Re: [LARTC] overlimit in tc statistics

From: Sergey Smirnov <svs@leaves.ru>
To: lartc@vger.kernel.org
Subject: Re: [LARTC] overlimit in tc statistics
Date: Mon, 07 Jul 2003 13:31:48 +0000	[thread overview]
Message-ID: <marc-lartc-105758477704881@msgid-missing> (raw)
In-Reply-To: <marc-lartc-105731795218938@msgid-missing>

Stef Coene wrote:
> On Friday 04 July 2003 13:18, Sergey Smirnov wrote:
> 
>>There is overlimit in queueing disciplines statistics:
>>
>># tc -s qdisc show dev eth1|grep overlimit
>>  Sent 2405317 bytes 17168 pkts (dropped 0, overlimits 0)
>>  Sent 145699 bytes 872 pkts (dropped 0, overlimits 0)
>>  Sent 0 bytes 0 pkts (dropped 0, overlimits 0)
>>  Sent 25374 bytes 160 pkts (dropped 0, overlimits 0)
>>  Sent 842 bytes 12 pkts (dropped 0, overlimits 0)
>>  Sent 0 bytes 0 pkts (dropped 0, overlimits 0)
>>  Sent 0 bytes 0 pkts (dropped 0, overlimits 0)
>>  Sent 0 bytes 0 pkts (dropped 0, overlimits 0)
>>  Sent 0 bytes 0 pkts (dropped 0, overlimits 0)
>>  Sent 38061 bytes 73 pkts (dropped 0, overlimits 0)
>>  Sent 35727 bytes 90 pkts (dropped 0, overlimits 0)
>>  Sent 2651738 bytes 18380 pkts (dropped 0, overlimits 318)
>>
>>But traffic classes statistics without overlimits:
>># tc -s class show dev eth1|grep overlimit
>>  Sent 2373030 bytes 16963 pkts (dropped 0, overlimits 0)
>>  Sent 35727 bytes 90 pkts (dropped 0, overlimits 0)
>>  Sent 2616582 bytes 18159 pkts (dropped 0, overlimits 0)
>>  Sent 38061 bytes 73 pkts (dropped 0, overlimits 0)
>>  Sent 0 bytes 0 pkts (dropped 0, overlimits 0)
>>  Sent 0 bytes 0 pkts (dropped 0, overlimits 0)
>>  Sent 0 bytes 0 pkts (dropped 0, overlimits 0)
>>  Sent 0 bytes 0 pkts (dropped 0, overlimits 0)
>>  Sent 0 bytes 0 pkts (dropped 0, overlimits 0)
>>  Sent 25374 bytes 160 pkts (dropped 0, overlimits 0)
>>  Sent 842 bytes 12 pkts (dropped 0, overlimits 0)
>>  Sent 143548 bytes 861 pkts (dropped 0, overlimits 0)
>>
>>I don't understand this.
>>What does this mean?
> 
> We are maybe guru's, but this is not enough information.  What's your setup, 
> used commands, network topology, what you want to do, tests you did, full 
> output of you tc commands, .........
> 
> Stef
> 
O'k. I have Debian stable Linux router to Internet.
I want manage Internet traffic on it:
Increase priority and speed of some protocols and decrease for others.
I know that I can change this only for outgoing packets.
First question:
If I decrease priority and speed for local network interface
does this mean that priority and speed for Internet interface
for the same protocol also decreases?
For example:
eth0 local network
eth1 Internet

tc qdisc del dev eth0 root
tc qdisc add dev eth0 root handle 1 htb default 99
tc class add dev eth0 parent 1: classid 1:2 htb rate 256Mbit burst 15k
tc class add dev eth0 parent 1:2 classid 1:10 htb rate 64Kbit ceil 
128Kbit burst 15k prio 4
tc qdisc add dev eth0 parent 1:10 handle 10 sfq perturb 10
tc filter add dev eth0 parent 1:0 protocol ip prio 100 u32 match ip 
sport 80 0xffff match ip dst 192.168.1.0/24 classid 1:10

Does this means that incoming traffic on eth1 to port 80 will be in 
64Kbit-128Kbit with prio 4?
-- 
Sergey Smirnov

_______________________________________________
LARTC mailing list / LARTC@mailman.ds9a.nl
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/