
    We have finally figured out the solution.

    It turns out his PC had some sort of hardware issue.

    For some reason his PC had something where it randomly shut off and then he turned it back on and it shut down after like 10 min.

    We simply swaped the HD's into a similar make/model PC and fixed some networking config but basically had all the same settings as we used the same PCI NIC and the Integrated was same model.

    Everything is working perfectly and his QoS works identical to my performance.

    We aren't sure what hardware issue he had but it oddly only affected his Internet if he had QoS on, LoL.
    The reply is currently minimized Show
  • Accepted Answer

    Thursday, January 09 2014, 12:04 PM - #Permalink
    Resolved
    0 votes
    Hmm could be an issue with one of the rules? particularly if it works when they are not all enabled

    Grouping IP's together by subnet I think used to create a rule for each entry (i.e. 254 for each) because of the way subnets were enumerated by the webconfig with each rules reserving a small proportion of bandwidth rather than having a shared pool for all IPs. Can you just try changing them to one rule to prioritise all traffic leaving your network on port 1935, rather than using the source address?

    In addition you usually want to prioritise traffic that's leaving your network as this is the bottleneck, so prioritise time delay sensitive traffic what you want to use (SSH, DNS, VOIP, streaming etc) and remove the download rules?
    The reply is currently minimized Show
  • Accepted Answer

    Wednesday, January 08 2014, 12:53 PM - #Permalink
    Resolved
    0 votes
    Yes I have done the speedtests with an average to put in the values.

    We did each 3 speedtests then averaged the values and put in 95%.

    Also if it helps somewhat, when he had the RTMP 1935 rule in only before the QoS settings worked fine and he had no issues at all.

    It seems the problem only arose after we added more rules.

    I will definitely look into the NIC drivers though.

    NIC Card Results:


    01:06.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL-8139/8139C/8139C+ (rev 10)

    01:0b.0 Ethernet controller: ADMtek NC100 Network Everywhere Fast Ethernet 10/100 (rev 11)



    alias eth0 tulip

    alias eth1 8139too



    2.6.18-308.1.1.v5

    The reply is currently minimized Show
  • Accepted Answer

    Wednesday, January 08 2014, 12:47 PM - #Permalink
    Resolved
    0 votes
    These are usually cuased by having an upload speed set above your actual available upload bandwidth

    Have you done any speed tests to confirm what your actual bandwidth is? (not the advertised rate) then set the bandwidth value at ~90% of this...

    I have seen similar threads a long time ago, but can't remember if there was a resolution - have you disabled the protocol filter that was known to cause problems

    Others found to be a bad NIC or driver module (r8168/r8169)
    The reply is currently minimized Show
  • Accepted Answer

    Wednesday, January 08 2014, 12:25 PM - #Permalink
    Resolved
    0 votes
    Unlikely, but can you check his NIC drivers:
    lspci | grep Eth
    grep eth /etc/modprobe.conf
    uname -r
    The reply is currently minimized Show
Your Reply