User Tools

Site Tools


doc:uci:qos

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
Next revision
Previous revision
doc:uci:qos [2012/09/14 05:43]
zuirch
doc:uci:qos [2015/11/29 00:14] (current)
KyleS
Line 1: Line 1:
 +====== Quality of Service (qos-scripts) configuration ======
 +This is the documentation for the UCI configuration file ''/​etc/​config/​qos''​. It is used by the package ''​qos-scripts''​ only.
  
 +| {{:​meta:​icons:​tango:​48px-emblem-important.svg.png?​nolink}} | Do NOT install multiple QoS-packages simultaneously! Uninstall the old package before installing a new one. \\ There are at least two other QoS/ToS packages in the OpenWrt repositories regarding: ''​[[https://​github.com/​openwrt/​packages/​blob/​master/​net/​sqm-scripts|sqm-scripts]]''​ and ''​[[https://​github.com/​openwrt/​packages/​tree/​master/​net/​wshaper|wshaper]]''​. They do NOT use this file.\\ \\ ''​sqm-scripts''​ is the most modern and has Luci support. Configuration advice for it can be found at [[http://​www.bufferbloat.net/​projects/​cerowrt/​wiki/​Setting_up_SQM_for_CeroWrt_310]] \\ ''​qos-scripts''​ is written in AWK/shell script and uses [[doc/​howto/​packet.scheduler/​sch_hfsc]] and [[doc/​howto/​packet.scheduler/​sch_fq_codel]]\\ ''​wshaper''​ uses [[doc/​howto/​packet.scheduler/​sch_sfq]] [[doc/​howto/​packet.scheduler/​sch_htb]] [[doc/​howto/​packet.scheduler/​act_police]];​ [[http://​lartc.org/​wondershaper/​]] (Last release has been in 2002, so it is mostly unmaintained) \\ \\ For help writing your own script please see [[doc:​howto:​packet.scheduler:​packet.scheduler|Traffic Control on OpenWrt: configuring the Linux Network Scheduler]]. |
 +
 +| {{:​meta:​icons:​tango:​dialog-information.png?​nolink}} | You can browse the scripts here: ''​[[https://​dev.openwrt.org/​browser/​trunk/​package/​network/​config/​qos-scripts|qos-scripts]]''​\\ There is direct LuCI-support for ''​qos-scripts''​ called: ''​luci-app-qos''​.\\ NOTE: ''​luci-app-qos''​ won't start until you enable the ''​qos''​ Initscript within the System-->​Startup tab as well as enable qos under Network-->​QoS |
 +
 +
 +| {{:​meta:​icons:​tango:​48px-outdated.svg.png?​nolink}} | As of [[https://​dev.openwrt.org/​changeset/​31759|r31759]] ''​qos-scripts''​ replaced sfq/red with fq_codel to massively improve latency under load. \\ \\ As of [[https://​dev.openwrt.org/​changeset/​25641/​trunk|r25641]] ''​qos-scripts''​ dropped the use of IMQ (package ''​iptables-mod-imq''​ – Intermediate Queueing Device). Its successor is [[http://​www.linuxfoundation.org/​collaborate/​workgroups/​networking/​ifb|IFB (Intermediate Functional Block device)]], (requires package: ''​kmod-ifb''​ and the scheduler action //​[[https://​dev.openwrt.org/​browser/​trunk/​package/​iproute2/​patches/​200-act_connmark.patch?​rev=25639|act_connmark]]//​ included).|
 +
 +
 +
 +===== Sections =====
 +A minimal QoS configuration usually consists of:
 +  * one //​interface//​ section
 +  * some //rules// allocating packets to at least two buckets
 +  * //​configuration//​ of the buckets.
 +
 +==== Interface ====
 +Each Interface can have its own buffer. The ''​interface''​ section declares global characteristics of the connection on which the specified interface is communicating. The following options are defined within this section:
 +
 +<​code>​
 +config interface dsl
 +        option enabled ​     1
 +        option classgroup ​ "​Default"​
 +        option overhead ​    1
 +        option upload ​      512
 +        option download ​    4096
 +</​code>​
 +
 +^ Name ^ Type ^ Required ^ Default ^ Description ^
 +| ''​enabled''​ | boolean | yes | ''​1''​ | Enable/​Disable QoS |
 +| ''​classgroup''​ | string | yes | ''​Default''​ | Specify ''​classgroup''​ used for this interface (see description of ''​classgroup''​ below) |
 +| ''​overhead''​ | boolean | yes | ''​1''​ | decrease upload and download ratio to prevent link saturation |
 +| ''​download''​ | integer | yes | ''​4096''​ | Download limit in ''​kilobits/​second''​ |
 +| ''​upload''​ | integer | yes | ''​512''​ | Upload limit in ''​kilobits/​second''​ |
 +
 +==== Rules ====
 +Each ''​classify''​ section defines one group of packets and which target (i.e. bucket) this group belongs to. All the packets share the bucket specified.
 +
 +^ Name ^ Type ^ Required ^ Default ^ Description ^
 +| ''​target''​ | bucket | yes | //(none)// | The four defaults are: ''​Priority,​ Express, Normal''​ and ''​Bulk''​ |
 +| ''​proto''​ | string | no | ''​0''​ | Packets matching this protocol belong to the bucket defined in target |
 +| ''​srchost''​ | string | no | //(none)// | Packets matching this source host(s) (single IP or in CIDR notation) belong to the bucket defined in target |
 +| ''​dsthost''​ | string | no | //(none)// | Packets matching this destination host(s) (single IP or in CIDR notation) belong to the bucket defined in target |
 +| ''​ports''​ | integer | no | //(none)// | Packets matching this, belong to the bucket defined in target |
 +| ''​srcports''​ | integer | no | //(none)// | Packets matching this, belong to the bucket defined in target |
 +| ''​dstports''​ | integer | no | //(none)// | Packets matching this, belong to the bucket defined in target |
 +| ''​portrange''​ | integer | no | //(none)// | Packets matching this, belong to the bucket defined in target |
 +| ''​pktsize''​ | integer | no | //(none)// | Packets matching this, belong to the bucket defined in target |
 +| ''​tcpflags''​ | string | no | //(none)// | Packets matching this, belong to the bucket defined in target |
 +| ''​mark''​ | string | no | //(none)// | Packets matching this, belong to the bucket defined in target |
 +| ''​connbytes''​ | int | no | //(none)// | Packets matching this, belong to the bucket defined in target |
 +| ''​tos''​ | string | no | //(none)// | Packets matching this, belong to the bucket defined in target |
 +| ''​dscp''​ | string | no | //(none)// | Packets matching this, belong to the bucket defined in target |
 +| ''​direction''​ | string | no | //(none)// | Packets matching this traffic direction (''​in''​ or ''​out''​) belong to the bucket defined in target |
 +
 +Note: the already broken '​layer7'​ option was removed by r45425.
 +==== Classgroup ====
 +As we can have more then one interface, we can have more then one classgroup.
 +
 +<​code>​
 +config classgroup "​Default"​
 + option classes ​     "​Priority Express Normal Bulk"
 + option default ​     "​Normal"​
 +</​code>​
 +
 +^ Name ^ Type ^ Required ^ Default ^ Description ^
 +| ''​classes''​ | bucket names | yes | //(none)// | Specifies the list of  names of //​classes// ​ |
 +| ''​default''​ | bucket name | yes | //(none)// | Defines which //class// is considered default |
 +
 +
 +==== Classes ====
 +Each Bucket has its own configuration.
 +
 +Example:
 +<​code>​
 +config class "​Normal"​
 + option packetsize ​ 1500
 + option packetdelay 100
 + option avgrate ​    10
 + option priority ​   5
 +</​code>​
 +
 +^ Name ^ Type ^ Required ^ Default ^ Description ^
 +| ''​packetsize''​ | integer | yes | //(none)// | in bytes |
 +| ''​packetdelay''​ | integer | yes | //(none)// | in ms |
 +| ''​maxsize''​ | integer | yes | //(none)// | in bytes |
 +| ''​avgrate''​ | integer | yes | //(none)// | Average rate for this class, value in % of bandwidth (this value uses for calculate vaues '​Nx'​ of '''​tc ... hfsc rt m1 N1 d N2 m2 N3'''​) |
 +| ''​limitrate''​ | integer | no  | 100 | Defines to how much percent of the available bandwidth this class is capped to, value in % |
 +| ''​maxsize''​ | integer | yes | //(none)// | in bytes |
 +| ''​priority''​ | integer | yes | //(none)// | in % |
 +
 +==== Classes (For Advanced Users) ====
 +Below is unverified technical breakdown of each /​etc/​config/​qos class parameters. Source: [[http://​pastebin.com/​YL55na2E]]
 +<​code>​
 +### Params:
 +#
 +# maxsize:
 +#       ​limits packet size in iptables rule
 +#
 +# avgrate: (note: sum(avgrates) ~ 100)
 +#       rt m1 = avgrate / sum (avgrate) * max_bandwidth
 +#       rt m2 = avgrate * max_bandwidth / 100
 +#       ls m1 = rt m1
 +#
 +# packetsize & packetdelay:​ (only works if avgrate is present)
 +#       rt d = max( packetdelay,​ 'time required for packetsize to transfer'​ ) (smaller ps -> smaller d)
 +#       ls d = rt d
 +#
 +# priority:
 +#       ls m2 = priority / sum (priority) * max_bandwidth
 +#
 +# limitrate:
 +#       ul rate = limitrate * max_bandwidth / 100
 +</​code>​
 +
 +===== Quick start guide =====
 +
 +1. Install the qos-scripts package:
 +
 +<​code>​
 +opkg install qos-scripts
 +</​code>​
 +
 +2. Basic configuration using UCI command line:
 +
 +<​code>​
 +uci set qos.wan.upload=1000 ​           # Upload speed in kBits/s
 +uci set qos.wan.download=16000 ​        # Download speed in kBits/s
 +uci set qos.wan.enabled=1
 +uci commit qos</​code>​
 +
 +3. Start it and look for error output and test):
 +
 +<​code>​
 +/​etc/​init.d/​qos start
 +</​code>​
 +
 +4. Make script run at every boot up:
 +
 +<​code>​
 +/​etc/​init.d/​qos enable
 +</​code>​
 +
 +===== Troubleshooting =====
 +
 +(Last updated for: Barrier Breaker 14.07)
 +
 +If your QoS doesn'​t seem to be working, it may be an error or typo in the config file is preventing it from loading properly.
 +
 +  * Check //enabled// is set to 1 in ///​etc/​config/​qos//​(!)
 +
 +  * Run //​iptables-save//​ and check there are lines near the top prefixed with either //-A qos_Default//​ or //-A qos_Default_ct//,​ and featuring the //​--set-xmark//​ directive. Here's an example:
 +<​code>​
 +-A qos_Default -p tcp -m mark --mark 0x0/0xf0 -m tcp --sport 1024:65535 --dport 1024:65535 -j MARK --set-xmark 0x44/0xff
 +</​code>​
 +The //​--set-xmark//​ is what flags the packet so it is picked up the traffic control subsystem.
 +
 +  * Look at the generated traffic control qdisc settings by running:
 +
 +<​code>​
 +tc qdisc
 +</​code>​
 +
 +The default (ie no-QoS-applied) values for any interface look like this:
 +
 +<​code>​
 +qdisc fq_codel 0: dev eth0 root refcnt 2 limit 1024p flows 1024 quantum 300 target 5.0ms interval 100.0ms ecn
 +</​code>​
 +
 +... Any interface with only a single qdisc line printed, showing the same settings as this line (this one is for //dev eth0//), indicates no QoS on that interface.
 +
 +Network interfaces with QoS enabled will have multiple qdisc lines printed, each corresponding to a QoS class, etc.
 +
 +  * If the printed qdisc settings don't seem to be correct, you can preview the //tc// commands generated from the OpenWRT ///​etc/​config/​qos//​ by running:
 +
 +<​code>​
 +/​usr/​lib/​qos/​generate.sh interface wan
 +</​code>​
 +
 +(Replace '​wan'​ with the OpenWRT interface name you're debugging, as given in the ///​etc/​config/​qos//​ file.)
 +
 +This should print a series of //insmod// and //tc// commands used to set up the QoS subsystem. You can debug any errors caused by running these commands by running:
 +
 +<​code>​
 +/​usr/​lib/​qos/​generate.sh interface wan | sh -x
 +</​code>​
 +
 +(Note //-x// option which tells //sh// to print each line as it is executed.)
 +
 +The output of ///​usr/​lib/​qos/​generate.sh//​ is normally executed automatically as part of ///​etc/​hotplug.d/​iface/​10-qos//​.
 +
 +===== txqueuelen =====
 +//Recent versions of trunk uses [[http://​www.bufferbloat.net/​projects/​codel/​wiki|CoDel]] (pronounced:​ Coddle), so this should not be needed. [[http://​www.bufferbloat.net/​projects/​bloat/​wiki/​Linux_Tips#​Reduce-transmit-queue-length|bufferbloat.net]]//​
 +
 +/* This whole section seems a bit rambling /Cybjit */
 +
 +( Note: after i know about bufferbloat - websearchd and many misguided users are asking about //raising// txqueuelen, or how to set it, and this was a recommended read posted for them -, and that just enabling QoS and setting up the rules i wanted didnt produce results, i thought of reduce txqueuelen from the default size of 1000 packets. On most SOHO applications the upload speed is much-much lower than the interface speed (100M or 1G), and it's written that the def buffer is tailored for enterprise size usage. I used values of 90 where[only] it was 1000 and it is wonderfully responsive and effective now, on a 256k connection. I also raised lenghts of 3 and 5 to 20 as i imagined it might be hard to do any queuing/​shaping on such ultra short buffers. Leave the 0-s as 0. Btw, the 90 came by, that with 1500byte packet size and 256kbit speed, it takes just half a sec to empty it. Well, i mistakenly took mtu as bits so it's 4 second really but i didnt have the courage for a drastically smaller than default value, 90 already seem so smallish, and it works nicely anyway, awesome lack of latency, jitter and packetloss on the other machine, according to [[http://​pingtest.net]],​ before of this there was no difference between mine and that. Note that this time doesnt correspond to ping values. The point is to allow built-in TCP congestion control //to work// to reduce "​spamming",​ set speeds as it was envisioned, and keep things leveled out rather than fluctuating widely. While this doesnt directly effects QoS-ing itself, it is extremely beneficial, even essential, to the results usually expected from employing it. It sets a foundation, a healthy network environment over which QoS to function.
 +
 +First we list the stuff with
 +<​code>​ifconfig</​code>​.
 +Then apply for example
 +<​code>​ifconfig eth0 txqueuelen 90</​code>​ and
 +<​code>​ifconfig pppoe-wan txqueuelen 20</​code>​ etc accordingly. Finally an
 +<​code>​uci commit
 +/​etc/​init.d/​network reload</​code>​ might be needed, i did a network restart too after this.)
 +
 +===== Types and Groups =====
 +
 +The ''​qos-scripts''​ package didn't come with documentation and there has been some confusion about its features, among users. The information in this section comes straight from nbd (the developer), so it should come a long way to clearing some confusion on two major issues.
 +
 +The biggest item of contention was which group setting gives better performance,​ **Priority** or **Express**. As it turns out, it depends on the application. **Priority** boosts low-bandwidth small frames, such as TCP-ACKs and DNS more than Express. **Express** is for prioritizing bigger frames, which would include stuff like VoIP (port 5060).
 +
 +Another biggie was the exact meaning of each type. Types are necessary for connection tracking. By default, **Classify** is not run on a connection that had already been assigned a traffic class, so it is the initial connection-tracked classification. **Reclassify** can override the traffic class per packet, without altering the connection tracking mark. **Default** is a fall-back for everything that has not been marked by Classify/​Reclassify. Rules get processed by type first (Classify gets processed first, then Reclassify and finally Default) and then based on the order in the configuration file (top to bottom).
 +
 +===== Tags =====
 +{{tag>​QoS}}