Discussion:
[Ryu-devel] Fast Failover in mininet not working !
陳家錡
2017-07-12 18:06:36 UTC
Permalink
Hi,

I am trying to use mininet to test my fast failover app.
(with Openflow 1.3, Open vSwitch 2.6.1)

My testing topology is shown as follow.

/ 2 \ / 5 \
h1 - 1 4 7 - h2
\ 3 / \ 6 /



It is fine when I using mininet to pingall.
(The path is *s1-s2-s4-s5-s7*)

But if I using mininet with command *link s1 s2 down*
I think it should trigger the ff and reroute from s1-s2 to s1-s3
However, it's not.
(I pretty sure the rule is matched and go to the group table,
dl_dst=00:00:00:00:00:02 actions=group:2)

The following is my group table in s1.
​group_id=2,type=ff,
bucket=watch_port:2,watch_group:2,actions=output:2,
bucket=watch_port:3,watch_group:2,actions=output:3

group_id=1,type=ff,
bucket=watch_port:1,watch_group:1,actions=output:1

How should I deal with it?
Is mininet should set some options?


Thanks for your help.
I truly appreciate it.

Rex.
--
[image: photo]

<http://www.facebook.com/xPlusSeven> <http://github.com/rex5207>
<http://www.instagram.com/ab_ab533/>
*陳家錡 Rex*
國立枅華倧孞資蚊工皋孞系碩士班

03-5715131#33564
***@gmail.com
高速網路寊驗宀 <http://hsnl.cs.nthu.edu.tw/>
枅華倧孞綜合二通705宀
<https://www.google.com/maps/search/%E6%B8%85%E8%8F%AF%E5%A4%A7%E5%AD%B8%E7%B6%9C%E5%90%88%E4%BA%8C%E9%A4%A8/@24.794716,120.992377,17.5z>
Fujimoto Satoshi
2017-07-13 02:21:18 UTC
Permalink
Hi!

I think fast-failover in s1 was succeeded.
But, for the packet from h2 to h1,
s4 may send it to s2, because s4 cannot notice the link between s1 and
s2 is down.
So, the packet from h2 to h1 is dropped in s2, and ping failed.

To avoid this, how about you have a link between s2 and s3?
And put a fast-failover action to s2 so that s2 send packets to s3 if
the link between s1 and s2 is down.
It can be said for s5, too.


Thanks,
Fujimoto
Post by 陳家錡
Hi,
I am trying to use mininet to test my fast failover app.
(with Openflow 1.3, Open vSwitch 2.6.1)
My testing topology is shown as follow.
/ 2 \ / 5 \
h1 - 1 4 7 - h2
\ 3 / \ 6 /
It is fine when I using mininet to pingall.
(The path is *s1-s2-s4-s5-s7*)
But if I using mininet with command *_link s1 s2 down_*
I think it should trigger the ff and reroute from s1-s2 to s1-s3
However, it's not.
(I pretty sure the rule is matched and go to the group table,
dl_dst=00:00:00:00:00:02 actions=group:2)
The following is my group table in s1.
​group_id=2,type=ff,
bucket=watch_port:2,watch_group:2,actions=output:2,
bucket=watch_port:3,watch_group:2,actions=output:3
group_id=1,type=ff,
bucket=watch_port:1,watch_group:1,actions=output:1
How should I deal with it?
Is mininet should set some options?
Thanks for your help.
I truly appreciate it.
Rex.
--
photo
<http://www.facebook.com/xPlusSeven> <http://github.com/rex5207>
<http://www.instagram.com/ab_ab533/>
*陳家錡 Rex*
國立枅華倧孞資蚊工皋孞系碩士班
03-5715131#33564 <tel:03-5715131#33564>
高速網路寊驗宀 <http://hsnl.cs.nthu.edu.tw/>
枅華倧孞綜合二通705宀
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Ryu-devel mailing list
https://lists.sourceforge.net/lists/listinfo/ryu-devel
陳家錡
2017-07-13 02:36:05 UTC
Permalink
Dear Fujimoto,

I think the problem you mentioned occurred in my app indeed.
I will try to change my app or the topology to deal with this situation.

Thanks for your help, again.
I truly appreciate it.

Rex.
Post by Fujimoto Satoshi
Hi!
I think fast-failover in s1 was succeeded.
But, for the packet from h2 to h1,
s4 may send it to s2, because s4 cannot notice the link between s1 and s2
is down.
So, the packet from h2 to h1 is dropped in s2, and ping failed.
To avoid this, how about you have a link between s2 and s3?
And put a fast-failover action to s2 so that s2 send packets to s3 if the
link between s1 and s2 is down.
It can be said for s5, too.
Thanks,
Fujimoto
Hi,
I am trying to use mininet to test my fast failover app.
(with Openflow 1.3, Open vSwitch 2.6.1)
My testing topology is shown as follow.
/ 2 \ / 5 \
h1 - 1 4 7 - h2
\ 3 / \ 6 /
It is fine when I using mininet to pingall.
(The path is *s1-s2-s4-s5-s7*)
But if I using mininet with command *link s1 s2 down*
I think it should trigger the ff and reroute from s1-s2 to s1-s3
However, it's not.
(I pretty sure the rule is matched and go to the group table,
dl_dst=00:00:00:00:00:02 actions=group:2)
The following is my group table in s1.
​group_id=2,type=ff,
bucket=watch_port:2,watch_group:2,actions=output:2,
bucket=watch_port:3,watch_group:2,actions=output:3
group_id=1,type=ff,
bucket=watch_port:1,watch_group:1,actions=output:1
How should I deal with it?
Is mininet should set some options?
Thanks for your help.
I truly appreciate it.
Rex.
--
[image: photo]
<http://www.facebook.com/xPlusSeven> <http://github.com/rex5207>
<http://www.instagram.com/ab_ab533/>
*陳家錡 Rex*
國立枅華倧孞資蚊工皋孞系碩士班
03-5715131#33564
高速網路寊驗宀 <http://hsnl.cs.nthu.edu.tw/>
枅華倧孞綜合二通705宀
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
--
[image: photo]

<http://www.facebook.com/xPlusSeven> <http://github.com/rex5207>
<http://www.instagram.com/ab_ab533/>
*陳家錡 Rex*
國立枅華倧孞資蚊工皋孞系碩士班

03-5715131#33564
***@gmail.com
高速網路寊驗宀 <http://hsnl.cs.nthu.edu.tw/>
枅華倧孞綜合二通705宀
<https://www.google.com/maps/search/%E6%B8%85%E8%8F%AF%E5%A4%A7%E5%AD%B8%E7%B6%9C%E5%90%88%E4%BA%8C%E9%A4%A8/@24.794716,120.992377,17.5z>
Loading...