15
1. ETH=$1
2. LATENCY=$2
3. LOSS=$3
4. JITTER=$4
5. BW=$5
6. sudo /sbin/tc qdisc del dev eth0 root
7. sudo /sbin/tc qdisc add dev eth0 root handle 1: netem delay $LATENCY $JITTER 
8. sudo /sbin/tc qdisc add dev eth0 parent 1:1 handle 10: netem loss $LOSS
9. sudo /sbin/tc qdisc add dev eth0 parent 10:1 handle 20: htb default 1
10.sudo /sbin/tc class add dev eth0 parent 20: classid 0:1 htb rate $BW ceil $BW
11.sudo /sbin/tc qdisc show

The above code results in:

RTNETLINK answers :No such file or directory

error on line 8,9,10 upon execution.

tshepang
  • 12,111
  • 21
  • 91
  • 136
kalpan bhargav
  • 159
  • 1
  • 1
  • 3
  • 1
    duplicate of http://stackoverflow.com/questions/4803069/rtnetlink-answers-no-such-file-or-directory?rq=1 ? – David Dombrowsky Apr 26 '13 at 10:59
  • Have you tried using `tc qdisc list` to see what is in effect? This error sometimes means you are referring to a non-existent entry. – Neil Nov 14 '13 at 17:45
  • For anyone else who gets this on `tc qdisc del dev ...` and runs acress this question, the problem can also be that everything was already deleted. – Cel Skeggs Jul 08 '15 at 21:10

5 Answers5

12

The problem is with missing kernel modules and/or kernel support. Make sure you can run 

modprobe sch_netem

I also had to rebuild the kernel after I enabled because there's something built-in that is enabled after you enable network emulation.

See also http://forums.fedoraforum.org/showthread.php?t=285408 and http://www.linuxfoundation.org/collaborate/workgroups/networking/netem

David Dombrowsky
  • 1,655
  • 3
  • 20
  • 32
  • 22
    BTW, in my case I was also getting a frustrating error that looks identical to this one. But I was trying to use `tc qdisc change...` when I meant to use `tc qdisc add...`. The error message is the same. Go figure. – David Dombrowsky Apr 26 '13 at 15:27
  • 2
    I can run modprobe `sch_netem` command, it doesn't give any errors. Also `lsmod | grep -i sch_netem` returns `sch_netem 17306 0`. But still `qdisk change` command gives me error: `RTNETLINK answers :No such file or directory` – Aqeel Ashiq Aug 10 '15 at 12:14
  • 1
    @DavidDombrowsky, thanks for spotting it, I was just trying the second example line from http://www.linuxfoundation.org/collaborate/workgroups/networking/netem, which "change"s instead of "add"ing. – Jaime Hablutzel Aug 31 '15 at 23:33
6

RTNETLINK answers :No such file or directory comes for 2 reason.

  • Executing tc qdisc with wrong options
  • Or Kernel module sch_netem is missing

So check first sch_netem is installed in your machine by using command lsmod -l. If it is missing install using command modprobe sch_netem.

Even after installing sch_netem, if you get this same error means you are executing with wrong option. For example you can add a impairment on a network interface using ip qdisc add dev .... Similarly for removing this you can do ip qdisc delete dev.... If you are trying to delete an impairment without any prior add means, then also you get this error. And also for wrong options also you get this error.

rashok
  • 12,790
  • 16
  • 88
  • 100
  • 1
    I ran into this and a missing module was indeed the culprit. In my case I needed restart after kernel upgrade had cleared out all the modules for the kernel verison I was currently running (this was on raspbian). – morganwahl Feb 23 '19 at 17:37
2

Your handles and parent handles don't match. e.g change to:

7. sudo /sbin/tc qdisc add dev eth0 root handle 1:1 netem delay $LATENCY $JITTER 
8. sudo /sbin/tc qdisc add dev eth0 parent 1:1 handle 10:1 netem loss $LOSS
9. sudo /sbin/tc qdisc add dev eth0 parent 10:1 handle 20:1 htb default 1
10.sudo /sbin/tc class add dev eth0 parent 20:1 classid 0:1 htb rate $BW ceil $BW

and it should work.

Kimvais
  • 38,306
  • 16
  • 108
  • 142
1

I am running Fedora Core 27. In order to get the module installed, I had to use

sudo dnf install install kernel-modules-extra

That's not intuitively obvious.

Jeff Silverman
  • 692
  • 1
  • 8
  • 15
0

It's an old question, but the problem still occur, so.

To understand more about the problem you can check this link and this .

As a workaround to solve the problem, flush the interface config with :

Please read this before executing the command, you maybe locked when you are remotely connecting through the interface you are working on.

sudo ip addr flush dev eth0
mbareck7
  • 31
  • 3