| *** uck_ has quit IRC | 00:10 | |
| *** uck has joined #tacker | 00:11 | |
| *** uck has quit IRC | 00:16 | |
| *** sripriya has joined #tacker | 00:23 | |
| *** yifei has quit IRC | 00:26 | |
| *** bobh has joined #tacker | 00:43 | |
| *** sripriya has quit IRC | 00:53 | |
| openstackgerrit | vishwanath jayaraman proposed openstack/tacker: Devref for event logging feature https://review.openstack.org/370428 | 01:48 |
|---|---|---|
| *** gongysh has joined #tacker | 01:57 | |
| openstackgerrit | gongysh proposed openstack/tacker: Device refactor part4: rename vm plugin and db into vnfm plugin and db https://review.openstack.org/363351 | 02:05 |
| gongysh | sridhar_ram, done with part4 | 02:08 |
| *** bobh has quit IRC | 02:10 | |
| *** s3wong has quit IRC | 02:14 | |
| *** gongysh has quit IRC | 02:16 | |
| *** gongysh has joined #tacker | 02:20 | |
| gongysh | sridhar_ram, hi | 02:23 |
| *** vishnoianil has quit IRC | 02:29 | |
| *** _achatterjee_ has quit IRC | 03:00 | |
| *** achatterjee has joined #tacker | 03:00 | |
| gongysh | trozet, hi | 03:13 |
| *** uck has joined #tacker | 03:19 | |
| gongysh | sridhar_ram, hi | 03:25 |
| sridhar_ram | gongysh: hi | 03:54 |
| gongysh | sridhar_ram, besides the part4 patch, I want to test vnffg, how to start? | 03:55 |
| sridhar_ram | gongysh: sure, here is a quick run down.. | 03:56 |
| sridhar_ram | start with ubuntu 16.04 | 03:56 |
| gongysh | sridhar_ram, thank you your slides on which I used some of them and indicated 'modified from PTL sridhar's slides'. | 03:56 |
| sridhar_ram | gongysh: you are welcome! | 03:56 |
| sridhar_ram | gongysh: need to make a better one for Barcelona | 03:57 |
| gongysh | sridhar_ram, I will not be able to be there, hope tacker will impress stacker. | 03:57 |
| sridhar_ram | gongysh: :) | 03:57 |
| sridhar_ram | For ffg: you need tacker with Stephen Wong + my integrations fixes | 03:58 |
| gongysh | so, the networking-sfc needs the OVS 2.5, right? | 03:58 |
| sridhar_ram | yes | 03:59 |
| gongysh | sridhar_ram, and then, what is the exact steps to test vnffg? | 03:59 |
| gongysh | I cannot find a devref or get started with vnffg. | 04:00 |
| *** hparekh has joined #tacker | 04:03 | |
| *** KanagarajM has joined #tacker | 04:06 | |
| gongysh | if we does not have one, after testing, I can get one uploaded. | 04:06 |
| *** KanagarajM_ has joined #tacker | 04:09 | |
| *** KanagarajM has quit IRC | 04:09 | |
| *** uck has quit IRC | 04:09 | |
| * sridhar_ram coming back after family interrupt | 04:12 | |
| sridhar_ram | gongysh: let me write this in a etherpad.. hang on | 04:12 |
| sridhar_ram | gongysh: btw, before i forget .. can you follow on uniqueness constraint patchset - https://review.openstack.org/329759 ? | 04:17 |
| sridhar_ram | gongysh: my opinion is db based constraint is ideal but we don't have time to reimplement this for Newton... | 04:17 |
| gongysh | sridhar_ram, ok | 04:18 |
| sridhar_ram | gongysh: .. and it is risky to allow duplicate VNFD names with VNFFG feature | 04:18 |
| *** KanagarajM_ has quit IRC | 04:18 | |
| *** gongysh has quit IRC | 04:18 | |
| sridhar_ram | gongysh: .. so i'd suggest we get this in mostly for the functional need | 04:18 |
| *** janki has joined #tacker | 04:26 | |
| *** neel has joined #tacker | 04:30 | |
| openstackgerrit | Sridhar Ramaswamy proposed openstack/tacker: VNFFG: neutron networking-sfc driver and plugin fixes https://review.openstack.org/347568 | 04:35 |
| *** sripriya has joined #tacker | 04:53 | |
| openstackgerrit | Merged openstack/tacker: Device refactor part4: rename vm plugin and db into vnfm plugin and db https://review.openstack.org/363351 | 05:09 |
| *** KanagarajM_ has joined #tacker | 05:16 | |
| *** sripriya has quit IRC | 05:22 | |
| *** vishnoianil has joined #tacker | 05:23 | |
| *** janki has quit IRC | 05:39 | |
| *** Ravikiran_K has joined #tacker | 05:45 | |
| openstackgerrit | Sridhar Ramaswamy proposed openstack/tacker: VNFFG: neutron networking-sfc driver and plugin fixes https://review.openstack.org/347568 | 06:00 |
| tung_doan | sridhar_ram: ping | 06:01 |
| sridhar_ram | tung_doan: pong | 06:01 |
| tung_doan | sridhar_ram: at the begging, ceilometer can trigger well (dkushwaha coulf confirm).. I suspect that the new version of ceilometer/aodhcreated some issue... | 06:02 |
| *** sripriya has joined #tacker | 06:03 | |
| tung_doan | sridhar_ram: i am scared that i dont have enough time to figure out what happend to ceilometer/aodh.. | 06:03 |
| sridhar_ram | repeating for sripriya... | 06:03 |
| sridhar_ram | sridhar_ram: at the begging, ceilometer can trigger well (dkushwaha coulf confirm).. I suspect that the new version of ceilometer/aodhcreated some issue... | 06:03 |
| sripriya | sridhar_ram: tung_doan: catching up | 06:04 |
| sridhar_ram | tung_doan: sripriya: given the state of things.. i was wondering to merge the alarm change and continue to debug the ceilometer issue... | 06:04 |
| tung_doan | sridhar_ram: totally we can try with curl | 06:04 |
| sripriya | tung_doan: so does that mean the metadata handling works fine as is indepedent of ceilometer issue | 06:05 |
| sridhar_ram | tung_doan: sripriya: in fact we shd try heat+ceilometer WITHOUT tacker and see if the webhook calls are firing on say high CPU | 06:05 |
| *** lulei has quit IRC | 06:05 | |
| tung_doan | sridhar_ram: fron now to the end of this week I will fix | 06:05 |
| sripriya | tung_doan: referring to tacker code’s metadat change | 06:05 |
| *** lulei has joined #tacker | 06:06 | |
| sripriya | sridhar_ram: fine with me to merge | 06:06 |
| dkushwaha | sripriya, sridhar_ram we are investigating the things, but i could not found the route cause, but i think it sould not due to metadata | 06:06 |
| sripriya | dkushwaha: okay | 06:06 |
| tung_doan | dkushwaha: maybe.. because actually ceilometer-show totally correct.. | 06:07 |
| tung_doan | dkushwaha: but cannot figure out why it cannot trigger.. | 06:07 |
| sridhar_ram | Hmm.. given the bulk of this code.. i'd rather bring this in now and get few days of soak time (including the weekend) than waiting say until monday to fix Ceilometer issue and *THEN* merge this big alarm patchset. That is too risky for me. | 06:08 |
| sridhar_ram | meanwhile, tung_doan and dkushwaha - please continue full speed in fixing this issue even if the alarm mon code merges .. | 06:08 |
| tung_doan | sridhar_ram: thanks.. i know it's too late for newton cycle.. will try my best | 06:09 |
| dkushwaha | sridhar_ram, yes sure | 06:09 |
| sridhar_ram | sripriya: ?? | 06:09 |
| tung_doan | sridhar_ram: dkushwaha: please help me.. | 06:09 |
| sripriya | sridhar_ram: i acknowledged already | 06:10 |
| sridhar_ram | sripriya: ah, u did indeed ! | 06:10 |
| tung_doan | dkushwaha: to figure out what happend to ceilometer.. we dont have enough time | 06:10 |
| dkushwaha | tung_doan, yups, Still investigating it | 06:11 |
| sridhar_ram | dkushwaha: tung_doan: your patchset is in merge conflict .. can you quickly respin ? | 06:11 |
| tung_doan | sridhar_ram: np | 06:12 |
| *** santoshk has joined #tacker | 06:12 | |
| dkushwaha | sridhar_ram, will do it quickly | 06:12 |
| sripriya | tung_doan: the best we can do is to send the right params whatever ceilometer resource expects to work with heat autoscaling policy, ceilometer behaving weird is something we can track down may be with the celiometer team itself if required | 06:12 |
| sridhar_ram | tung_doan: dkushwaha: thanks.. please note this conflict is due to https://review.openstack.org/#/c/363351/ merge.. which moved the plugin file to vnfm dir | 06:12 |
| sridhar_ram | sripriya: +1 | 06:13 |
| openstackgerrit | Neeldhwaj Pathak proposed openstack/tacker: Validate the Tosca VNFD templates https://review.openstack.org/333852 | 06:13 |
| sripriya | tung_doan dkushwaha appreciate you guys pulling this through. way to go! | 06:14 |
| tung_doan | sripriya: still follow that way | 06:14 |
| tung_doan | sridhar_ram: sriprya: thanks | 06:14 |
| tung_doan | dkushwaha: could you help me please respin our patch? i am bussy to test... | 06:15 |
| dkushwaha | tung_doan, sure, will do | 06:16 |
| *** gongysh has joined #tacker | 06:16 | |
| tung_doan | dkushwaha: thanks | 06:16 |
| tung_doan | gongysh: hi | 06:16 |
| gongysh | sridhar_ram, progress on etherpad vnffg? | 06:17 |
| gongysh | tung_doan, hi | 06:17 |
| tung_doan | gongysh: also, our patch (alarm monitor) had problem with ceilometer. Currently, ceilometer cannot trigger alarms.. please help me to look into this... | 06:17 |
| gongysh | tung_doan, bad news | 06:18 |
| sridhar_ram | gongysh: i didn't see u in the channel.. here it is https://etherpad.openstack.org/p/tacker-newton-vnffg-testing | 06:18 |
| tung_doan | gongysh: yes :( | 06:18 |
| gongysh | sridhar_ram, just be offline. | 06:18 |
| sridhar_ram | gongysh: just to update the chat we just had among.. sripriya tung_doan dkushwaha and me... | 06:19 |
| sridhar_ram | gongysh: i've suggested to merge alarm code now and continue to debug this ceilometer issue over next day or two instead of waiting to merge this big patchset on say over the weekend or Monday | 06:19 |
| gongysh | sridhar_ram, agree | 06:20 |
| gongysh | tung_doan, hi, can u fix the merge conflict and let us merge it for the time being. | 06:21 |
| tung_doan | gongysh: dkushwaha will do | 06:21 |
| dkushwaha | gongysh, fixing merge conflict | 06:21 |
| gongysh | dkushwaha, tung_doan thanks. | 06:22 |
| openstackgerrit | vishwanath jayaraman proposed openstack/tacker: Devref for event logging feature https://review.openstack.org/370428 | 06:22 |
| openstackgerrit | Sridhar Ramaswamy proposed openstack/tacker: VNFFG: neutron networking-sfc driver and plugin fixes https://review.openstack.org/347568 | 06:23 |
| openstackgerrit | Lu lei proposed openstack/python-tackerclient: Stop using mox in tackerclient (1) https://review.openstack.org/359132 | 06:23 |
| sridhar_ram | gongysh: sripriya: folks, the next thing to line up to merge is FFG Horizon.. https://review.openstack.org/347779 | 06:24 |
| sridhar_ram | gongysh: sripriya: I verified VNFFGD Catalog and it works fine.. | 06:25 |
| sridhar_ram | gongysh: sripriya: VNFFG has underlying non-horizon issues hence it is unverified | 06:25 |
| *** vishwanathj has quit IRC | 06:25 | |
| sridhar_ram | gongysh: sripriya: again, please go and merge it .. assuming no major blockers | 06:25 |
| sripriya | sridhar_ram: will do | 06:28 |
| sridhar_ram | sripriya: thanks | 06:28 |
| *** janki has joined #tacker | 06:31 | |
| openstackgerrit | dharmendra kushwaha proposed openstack/tacker: Implement Alarm monitor https://review.openstack.org/365435 | 06:31 |
| openstackgerrit | Merged openstack/tacker-horizon: Adds VNFFG support in Tacker-horizon https://review.openstack.org/347779 | 06:32 |
| *** manikanta_tadi has quit IRC | 06:38 | |
| openstackgerrit | gengchc2 proposed openstack/tacker: Translate log messages https://review.openstack.org/374560 | 06:52 |
| *** priya__ has quit IRC | 06:56 | |
| *** Ravikiran_K has quit IRC | 06:57 | |
| openstackgerrit | Sripriya Seetharam proposed openstack/tacker: Introduce uniqueness constraint on resource names https://review.openstack.org/329759 | 07:03 |
| *** santoshk has quit IRC | 07:16 | |
| *** Ravikiran_K has joined #tacker | 07:17 | |
| openstackgerrit | Kanagaraj Manickam proposed openstack/tacker: Adds unit test case for scaling https://review.openstack.org/371368 | 07:30 |
| *** neel has quit IRC | 07:41 | |
| tung_doan | sridhar_ram: sripriya: gongysh: dkushwaha: godbless.. try with the older version of Ceilometer/Aodh... alarm was triggered well and scaling worked well too.. will deeply test it again | 07:42 |
| *** bluelime has joined #tacker | 07:47 | |
| *** bluelime has left #tacker | 07:48 | |
| *** dkushwaha has quit IRC | 08:04 | |
| *** hparekh_ has joined #tacker | 08:05 | |
| *** Ravikiran_K has quit IRC | 08:06 | |
| *** hparekh has quit IRC | 08:08 | |
| *** sripriya has left #tacker | 08:09 | |
| *** neel has joined #tacker | 08:40 | |
| openstackgerrit | Lu lei proposed openstack/tacker: Remove virtualenv management scripts from oslo-incubator https://review.openstack.org/306927 | 08:40 |
| *** manikanta_tadi has joined #tacker | 08:53 | |
| manikanta_tadi | Hi All, my devstack is broken after pulling latest code changes in master. Error Logs : http://paste.openstack.org/show/582531/ | 08:57 |
| manikanta_tadi | Anybody facing same issue ? | 08:57 |
| *** neel has quit IRC | 09:20 | |
| *** tung_doan has quit IRC | 09:21 | |
| *** gongysh has quit IRC | 09:27 | |
| *** achatterjee has quit IRC | 09:51 | |
| *** neel has joined #tacker | 10:31 | |
| *** tbh has joined #tacker | 10:38 | |
| *** tung_doan has joined #tacker | 11:02 | |
| *** xiayu has quit IRC | 11:02 | |
| janki | trozet, ping | 11:07 |
| openstackgerrit | Himanshu Kumar proposed openstack/tacker: changes in file post_test_hook.sh https://review.openstack.org/372897 | 11:20 |
| janki | sridhar_ram, ping | 11:22 |
| openstackgerrit | Lu lei proposed openstack/python-tackerclient: Base OpenStackClient(OSC) plugin support https://review.openstack.org/374715 | 11:24 |
| *** bobh has joined #tacker | 11:57 | |
| *** neel has quit IRC | 11:58 | |
| *** bobh has quit IRC | 12:22 | |
| *** hparekh_ has quit IRC | 12:29 | |
| openstackgerrit | Janki Chhatbar proposed openstack/tacker-horizon: VNFFG Manager fixes https://review.openstack.org/374767 | 12:36 |
| *** amotoki has joined #tacker | 13:00 | |
| *** KanagarajM_ has quit IRC | 13:10 | |
| trozet | hi janki | 13:29 |
| *** lamt has joined #tacker | 13:35 | |
| janki | trozet, hey, you there? | 13:50 |
| trozet | hi janki | 13:50 |
| janki | trozet, I wanted to understand vnf_mapping | 13:51 |
| janki | trozet, how it correlates to VNFFGD and which VNFDs to mention in the mapping | 13:51 |
| *** bobh has joined #tacker | 13:52 | |
| trozet | janki: the vnf mapping matches VNFs to VNFDs, those VNFDs are listed in the chain path of the NFP | 13:54 |
| trozet | janki: so if you provide vnf mapping, you can directly tell the plugin which VNFs you want to use in the chain | 13:55 |
| trozet | janki: otherwise it will pick for you by searching for VNFs that are spawned with the same VNFD | 13:55 |
| janki | trozet, aren't these specified in VNFFGD? | 13:55 |
| trozet | janki: in real NSD, it is all defined in one template | 13:56 |
| trozet | janki: but in Tacker now we declare VNFD and VNFFGD as 2 separate things | 13:57 |
| trozet | janki: there is nothing in the VNFFGD that says which VNF instance to use | 13:57 |
| janki | trozet, ohh, ohk. Thanks for explaining... | 13:57 |
| trozet | janki: np | 13:57 |
| janki | trozet, I was experimenting with this. I noticed in client while getting vnfs and vnfds from mapping, the list brackets are included too | 13:58 |
| trozet | janki: oh | 13:58 |
| janki | trozet, meaning if vnf_mapping = [vnfd1: vnf1], vnfd = [vnfd1 and vnf = vnf1] | 13:59 |
| *** bobh has quit IRC | 13:59 | |
| trozet | janki: can you fpaste hte output? | 13:59 |
| janki | trozet, give me 1 moment please | 14:00 |
| janki | trozet, the code and output is at http://paste.openstack.org/show/582584/ | 14:06 |
| janki | trozet, added 2 prints and its corresponding output. The solution would be to strip the "[ ]". Tried doing this (line commented in above paste) | 14:08 |
| janki | trozet, need to leave. attedning PyCon. I would be checking my mails in case you have any queries. Thanks | 14:16 |
| *** janki has quit IRC | 14:16 | |
| *** vishwanathj has joined #tacker | 14:19 | |
| openstackgerrit | vishwanath jayaraman proposed openstack/tacker: Devref for event logging feature https://review.openstack.org/370428 | 14:28 |
| openstackgerrit | Merged openstack/tacker: Introduce uniqueness constraint on resource names https://review.openstack.org/329759 | 14:43 |
| *** bobh has joined #tacker | 14:55 | |
| *** bobh has quit IRC | 14:59 | |
| *** lamt has quit IRC | 15:18 | |
| *** tung_doan has quit IRC | 15:50 | |
| *** uck has joined #tacker | 15:53 | |
| *** bobh has joined #tacker | 15:55 | |
| *** bobh has quit IRC | 16:00 | |
| trozet | sridhar_ram: ping? | 16:03 |
| *** KanagarajM_ has joined #tacker | 16:18 | |
| openstackgerrit | Kanagaraj Manickam proposed openstack/tacker: Adds unit test case for scaling https://review.openstack.org/371368 | 16:19 |
| *** Vijayendra has quit IRC | 16:23 | |
| openstackgerrit | Kanagaraj Manickam proposed openstack/tacker: Devref for event logging feature https://review.openstack.org/370428 | 16:25 |
| vishwanathj | KanagarajM_ FYI:the dsvm is failing for the devref patchset above..its a monitoring functional test case issue totally unrelated to our devref patchset | 16:40 |
| *** KanagarajM_ has quit IRC | 16:41 | |
| *** bobh has joined #tacker | 16:56 | |
| *** bobh has quit IRC | 17:01 | |
| sridhar_ram | trozet: pong | 17:07 |
| tbh | sridhar_ram, trozet while playing with VNFFG, I have some issues reported here https://bugs.launchpad.net/tacker/+bug/1626638. Let me know, if anyone already working on it or if it is not valid then I will mark it as invalid otherwise I will be working on it | 17:10 |
| openstack | Launchpad bug 1626638 in tacker "Few bugs in vnffgd code" [Undecided,New] - Assigned to bharaththiruveedula (bharath-ves) | 17:10 |
| sridhar_ram | tbh: trozet: most of these things are getting fixed using https://review.openstack.org/#/c/347568/ | 17:11 |
| sridhar_ram | tbh: trozet: btw, is the etherpad with my notes on how to test FFG with neutron-sfc ... https://etherpad.openstack.org/p/tacker-newton-vnffg-testing | 17:12 |
| tbh | sridhar_ram, ah, okay, I will mark it as invalid and will review this patch | 17:12 |
| tbh | sridhar_ram, will follow this etherpad | 17:14 |
| sridhar_ram | tbh: thanks.. we don't have time to float all the ffg fixes across diff patchset.. hence i've consolidated everything into this one.. | 17:15 |
| sridhar_ram | tbh: so that we can share each others fixes | 17:15 |
| tbh | sridhar_ram, sure no problem, yup that saves time | 17:15 |
| sridhar_ram | tbh: btw, the list_net / list_tenant_net is not solved.. i've remove network name in the ACL to sidestep that issue for now | 17:16 |
| tbh | sridhar_ram, ah, okay | 17:19 |
| *** uck has quit IRC | 17:21 | |
| *** uck has joined #tacker | 17:25 | |
| *** lamt has joined #tacker | 17:44 | |
| *** bobh has joined #tacker | 17:57 | |
| *** s3wong has joined #tacker | 17:58 | |
| *** bobh has quit IRC | 18:02 | |
| *** sripriya has joined #tacker | 18:10 | |
| *** tbh has quit IRC | 18:13 | |
| trozet | sridhar_ram: hey | 18:14 |
| vishwanathj | sridhar_ram and other cores, would appreciate review of event logging devref https://review.openstack.org/#/c/370428/ | 18:17 |
| *** uck has quit IRC | 18:38 | |
| *** bobh has joined #tacker | 18:58 | |
| *** bobh has quit IRC | 19:03 | |
| *** uck has joined #tacker | 19:39 | |
| *** uck has quit IRC | 19:45 | |
| sridhar_ram | vishwanathj: reviewed, pls checked my comments.. | 19:54 |
| vishwanathj | thanks | 19:54 |
| *** bobh has joined #tacker | 19:59 | |
| *** bobh has quit IRC | 20:03 | |
| *** santoshk has joined #tacker | 20:13 | |
| openstackgerrit | Sridhar Ramaswamy proposed openstack/tacker: VNFFG: neutron networking-sfc driver and plugin fixes https://review.openstack.org/347568 | 20:15 |
| *** sripriya has quit IRC | 20:19 | |
| *** sripriya has joined #tacker | 20:48 | |
| *** bobh has joined #tacker | 21:00 | |
| trozet | sridhar_ram: hi | 21:02 |
| sridhar_ram | trozet: hi | 21:03 |
| trozet | sridhar_ram: sorry for not being very useful past couple days...release blocking bugs popped up in OPNFV | 21:03 |
| trozet | sridhar_ram: we release tmrw...we really shouldnt release at same time as OS | 21:03 |
| trozet | sridhar_ram: I've been following some of the emails you sent and the patches | 21:03 |
| sridhar_ram | trozet: i agree, it is better if staggered.. | 21:03 |
| trozet | sridhar_ram: I saw you just posted another patch, can you give me a quick run down of where you are at/where can I help? | 21:04 |
| sridhar_ram | trozet: i was just trying to make sure all the hard work that went into FFG comes out in a useful way for newton release candidate | 21:04 |
| trozet | sridhar_ram: opnfv is all set to release now, so I'm pretty much free | 21:04 |
| *** bobh has quit IRC | 21:04 | |
| s3wong | trozet, sridhar_ram: I am about to post another patch --- though at the moment the base patch doesn't pass Jenkins | 21:04 |
| trozet | sridhar_ram, s3wong: i saw comments on email that the classification for matching net name etc would be removed because the functionality doesnt work to find the ID, is that right? | 21:05 |
| sridhar_ram | trozet: s3wong: we are just getting past neutron client auth related issues.. | 21:05 |
| sridhar_ram | trozet: that is correct | 21:05 |
| trozet | sridhar_ram: ok I think that is fine | 21:05 |
| s3wong | sridhar_ram: your change still has a problem --- you used auth1_sess, in which you only import session | 21:06 |
| trozet | sridhar_ram, s3wong: so now you have neutron client auth sorted out? | 21:06 |
| s3wong | sridhar_ram: so I changed that in my patchset | 21:06 |
| trozet | sridhar_ram: ok so what are you hitting now? | 21:06 |
| sridhar_ram | s3wong: sure, push an update and i'll pull it from there | 21:07 |
| s3wong | trozet: my upcoming patchset (hopefull | 21:07 |
| s3wong | hopefully syncs up with most of your flow_classifier attribute names ... | 21:07 |
| sridhar_ram | trozet: just to level set.. i'm using the cmds as listed here https://etherpad.openstack.org/p/tacker-newton-vnffg-testing#L41 | 21:08 |
| trozet | s3wong, sridhar_ram: so after htat patch, are you able to create a chain? or what is blocking it from working? | 21:09 |
| sridhar_ram | trozet: i still haven't got past the neutron auth issue.. | 21:09 |
| sridhar_ram | trozet: i was just about trying.. there is a pesky devstack issue that cropped up right now, which is burning some of my time | 21:10 |
| trozet | sridhar_ram: ok then I can download the current patch and try the neutron auth | 21:10 |
| s3wong | trozet: please wait until I post my patch :-) | 21:11 |
| s3wong | sridhar_ram, trozet: my ubuntu 16.04 based setup is still mess up --- I have DB error when I try to create fc | 21:11 |
| sridhar_ram | trozet: yes, please do.. at least it will nice if you can hit the same error as we are all hitting .. so that we can help each other and iterate | 21:11 |
| openstackgerrit | vishwanath jayaraman proposed openstack/tacker: Devref for event logging feature https://review.openstack.org/370428 | 21:13 |
| trozet | sridhar_ram, s3wong: ok I have a CentOS7 deployment, so should be another datapoint | 21:14 |
| sridhar_ram | trozet: +2 | 21:14 |
| s3wong | trozet: good luck :-) | 21:15 |
| trozet | sridhar_ram, s3wong: i have to go afk for a bit to get kids dinner. Will be back and try it later tonight. | 21:17 |
| sridhar_ram | trozet: sounds good.. | 21:17 |
| sridhar_ram | s3wong: looks your patchset needs to be rebased | 21:19 |
| sridhar_ram | s3wong: oops, sorry .. there is typo that is severely mess up the patchset.. | 21:20 |
| sridhar_ram | s3wong: can i quickly push to remove two chars in your patchset ? | 21:21 |
| s3wong | sridhar_ram: go for it | 21:26 |
| s3wong | sridhar_ram: was it the auth1_sess thing? | 21:26 |
| sridhar_ram | s3wong: no, pure fat finger :( | 21:27 |
| openstackgerrit | vishwanath jayaraman proposed openstack/tacker: Devref for event logging feature https://review.openstack.org/370428 | 21:27 |
| s3wong | sridhar_ram: OK :-) | 21:27 |
| sridhar_ram | s3wong: btw, what is the issue with auth1_sess ? | 21:27 |
| sridhar_ram | s3wong: did you look at the latest https://review.openstack.org/#/c/347568/16/tacker/nfvo/drivers/vim/openstack_driver.py ? | 21:29 |
| sridhar_ram | s3wong: it has some auth related fixes | 21:29 |
| s3wong | sridhar_ram: in line 467 of https://review.openstack.org/#/c/347568/16/tacker/nfvo/drivers/vim/openstack_driver.py,unified | 21:29 |
| s3wong | auth1_sess is not defined | 21:29 |
| sridhar_ram | s3wong: got it.. that was my bad | 21:29 |
| s3wong | session if imported in line 21, so the other change is to put this into session instead of auth1_sess | 21:29 |
| s3wong | sridhar_ram: easy fix :-) | 21:29 |
| s3wong | sridhar_ram: mine is ready --- do you want me to post mine first? | 21:30 |
| sridhar_ram | s3wong: if you are pushing one can u also fix L470 | 21:30 |
| s3wong | sridhar_ram: oh, that one --- yes, I have already fixed that in my view also | 21:31 |
| s3wong | sridhar_ram: so I will just push that in also | 21:31 |
| sridhar_ram | s3wong: cool thanks, go ahead | 21:31 |
| s3wong | sridhar_ram: I have not fixed the various Jenkins error, though | 21:31 |
| sridhar_ram | s3wong: lets fix this obvious ones first.. | 21:31 |
| *** sripriya has quit IRC | 21:31 | |
| s3wong | sridhar_ram: I tried to fix some last night with some pep8; yeah, let's get the functionality down first | 21:32 |
| openstackgerrit | Stephen Wong proposed openstack/tacker: VNFFG: neutron networking-sfc driver and plugin fixes https://review.openstack.org/347568 | 21:33 |
| *** veena has quit IRC | 21:37 | |
| *** santoshk has quit IRC | 21:37 | |
| *** amotoki has quit IRC | 21:37 | |
| *** Qiming has quit IRC | 21:37 | |
| *** vishwanathj has quit IRC | 21:37 | |
| *** vishnoianil has quit IRC | 21:37 | |
| *** amit213 has quit IRC | 21:37 | |
| s3wong | Lost a bunch of people just as I posted a patchset | 21:37 |
| s3wong | :-) | 21:37 |
| *** vishnoianil has joined #tacker | 21:40 | |
| *** amit213 has joined #tacker | 21:40 | |
| *** veena has joined #tacker | 21:41 | |
| *** santoshk has joined #tacker | 21:41 | |
| *** amotoki has joined #tacker | 21:41 | |
| *** Qiming has joined #tacker | 21:41 | |
| *** uck has joined #tacker | 21:42 | |
| *** vishwanathj has joined #tacker | 21:42 | |
| *** amit213 has quit IRC | 21:44 | |
| *** uck has quit IRC | 21:47 | |
| sridhar_ram | s3wong: looks like a co-ordinated network flap ;-) | 22:00 |
| sridhar_ram | s3wong: btw, i just fixed pep8 issues .. pushing it now | 22:00 |
| s3wong | sridhar_ram: cool! | 22:00 |
| openstackgerrit | Sridhar Ramaswamy proposed openstack/tacker: VNFFG: neutron networking-sfc driver and plugin fixes https://review.openstack.org/347568 | 22:01 |
| *** bobh has joined #tacker | 22:01 | |
| openstackgerrit | Sridhar Ramaswamy proposed openstack/tacker: VNFFG: neutron networking-sfc driver and plugin fixes https://review.openstack.org/347568 | 22:01 |
| sridhar_ram | s3wong: i also rebased to latest master | 22:02 |
| s3wong | sridhar_ram: OK | 22:02 |
| *** bobh has quit IRC | 22:05 | |
| *** bobh has joined #tacker | 22:08 | |
| *** sripriya has joined #tacker | 22:13 | |
| *** amit213 has joined #tacker | 22:14 | |
| *** vishwanathj has quit IRC | 22:18 | |
| *** bobh has quit IRC | 22:22 | |
| sridhar_ram | s3wong: now, with latest patchset.. port pairs are created.. still there are errors but at least it is a progress | 22:23 |
| *** lamt has quit IRC | 22:24 | |
| s3wong | sridhar_ram: cool | 22:27 |
| s3wong | sridhar_ram: flow-classifier created also? | 22:27 |
| * sridhar_ram checking | 22:27 | |
| sridhar_ram | s3wong: no | 22:28 |
| sridhar_ram | s3wong: no flow-classifier got created | 22:28 |
| s3wong | sridhar_ram: damnit! | 22:28 |
| sridhar_ram | s3wong: a dumb question.. | 22:28 |
| s3wong | sridhar_ram: still error there then --- fc is the one with the most parameters, I think port-pair, port-pair-group should be fine. The next challenge after fc would be chain | 22:29 |
| s3wong | sridhar_ram: go ahead | 22:29 |
| sridhar_ram | s3wong: on which "neutron network" this flow-classifier will be applied ? | 22:29 |
| s3wong | sridhar_ram: no where when you just create a fc | 22:29 |
| s3wong | sridhar_ram: it wouldn't be until the chain is resolved would a flow-classifier be applied | 22:30 |
| sridhar_ram | s3wong: i know ultimately there are neutron port (vNICs).. | 22:30 |
| s3wong | sridhar_ram: but fc is only programmed on ingress and egress | 22:30 |
| s3wong | sridhar_ram: NOT throughout the chain | 22:30 |
| sridhar_ram | s3wong: ingress and egress of what ? | 22:30 |
| * sridhar_ram reminds this is a dumb question | 22:30 | |
| s3wong | sridhar_ram: hence the REQUIRED parameters of source Neutron port and destination Neutron port for fc create | 22:30 |
| s3wong | sridhar_ram: of the chain | 22:31 |
| s3wong | sridhar_ram: you create a chain of Neutron ports (presumably of VNFs) | 22:31 |
| sridhar_ram | s3wong: oh, this "ACL" is applied on a per neutron port basis ? | 22:32 |
| s3wong | sridhar_ram: then the fc with its source/dst Neutron ports would in essence redirect traffic to the chain | 22:32 |
| s3wong | sridhar_ram: not throughout the chain | 22:32 |
| s3wong | sridhar_ram: only on ingress / egress --- within the chain, the MPLS label would specify the next element of the chain | 22:33 |
| s3wong | sridhar_ram: the idea is we don't do fully flow-classifier lookup on every hop of the chain | 22:33 |
| sridhar_ram | s3wong: okay, i'll re-read the n-sfc literature with this in mind | 22:33 |
| sridhar_ram | s3wong: make sense | 22:33 |
| s3wong | sridhar_ram: so... as gongshi once asked me --- that the logic source/dest ports are REQUIRED for fc creation --- are those specified in your test? | 22:34 |
| sridhar_ram | s3wong: i need to check.. | 22:34 |
| sridhar_ram | s3wong: but.. i still get "flow classifier match not supported by networking-sfc driver" | 22:34 |
| s3wong | sridhar_ram: do you have the fc_dict is ... output? | 22:35 |
| sridhar_ram | s3wong: fc passed is {'match': {'ip_proto': 6, 'ip_dst_prefix': u'192.168.1.2/24', 'destination_port_min': 80, 'destination_port_max': 1024}} | 22:35 |
| s3wong | sridhar_ram: oh, I see | 22:35 |
| s3wong | sridhar_ram": the incoming dict is fc['match'] instead of fc | 22:35 |
| s3wong | sridhar_ram: so you need to extract the 'match' dictionary from fc | 22:36 |
| s3wong | sridhar_ram: why would trozet pass to driver with a 'match' key to the flow classifier dictionary? :-) | 22:37 |
| sridhar_ram | trozet: not sure, i'm now trying with this.. http://paste.openstack.org/show/582676/ | 22:38 |
| s3wong | sridhar_ram: that actually won't work --- the parsing happen before invoking NeutronClient | 22:39 |
| sridhar_ram | s3wong: okay, can u take a stab at this ? | 22:40 |
| s3wong | sridhar_ram: you need to do that in create_flow_classifier itself :-) | 22:40 |
| s3wong | sridhar_ram: OK | 22:40 |
| sridhar_ram | s3wong: if it is something simple.. i can quickly give it a try | 22:41 |
| s3wong | sridhar_ram: simple, yet tedious | 22:42 |
| s3wong | sridhar_ram: between line 316 and 317, do fc_dict = fc['match'] | 22:42 |
| s3wong | sridhar_ram: then replace all subsequent fc with fc_dict | 22:42 |
| s3wong | (which are only three instances, so not so bad) | 22:43 |
| sridhar_ram | s3wong: making the changes.. hang on | 22:47 |
| sridhar_ram | s3wong: proceeding further to get "BadRequest: FlowClassifier 0ea70908-c70b-4419-bf45-5a9bb38ee6cd does not set logical source port in ovs driver" | 22:51 |
| sridhar_ram | s3wong: based on the discussion we just had .. is logical_source_port and/or logical_destination_port required in flowclassifier ? | 23:03 |
| *** lamt has joined #tacker | 23:09 | |
| s3wong | sridhar_ram: sorry, stepped up for a moment | 23:10 |
| s3wong | sridhar_ram: yes, those are required | 23:10 |
| s3wong | sridhar_ram: and that is an AND (not or) --- so you need both source and destination logical ports | 23:11 |
| s3wong | sridhar_ram: (hang in there, you are about to get to ... second base) | 23:11 |
| s3wong | sridhar_ram: after that, you should pass networking-sfc DB / API, then you are entering into a territory that I myself had not gone: OVS flow tables :-) | 23:12 |
| *** sripriya has quit IRC | 23:35 | |
| *** sripriya has joined #tacker | 23:39 | |
| sridhar_ram | s3wong: i was afk.. back now | 23:41 |
| sridhar_ram | s3wong: these two ports go into the sam fc_dict ? | 23:42 |
| sridhar_ram | *same | 23:42 |
| sridhar_ram | s3wong: trozet: looks nfvo_plugin need to be modified to pass the neutron ports in the fc dict in the first place | 23:44 |
| s3wong | sridhar_ram: same one | 23:44 |
| *** uck has joined #tacker | 23:44 | |
| sridhar_ram | s3wong: trozet: btw, if there are multiple VNFs say with two neutron ports each (ingress & egress).. which neutron ports needs to be used as logical_source and logical_destination ? | 23:46 |
| s3wong | sridhar_ram: limitation of the API :-) | 23:47 |
| s3wong | sridhar_ram: for now, let's just pick the one where you want traffic to send out to chain | 23:49 |
| *** uck has quit IRC | 23:49 | |
| s3wong | sridhar_ram: (or of course, you can just create another fc, and bind it to the same chain) | 23:49 |
| sridhar_ram | s3wong: looking http://docs.openstack.org/developer/networking-sfc/api.html .. it doesn't show logical ports as mandatory args.. do you know if this behavior has recently changed ? | 23:51 |
| s3wong | sridhar_ram: no, it has always been like this --- it is just a documentation bug | 23:51 |
| sridhar_ram | s3wong: I see | 23:51 |
| s3wong | sridhar_ram: gongysh found out about that couple months ago, well before we cut stable/mitaka | 23:52 |
| sridhar_ram | s3wong: i'm worried this is going to be a non-trivial work on the plugin side.. particularly figured out the two edges / ports of a chain. Need to consult w/ trozet | 23:53 |
| sridhar_ram | *figuring | 23:53 |
| s3wong | sridhar_ram: yes. Need to talk to trozet on this | 23:53 |
| s3wong | sridhar_ram: I believe when you deploy a chain, the ingress / egress points needed to be specified for sure; but those may not be part of VNFFG template | 23:54 |
| *** KanagarajM_ has joined #tacker | 23:57 | |
| *** ksantoshk has joined #tacker | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!