| *** doonhammer has quit IRC | 00:15 | |
| *** janki has joined #networking-sfc | 04:12 | |
| *** mohankumar has joined #networking-sfc | 06:02 | |
| *** mbuil has joined #networking-sfc | 06:52 | |
| *** mardim has quit IRC | 08:20 | |
| *** mardim has joined #networking-sfc | 08:32 | |
| mardim | phuoc, hello | 08:45 |
|---|---|---|
| mardim | did you tried to create several chains which are attached to the same SF ? | 08:45 |
| mardim | and it worked ? | 08:45 |
| phuoc | mardim, it worked | 08:48 |
| mardim | how so ? | 08:48 |
| mardim | phuoc, | 08:48 |
| mardim | what was your topology ? | 08:49 |
| phuoc | I create 2 VNFFG | 08:49 |
| phuoc | first VNFFG goes through VNF1 VNF2 | 08:49 |
| phuoc | the second go through VNF2 | 08:49 |
| mardim | hmmm | 08:49 |
| mardim | I think this is not testing what we are talking about | 08:50 |
| phuoc | in networking-sfc domain, 2 port chain can have the same port pair group | 08:50 |
| mardim | because you are using the VNF2 which is the second in VNFFG | 08:50 |
| mardim | VNFFG_one | 08:50 |
| phuoc | I tested this scenario by using simple tcpdump command | 08:50 |
| mardim | MAybe you should try to use the the VNF1 as the common SF | 08:51 |
| mardim | and not VNF2 | 08:51 |
| mardim | Because in your case | 08:51 |
| mardim | when the traffic goes out form the ssecond VNF2 | 08:51 |
| mardim | it doesn't need to steered to another SF | 08:52 |
| mardim | so it exits the SFC domain | 08:52 |
| phuoc | I will test with VNF1 as common | 08:52 |
| mardim | so no SFF is involved after the VNF2 | 08:52 |
| mardim | bcafarel, ^ | 08:52 |
| mardim | bcafarel, What do you think ? ^ | 08:53 |
| phuoc | when I test, I need to make sure flow-classifiers in 2 VNFFGs are not conflicted | 08:54 |
| phuoc | I don't know to much about ovs, but I think whenever traffic go in/out VNF | 08:55 |
| phuoc | it will classifiers by ovs flow rules | 08:56 |
| mardim | phuoc, bcafarel I think that the traffic is not reclassified when goes out of the NVF | 08:57 |
| mardim | VNF* | 08:57 |
| mardim | if it reclassifies it's ok | 08:57 |
| mardim | but I think this is not the case here | 08:57 |
| phuoc | I think traffic need to re-classifier, because there are no NSH now xD | 09:00 |
| bcafarel | quick checking through code, we have https://github.com/openstack/networking-sfc/blob/master/networking_sfc/db/sfc_db.py#L266 check | 09:02 |
| bcafarel | so in networking-sfc terms, a PPG cannot be attached to more than one chain | 09:03 |
| mardim | phuoc, Yes it needs to reclassify | 09:08 |
| mardim | but this is not the current implementatiin in networking-sfc | 09:08 |
| phuoc | bcafarel, I think this function is used to check 2 port chain don't have the same path | 09:55 |
| phuoc | if 2 port chains have different list of ppg | 09:57 |
| phuoc | it will be allowed | 09:57 |
| *** mangelm has joined #networking-sfc | 10:16 | |
| *** mbuil has quit IRC | 10:54 | |
| *** mbuil has joined #networking-sfc | 12:49 | |
| *** mangel2 has joined #networking-sfc | 14:17 | |
| *** mangelm has quit IRC | 14:20 | |
| *** mangel3 has joined #networking-sfc | 14:24 | |
| *** mangel2 has quit IRC | 14:26 | |
| *** mohankumar has quit IRC | 14:38 | |
| *** geb has quit IRC | 14:45 | |
| *** geb has joined #networking-sfc | 14:46 | |
| *** mangel3 has quit IRC | 15:31 | |
| *** mangel3 has joined #networking-sfc | 15:32 | |
| *** doonhammer has joined #networking-sfc | 15:39 | |
| *** doonhammer has left #networking-sfc | 15:41 | |
| *** mbuil has quit IRC | 16:51 | |
| *** janki has quit IRC | 17:29 | |
| *** mangel3 has quit IRC | 19:00 | |
| *** doonhammer has joined #networking-sfc | 20:19 | |
| *** doonhammer has quit IRC | 22:34 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!