Hey, I just found the bug. By default, switch won't send packets out to the
in-port, however, openflow has provided OFPP_INPUT_PORT to work around it.Then,
I just changed the original output port in group to OFPP_INPUT_PORT, and it
works.
Thanks!
On Wednesday, May 31, 2017 3:34 PM, Iwase Yusuke <[email protected]>
wrote:
Hi,
> Question: Is this "feature" a bug of Ryu-4.13 or OpenvSwitch-2.4.1 ?
Please confirm whether your flow entries and group entries are installed
as expected or not.
If all entries are installed, I guess it is the bug of your switch side.
> Topology:
> / s2 \
> h1--s1 ----| |---- s4--h2
> \ s3 /
>
> To prevent from link s3-s4 being down, one Fast-Failover Group g1 was added
> into s3. Group g1 consists of two buckets, the first bucket was used to
> forward packets to s3-s4 when link s3-s4 is normal, while the second bucket
> was used to forward packets to s3-s1 when link s3-s4 is down.
In this case, it is required that s1 can forward packets to s2 if packets
is sent back from s3, but the flows for this situation are installed?
Or, I guess you need additional link s2-s3 and forward packets to this
link if s3-s4 goes down.
Thanks,
Iwase
On 2017年05月30日 22:18, malo lk wrote:
> Test environment: Ryu-4.13, OpenvSwitch-2.4.1
>
> Topology:
> / s2 \
> h1--s1 ----| |---- s4--h2
> \ s3 /
>
> To prevent from link s3-s4 being down, one Fast-Failover Group g1 was added
> into s3. Group g1 consists of two buckets, the first bucket was used to
> forward packets to s3-s4 when link s3-s4 is normal, while the second bucket
> was used to forward packets to s3-s1 when link s3-s4 is down.
>
> Experiment: hit down command `h1 ping h2`, everything is normal when link
> s3-s4 is up. Then tear it down, the ping command won't work. In this
> situation, the port in the second
> bucket is the in-port of the ping pakcet, but group g1 didn't forward packets
> to the port in second bucket, i.e. in-port of ping packets.
>
> Verification: this time, I add another link to the above topology.
> / s2 \
> h1--s1 ----| |---- s4--h2
> \ s3 /
> \ s5 /
> Then, the second bucket was changed to forward packets to s5-s4 when link
> s3-s4 was down. Repeat the above experiment, everything is ok when link s3-s4
> is down, and the group
> indeed had forward packets to the s5-s4.
>
> Question: Is this "feature" a bug of Ryu-4.13 or OpenvSwitch-2.4.1 ?
>
> Any experience sharing or suggestions would be great.
> Thanks.
>
>
>
>
> ------------------------------------------------------------------------------
> 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
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/ryu-devel
>
------------------------------------------------------------------------------
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
[email protected]
https://lists.sourceforge.net/lists/listinfo/ryu-devel
------------------------------------------------------------------------------
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
[email protected]
https://lists.sourceforge.net/lists/listinfo/ryu-devel
