Yep, this is a well-known issue/limitation:

https://bugs.opendaylight.org/show_bug.cgi?id=3345 
<https://bugs.opendaylight.org/show_bug.cgi?id=3345>

BR/Luis

> On Jun 9, 2016, at 5:26 PM, Jamo Luhrsen <[email protected]> wrote:
> 
> Luis, can you confirm for Selvakumar that l2switch is or is not able to handle
> these link down events?
> 
> I know you excluded these test cases from our CSIT in the past [0] because 
> this
> functionality is not really implemented.  I couldn't find the email about it,
> but I vaguely remember you sending one.
> 
> JamO
> 
> [0]  https://git.opendaylight.org/gerrit/#/c/34103/
> 
> 
> 
> On 06/08/2016 11:09 PM, Selvakumar Murugasekaran wrote:
>> Hello All,
>> 
>> Good day to you all.
>> 
>> Could any one help me to find the solution for the problem "ODL beryllium 
>> controller does not re-provisioning the data
>> traffic when one of the active link is down" as following,
>> 
>> Let me know if you need any more information.
>> 
>> 
>> Thanks and Regards,
>> Selvakumar Murugasekaran.
>> 
>> 
>> -------- Original Message --------
>> Subject:     Reg: ODL beryllium controller does not re-provisioning the data 
>> traffic when one of the active link is down.
>> Date:        Fri, 03 Jun 2016 15:16:54 +0530
>> From:        Selvakumar Murugasekaran 
>> <[email protected]>
>> To:  Daniel Farrell <[email protected]>
>> CC:  Jamo Luhrsen <[email protected]>, [email protected], 
>> "[email protected]"
>> <[email protected]>, "Bhuvan (Veryx Technologies)" 
>> <[email protected]>
>> 
>> 
>> 
>> Hello,
>> 
>> I am doing the ODL beryllium controller "Network re-provisioning Time" test 
>> with *miniedit*(*2.2.0.1* version).
>> 
>> The ODL controller(distribution-karaf-0.4.0-Beryllium) is running in 
>> reactive mode. OVS version is *2.3.2*.
>> I created the topology with 4 switch's and 2 host's(Refer: miniedit.jpg). 
>> The connection details are following in mininet,
>> 
>> mininet> net
>> h1 h1-eth0:s1-eth3
>> h2 h2-eth0:s4-eth3
>> s1 lo:  s1-eth1:s2-eth1 s1-eth2:s3-eth2 s1-eth3:h1-eth0
>> s2 lo:  s2-eth1:s1-eth1 s2-eth2:s4-eth2
>> s4 lo:  s4-eth1:s3-eth1 s4-eth2:s2-eth2 s4-eth3:h2-eth0
>> s3 lo:  s3-eth1:s4-eth1 s3-eth2:s1-eth2
>> c0
>> mininet> xterm h1
>> 
>> I started the ping(Ex: ping 10.0.0.2) from interface h1-eth0(10.0.0.1) to 
>> h2-eth0(10.0.0.2) .
>> ODL controller sends the Flow_Mod to OVS and ping data flows via h1-eth0 --> 
>> s1-eth3 --> S1-eth2 --> s3-eth2 ---> S3-eth1-->
>> S4-eth1 --> s4-eth3 --> h2-eth0.
>> 
>> The ping 10.0.0.2 is not succeed when I down the active path link between 
>> s3-eth1 - S4-eth1.
>> ODL controller does not re-provisioning data to redundant active path via 
>> h1-eth0 --> s1-eth3 --> S1-eth1 --> s2-eth1 --->
>> S2-eth2--> S4-eth2 --> s4-eth3 --> h2-eth0.
>> The switch S1, S2 & S3 send the packet-in(ARP) message to controller but the 
>> ODL controller send the packet-out(ARP) message
>> with old in-port and out-port to switch s1 & S3 only.
>> 
>> Could you please clarify Whether I am missing any configuration to 
>> re-provisioning the data traffic ?
>> 
>> Note: Only one time I saw the ODL controller re-provisioning the path out of 
>> 6 times.
>> 
>> Let me know if you need any more information.
>> 
>> 
>> Thanks and Regards,
>> Selvakumar Murugasekaran.
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> 

_______________________________________________
L2switch-dev mailing list
[email protected]
https://lists.opendaylight.org/mailman/listinfo/l2switch-dev

Reply via email to