| *** gongysh has joined #openstack-lbaas | 00:03 | |
| johnsom | Well that is good news. With event streamer enabled, neutron lbaas-loadbalancer-status is updating like it should.... | 00:12 |
|---|---|---|
| *** reedip has quit IRC | 00:47 | |
| *** KeithMnemonic1 has quit IRC | 01:00 | |
| *** armax has quit IRC | 01:08 | |
| *** reedip has joined #openstack-lbaas | 01:24 | |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Adds v2 load balancer API section https://review.openstack.org/438757 | 01:39 |
| *** gongysh has quit IRC | 01:40 | |
| johnsom | Ok, I think the load balancer API reference section is ready for review | 01:40 |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Adds v2 load balancer API section https://review.openstack.org/438757 | 01:43 |
| *** ndahiwade has joined #openstack-lbaas | 01:45 | |
| *** ianychoi has quit IRC | 01:48 | |
| *** ianychoi has joined #openstack-lbaas | 01:54 | |
| bbzhao | johnsom, hi micheal, I have left some comments in https://review.openstack.org/#/c/441912/6 , You said just accept qos_policy_id. Could we accept the Qos policy details (such as max_kbps and max_burst_kbps), we can use this parameters to create the qos in neutron? Is that OK? | 02:09 |
| reedip | johnson , rm_work( in case you are awake ) : Is octavia still going to be a part of NeutronStadium? | 02:17 |
| *** gongysh has joined #openstack-lbaas | 02:36 | |
| *** csomerville has joined #openstack-lbaas | 02:46 | |
| *** cody-somerville has quit IRC | 02:50 | |
| reedip | bbzhao : I responded to your spec :) | 02:50 |
| reedip | with my understanding .. | 02:50 |
| *** csomerville has quit IRC | 02:52 | |
| bbzhao | reedip, thanks for your comments, I 'm returned on the spec now. At first , You are welcome for the work of this spec. Many thanks for help us.:) | 02:53 |
| reedip | thanks bbzhao .. please let me know what you think about my idea ... I think we can directly tie in with Neutron's QoS than to do something new in LBaaS | 02:54 |
| johnsom | bbzhao Thanks for chatting | 02:55 |
| bbzhao | Wow | 02:55 |
| bbzhao | hi | 02:56 |
| johnsom | I was asking if it was ok to just use the qos_policy_id. It would make the code much easier | 02:56 |
| johnsom | It was an idea / question | 02:56 |
| bbzhao | oh Sorry for that, I see. | 02:57 |
| reedip | johnsom : Yes we can ,IMHO | 02:57 |
| *** bzhao_ has joined #openstack-lbaas | 02:58 | |
| reedip | johnsom : I had a query earlier : Is octavia still going to be a part of NeutronStadium? | 02:59 |
| reedip | after neutron-lbaas is deprecated? | 02:59 |
| *** sanfern has joined #openstack-lbaas | 03:00 | |
| bbzhao | sorry for a meeting.. | 03:00 |
| reedip | reason I am asking is, it seems to me that neutron cores would like the client side implementation of projects which are a part of neutronstadium, in neutronclient | 03:00 |
| johnsom | reedip In Ocata, octavia and neutron-lbaas spun out of neutron stadium and became a top level OpenStack project: https://governance.openstack.org/tc/reference/projects/octavia.html | 03:00 |
| reedip | johnsom : ok, sounds good enough for me now ... thanks :) | 03:01 |
| johnsom | I.e. We are in the big leagues now... We are at the same standing as neutron... | 03:01 |
| reedip | :) | 03:01 |
| johnsom | Grin | 03:01 |
| reedip | johnsom : btw I also had an idea | 03:02 |
| reedip | wanted to know if there is any use case for loadbalancers mirroring network packets for network analysis | 03:02 |
| reedip | just wanted to get an idea about the above idea :) | 03:02 |
| johnsom | bbzhao We meet on #openstack-meeting-alt | 03:03 |
| reedip | johnsom :when ?????? | 03:03 |
| reedip | oh I missed it :( | 03:04 |
| johnsom | reedip Hmm, I know neutron has some port mirroring support or proposals | 03:04 |
| reedip | johnsom : that is known as Tap-as-a-Service :D | 03:04 |
| johnsom | We meet every week at 20:00 UTC on Wednesday | 03:04 |
| johnsom | https://wiki.openstack.org/wiki/Octavia | 03:04 |
| reedip | yeah , I saw the chat now :( | 03:05 |
| johnsom | There is a full log of the meeting linked there as well if you missed the meeting and want to see what we discussed | 03:05 |
| johnsom | reedip So, port mirroring. Do you have a use case for end users doing this or just operators? | 03:08 |
| reedip | johnsom : currently , port-mirroring is required by telcos | 03:08 |
| reedip | johnsom : and thats where we are focussing for now, but I guess if need be, telecos who use loadbalancers may need port-mirroring as well. We are waiting for it as of now. (seems to be a very low priority ) | 03:09 |
| johnsom | bbzhao It seems more flexible if we just take the neutron qos_policy_id. For example if they add a new feature to neutron qos, like the method CoDel, etc., we don't have to update the octavia API | 03:11 |
| johnsom | reedip I get the value, I'm just trying to understand the "user". If it is an operator with admin rights, they can use the tap-as-a-service as they can see/access the VIP port owned by the neutron/octavia service account. | 03:12 |
| reedip | johnsom : operator with admin rights, for now... tenant probably later .... | 03:13 |
| reedip | johnsom: anyways I think the port created by octavia can be easily tied up with loadbalancer and verified | 03:14 |
| johnsom | So, operator should be good to go now assuming tap-as-a-service is functional in neutron | 03:14 |
| reedip | johnsom : yep | 03:14 |
| johnsom | Yeah, end users, that is a different story. I think once we get our load balancer logs story implemented, that will probably address most if not all of the end-user needs | 03:15 |
| reedip | johnsom : also , yes in bbzhao 's patch, we do not need to modify the tables , but I think minor Octavia API modification may be required. Also as a QoS policy may have more than one rule, it would be difficult to moderate the flow table filters implemented by the various QoS rules on the loadbalancer. | 03:16 |
| johnsom | We are trying to investigate the "right" way to expose the LB logs to end users during Pike. It's something we (as an operator) want. | 03:16 |
| reedip | johnsom: ok... | 03:18 |
| johnsom | reedip Yeah, where I am leaning is, user passes us a neutron qos_policy_id at LB create time. We take that and apply it to the LB VIP port for the reference driver, but if the vendor driver wants to do QoS in hardware, they can query the neutron qos_policy_id and sign up for change notifications. This keeps it simple for octavia/LBaaS | 03:18 |
| johnsom | As they add features to QoS in neutron we get them for free. | 03:19 |
| reedip | johnsom : this also introduces lot of verification at our end. Do we implement DSCP Marking? Do we assure Minimum Bandwidth Garuntee ? | 03:20 |
| reedip | johnsom: what happens if a QoS policy is modified? Would LB work will with it ?? ( and the scenarios increase with more features :D ) | 03:21 |
| johnsom | reedip Right, with just taking the ID and configuring neutron to manage that, we don't have to actually implement any of that | 03:21 |
| johnsom | If vendor wants to go there, they could, it just gets complicated... | 03:22 |
| reedip | johnsom: on the contrary, I think we might have to , else we may have a complicated scenario in our hand where in instead of limiting Loadbalancer to a max bandwidth of 100kbps, we assure that the port gets a minimum bandwidth of 200 kbps ( QoS Policies can change that way ) | 03:22 |
| reedip | because of the changing QoS Rules | 03:22 |
| reedip | johnsom : I guess the feature owner will have lot of things to read up once his meeting ends :D | 03:23 |
| johnsom | Right, if they change their QoS policy in neutron, the LB would automatically change | 03:23 |
| johnsom | Since octavia isnt' actually doing the QoS but neutron is, using the policy the user specified | 03:24 |
| *** fnaval has joined #openstack-lbaas | 03:24 | |
| johnsom | isnt' == is not, typo | 03:25 |
| reedip | kk | 03:25 |
| johnsom | Does that make sense? Does that work for you? | 03:25 |
| reedip | johnsom : Considering that we would still have to do a lot of FUnctional Test cases to do , for now it sounds ok.... I am a bit skeptical with the type of checks we need to do , but yes, it saves the extra implentation to be done on the Octavia side | 03:32 |
| johnsom | reedip What kind of checks? We are discussing this and trying to investigate all of the issues, so please voice concerns. | 03:33 |
| *** sticker_ has joined #openstack-lbaas | 03:33 | |
| reedip | johnsom : sorry, I am also having a parallel meeting.... can I discuss with you the same on the patch ? | 03:33 |
| johnsom | Sure, we can discuss another time. It is evening here anyway. I just want to make sure I hear/understand your perspective! | 03:36 |
| *** sticker has quit IRC | 03:36 | |
| *** links has joined #openstack-lbaas | 03:37 | |
| *** amotoki has quit IRC | 03:37 | |
| *** sticker_ is now known as sticker | 03:38 | |
| reedip | johnsom : hi .. my only concern is that qos policy must only have the bandwidth limit rules. It shouldnt have any other qos rule | 03:39 |
| johnsom | Ok, interesting. So, if they specify a DSCP rule, the load balancer is just going to ignore and strip those right? | 03:43 |
| johnsom | Is your concern about future rules? | 03:43 |
| *** armax has joined #openstack-lbaas | 03:49 | |
| *** amotoki has joined #openstack-lbaas | 03:50 | |
| reedip | johnsom : yes | 03:54 |
| reedip | johnsom : loadbalancer SHOULD .. but would it ? that needs to be checked | 03:54 |
| *** amotoki has quit IRC | 03:55 | |
| johnsom | It will with octavia driver for sure | 03:57 |
| johnsom | With vendor driver, wouldn't that be the user's issue. If their hardware honors DSCP, maybe they are using it for a reason | 03:58 |
| *** amotoki has joined #openstack-lbaas | 03:59 | |
| reedip | that sounds correct | 04:00 |
| *** sticker_ has joined #openstack-lbaas | 04:17 | |
| *** amotoki has quit IRC | 04:19 | |
| *** sticker__ has joined #openstack-lbaas | 04:19 | |
| *** sticker has quit IRC | 04:21 | |
| *** sticker__ is now known as sticker | 04:21 | |
| *** krypto has joined #openstack-lbaas | 04:21 | |
| *** krypto has quit IRC | 04:21 | |
| *** krypto has joined #openstack-lbaas | 04:21 | |
| *** sticker_ has quit IRC | 04:22 | |
| *** krypto has quit IRC | 04:27 | |
| *** krypto has joined #openstack-lbaas | 04:28 | |
| *** amotoki has joined #openstack-lbaas | 04:28 | |
| *** gongysh has quit IRC | 04:34 | |
| *** fnaval has quit IRC | 04:42 | |
| *** armax has quit IRC | 04:44 | |
| *** armax has joined #openstack-lbaas | 04:44 | |
| *** gongysh has joined #openstack-lbaas | 05:03 | |
| *** gcheresh_ has joined #openstack-lbaas | 05:54 | |
| openstackgerrit | huangshan proposed openstack/neutron-lbaas master: Remove log translations from neutron-lbaas https://review.openstack.org/447921 | 05:59 |
| bbzhao | johnsom, reedip Sorry, back right now for other meeting. reedip raise the issue that I'm concerned too. What we want the function is just the qos bandwith limitation, so the created qos_policy should contained the bandwith limit rule only. But I saw your chat log, only input qos_policy_id, and the more qos_rules in it should be considered by the reasons of the real usage from vendors. That sounds we just extend the octavia API to accept an Qos | 06:01 |
| bbzhao | policy, not just for bandwith limit.Right? | 06:01 |
| *** armax has quit IRC | 06:03 | |
| *** armax has joined #openstack-lbaas | 06:03 | |
| *** armax has quit IRC | 06:04 | |
| reedip | bbzhao : yeah | 06:04 |
| *** armax has joined #openstack-lbaas | 06:05 | |
| reedip | but we need to mention the limitation ( this looks like a limitation for third party drivers) | 06:05 |
| *** armax has quit IRC | 06:05 | |
| *** bedis has quit IRC | 06:07 | |
| *** bedis has joined #openstack-lbaas | 06:07 | |
| *** kbyrne_ has joined #openstack-lbaas | 06:09 | |
| *** kbyrne has quit IRC | 06:13 | |
| *** kbyrne_ is now known as kbyrne | 06:13 | |
| bbzhao | That means we need mentioned this policy is just for bandwith limitation. That sounds we need maintain the usage of qos policy for vip port. But neutron just allow 1 port can applied 1 policy.. | 06:13 |
| reedip | ok , so ? | 06:14 |
| bbzhao | If extend the API just accept the qos policy id, and maintain the usage about that, we can't know what the users will do in the next. For example, user create lb with qos policy with bandwith limit rule only, we store an "bandwith limitation" about that, then users want add some DSCP rule into this policy on neutron side. Octavia need know the qos usage info should be add some info about DSCP rule addition. So I think this may be a alternative | 06:21 |
| bbzhao | solution for lb port bandwith limitation. | 06:21 |
| *** bzhao_ has quit IRC | 06:38 | |
| reedip | bbzhao : this is exactly what I mentioned to johnsom | 06:47 |
| reedip | his idea is that octavia default driver would ignore the DSCP marking | 06:48 |
| reedip | but I am not sure about the minimum bandwidth garuntee rule | 06:48 |
| reedip | he is correct about dscp | 06:48 |
| reedip | but the min bandwidth may be an issue or a blessing | 06:48 |
| *** amotoki_ has joined #openstack-lbaas | 06:52 | |
| *** amotoki has quit IRC | 06:55 | |
| bbzhao | reedip, Yeah, that's why the spec input parameters is all about bandwith limit rule creation. I'm agree that default driver ignore the DSCP marking. I need to think about the effect of min bandwitdt. I need investigate it. | 06:56 |
| reedip | bbzhao : think of it as a DAM on a RIVER | 06:56 |
| reedip | right now, with max bandwidth, you are blocking the river on one end, if you put mim bandwidth, it will block at both the ends ... so the streaming data would be much better | 06:57 |
| reedip | :P | 06:57 |
| reedip | mind it , using both mim bw and max bw will make the VIP port function better if we have streamed data | 06:59 |
| bbzhao | reedip, Thanks for decription. make sence. It is for the big Network traffic Scenario. Yeah, make sence now | 07:05 |
| bbzhao | reedip, but I think is OK, eventhough I set the min bd rule is larger than max db rule. The effect should be owned by the users. As lb just read the qos_policy_id and use it. Lb dont create any rules. | 07:21 |
| *** eezhova has joined #openstack-lbaas | 07:22 | |
| bbzhao | reedip, but if there is no max bd rule in the policy..Wow..octavia need query this qos_policy for verify it contains a max bd rule? | 07:27 |
| reedip | bbzhao : When minimum bw rule is > max bw rule ... https://review.openstack.org/#/c/442375/ | 07:28 |
| reedip | bbzhao : octavia neednt query the QoS policy ... | 07:28 |
| reedip | bbzhao: whatever policy is set for QoS policy will drive the port data .. LBaaS may not be affected by iot | 07:29 |
| bbzhao | reedip, hah. I saw the patch. Ok, I forget octavia just extend to accept qos_policy... | 07:32 |
| reedip | bbzhao : yeah , the idea becomes pretty simple if we understand Octavia wont be responsible for the Data flow , that would be QoS | 07:32 |
| reedip | bbzhao : if the port is associated with incorrect QoS policies, then its a user misoperation, though the same needs to be mentioned in the RST and the document which would have the reno for this feature | 07:33 |
| reedip | so thats a ToDO from our end.. | 07:33 |
| bbzhao | reedip, Yeah. It's like we pass a parameter in port update.. | 07:39 |
| bbzhao | reedip, it's very simple if we understood the standing of both side. | 07:40 |
| reedip | bbzhao : I guess thats the intention of johnsom : make complicated things simpler :D | 07:41 |
| bbzhao | reedip, hah. That's correct. | 07:42 |
| *** tesseract has joined #openstack-lbaas | 07:44 | |
| *** eezhova has quit IRC | 08:03 | |
| *** pcaruana has joined #openstack-lbaas | 08:06 | |
| openstackgerrit | huangshan proposed openstack/neutron-lbaas master: Remove log translations from neutron-lbaas https://review.openstack.org/447921 | 08:10 |
| reedip | is it me or is FibreHome really looking into log translations | 08:11 |
| openstackgerrit | Merged openstack/octavia master: Look for all qemu-img/kvm package names in rhel-based distros https://review.openstack.org/446610 | 08:20 |
| *** kobis has joined #openstack-lbaas | 08:21 | |
| *** eezhova has joined #openstack-lbaas | 08:36 | |
| *** kobis1 has joined #openstack-lbaas | 08:42 | |
| *** kobis has quit IRC | 08:42 | |
| *** reedip has quit IRC | 08:49 | |
| *** aojea has joined #openstack-lbaas | 09:01 | |
| *** openstackgerrit has quit IRC | 09:03 | |
| *** belharar has joined #openstack-lbaas | 09:17 | |
| *** reedip has joined #openstack-lbaas | 09:25 | |
| *** aojea_ has joined #openstack-lbaas | 09:31 | |
| *** aojea has quit IRC | 09:34 | |
| *** gongysh has quit IRC | 09:38 | |
| *** links has quit IRC | 09:39 | |
| *** amotoki_ has quit IRC | 09:40 | |
| *** links has joined #openstack-lbaas | 09:56 | |
| *** amotoki has joined #openstack-lbaas | 09:59 | |
| *** openstackgerrit has joined #openstack-lbaas | 10:07 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/neutron-lbaas master: Imported Translations from Zanata https://review.openstack.org/449037 | 10:07 |
| *** links has quit IRC | 10:11 | |
| *** sanfern has quit IRC | 10:16 | |
| *** krypto has quit IRC | 10:19 | |
| *** links has joined #openstack-lbaas | 10:23 | |
| *** m-greene| has quit IRC | 10:30 | |
| openstackgerrit | Nir Magnezi proposed openstack/octavia master: Auto detect haproxy user_group https://review.openstack.org/429398 | 10:32 |
| *** m-greene has joined #openstack-lbaas | 10:34 | |
| *** m-greene- has quit IRC | 10:35 | |
| *** links has quit IRC | 10:35 | |
| *** m-greene_ has joined #openstack-lbaas | 10:36 | |
| *** amotoki has quit IRC | 10:46 | |
| *** amotoki has joined #openstack-lbaas | 10:49 | |
| *** links has joined #openstack-lbaas | 10:51 | |
| *** m-greene has quit IRC | 10:55 | |
| *** m-greene_ has quit IRC | 10:55 | |
| *** ndahiwade has quit IRC | 11:36 | |
| *** amotoki has quit IRC | 11:45 | |
| *** Guest3223 is now known as dasm | 12:03 | |
| *** sanfern has joined #openstack-lbaas | 12:13 | |
| *** catintheroof has joined #openstack-lbaas | 12:44 | |
| *** catintheroof has quit IRC | 13:03 | |
| *** amotoki has joined #openstack-lbaas | 13:13 | |
| rm_work | rebase time! | 13:31 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Octavia v2 API for listeners https://review.openstack.org/424744 | 13:31 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Introduce Octavia v2 API for pools https://review.openstack.org/405922 | 13:31 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Introduce Octavia v2 API for Members https://review.openstack.org/405932 | 13:31 |
| openstackgerrit | Merged openstack/octavia master: Replace six.iteritems() with .items() https://review.openstack.org/444694 | 13:39 |
| *** belharar has quit IRC | 13:49 | |
| *** armax has joined #openstack-lbaas | 14:18 | |
| *** fnaval has joined #openstack-lbaas | 14:29 | |
| *** links has quit IRC | 14:38 | |
| rm_work | hmmmmmm | 14:49 |
| rm_work | so on pools, having listener_id and loadbalancer_id seems redundant | 14:49 |
| rm_work | but i guess sometimes you won't have a listener, so you need at least one of the two | 14:49 |
| rm_work | and i do believe loadbalancer_id is mandatory in neutron-lbaas | 14:49 |
| rm_work | but i almost want to make it not? and do the validation manually | 14:49 |
| *** diltram_ is now known as diltram | 14:50 | |
| sshank | rm_work, listener_id in pools was made optional right? And loadbalancer_id as mandatory. | 15:02 |
| rm_work | right, but | 15:03 |
| rm_work | if you DO provide listener_id, loadbalancer_id seems redundant in that case | 15:03 |
| sshank | Hmm yeah. | 15:04 |
| *** armax has quit IRC | 15:09 | |
| ankur-gupta-f4 | rm_work: i totally forgot to push up my changes yesterday. My bad. Did u push up some changes to listeners? | 15:13 |
| rm_work | just a rebase this morning | 15:14 |
| rm_work | did you have changes to it? | 15:14 |
| rm_work | you should be able to do a diff against the reflog before you did your amend commit, and then re-apply those as a patch on top of the latest | 15:14 |
| rm_work | (I have that problem quite often) | 15:15 |
| ankur-gupta-f4 | Yea some small ones for neutron lbaas alignment. K. Still in bed tho so it might be a bit :) | 15:15 |
| rm_work | i mean | 15:16 |
| rm_work | that's where I do 100% of my work | 15:16 |
| *** aojea_ has quit IRC | 15:33 | |
| rm_work | https://gist.github.com/anonymous/da75fd830364c303a5f5c99b46de3762 | 15:34 |
| rm_work | ^^ if someone wants to look at outputs from the v2 patches in octavia, that's what they do | 15:34 |
| rm_work | in their *current* state | 15:34 |
| rm_work | let me know if you see something wrong | 15:34 |
| rm_work | I'm honestly not 100% sure what they look like in v2, i really don't like the API a whole lot, it feels very compromisey to me | 15:35 |
| johnsom | I have a neutron-lbaas VM running I use to compare as I go along | 15:38 |
| rm_work | ah that is a good idea | 15:40 |
| rm_work | also i should do deletes and unhappy-path | 15:40 |
| rm_work | but | 15:40 |
| rm_work | ugh | 15:40 |
| rm_work | johnsom: I am really hopeful we can merge listeners today! what do you think? | 15:41 |
| *** ndahiwade has joined #openstack-lbaas | 15:41 | |
| rm_work | I don't know if there are any major issues remaining? | 15:41 |
| ankur-gupta-f4 | The changes i put in take care of some of the null values returned by listeners | 15:41 |
| rm_work | ankur-gupta-f1: what was it that you had to add still? | 15:41 |
| ankur-gupta-f4 | Im out of my 1-1 | 15:41 |
| rm_work | err, wrong ankur-gupta-f1 apparently | 15:41 |
| rm_work | ankur-gupta-f4 | 15:41 |
| rm_work | there are 1-4 of you? :P | 15:41 |
| ankur-gupta-f4 | And unskipped 2 functional tests | 15:41 |
| rm_work | ah cool | 15:41 |
| rm_work | yeah if you want help with getting out your changes | 15:42 |
| ankur-gupta-f4 | Yea i think i left some clients running at homein texas | 15:42 |
| rm_work | i can help | 15:42 |
| *** armax has joined #openstack-lbaas | 15:42 | |
| rm_work | do you know how to use reflog and make a patch from a diff against that? | 15:42 |
| johnsom | Yeah, I will work on testing listeners today | 15:42 |
| ankur-gupta-f4 | Yea let me get my coffee n hop onto comp n will let u know. | 15:42 |
| johnsom | I finished up the load balancer API section last night | 15:43 |
| rm_work | k | 15:43 |
| ankur-gupta-f4 | I do not so will need ur help | 15:43 |
| rm_work | ok cool | 15:43 |
| rm_work | git diff $(git reflog | grep "commit (amend)" -A1 | head -n2 | tail -n1 | awk '{print $1}') | pbcopy | 15:49 |
| rm_work | should put your changes in your paste-buffer | 15:49 |
| rm_work | then just do a git review -d again to get the latest | 15:49 |
| rm_work | and: | 15:49 |
| rm_work | pbpaste | git apply | 15:49 |
| rm_work | magic | 15:49 |
| rm_work | my bash might not be super clean there, but that works for me basically 100% of the time | 15:50 |
| rm_work | basically that gets all your changes from your latest amend | 15:52 |
| *** KeithMnemonic has joined #openstack-lbaas | 16:02 | |
| johnsom | I just posted some initial issues with the listeners patch | 16:03 |
| ankur-gupta-f4 | Im on it boss! | 16:05 |
| rm_work | k | 16:05 |
| rm_work | you got it ankur-gupta-f4? | 16:05 |
| johnsom | We updated the types, but didn't update the controller | 16:05 |
| rm_work | ah when did functionals stop working | 16:05 |
| rm_work | i thought i did that? :/ | 16:05 |
| rm_work | oh | 16:06 |
| rm_work | i didn't even look at the stuff outside of post/put | 16:06 |
| rm_work | >_< | 16:06 |
| rm_work | so ankur-gupta-f4 are you going to fix these things too? or just post your updates? | 16:06 |
| rm_work | johnsom: some counter-comments | 16:08 |
| johnsom | Ok | 16:08 |
| ankur-gupta-f4 | I can do it. Fixing functional tests too | 16:08 |
| rm_work | k | 16:09 |
| rm_work | note my comments | 16:09 |
| ankur-gupta-f4 | Did u put on the patch? | 16:09 |
| rm_work | yes | 16:09 |
| johnsom | Ok, I was just checking on those. I guess I haven't got to the conversion code in my review yet | 16:09 |
| rm_work | next to johnsom's | 16:09 |
| rm_work | yeah it's in the db_prepare | 16:10 |
| rm_work | which happens before the validate | 16:10 |
| openstackgerrit | Nir Magnezi proposed openstack/octavia master: Auto detect haproxy user_group https://review.openstack.org/429398 | 16:10 |
| ankur-gupta-f4 | rm_work: if u r sick of api this cud use eyes till i do the listeners stuff, https://review.openstack.org/#/c/448331/ | 16:16 |
| rm_work | hmm k | 16:17 |
| rm_work | i don't understand why we need all this boilerplate, if this is just a plugin to OSC? | 16:18 |
| rm_work | can't we just import anything we need from OSC? | 16:18 |
| * rm_work doesn't understand how OSC plugins work, exactly | 16:18 | |
| *** gcheresh_ has quit IRC | 16:22 | |
| johnsom | ankur-gupta-f4 FYI, keep refreshing the listeners API, I am continuing to review so may post new comments | 16:22 |
| ankur-gupta-f4 | Yea. The functional tests may take awhile. Plus changing all return status codes as well | 16:23 |
| johnsom | Ok, just wanted to give you a heads up, I am still whining about the patch..... Grin | 16:24 |
| rm_work | ugh the status code changes are killing me | 16:26 |
| rm_work | going from correct statuses to purposefully incorrect statuses | 16:26 |
| johnsom | I know. | 16:26 |
| rm_work | to be backwards compatible... >_> | 16:27 |
| johnsom | But wrath has come upon those in other projects that tried to fix this exact issue | 16:27 |
| johnsom | Basically they are being forced into microversions | 16:27 |
| johnsom | Hmm, yeah, the page on "how to implement a plugin" | 16:31 |
| johnsom | https://docs.openstack.org/developer/python-openstackclient/plugins.html | 16:31 |
| johnsom | Talks about using osc_lib | 16:32 |
| rm_work | ok so yeah, this is the better way to do it then | 16:33 |
| johnsom | Yeah, duplicated code | 16:33 |
| rm_work | yeah k | 16:34 |
| rm_work | so my gut WAS right | 16:34 |
| ankur-gupta-f4 | Yea. That stuff sits in my WIP patch | 16:34 |
| ankur-gupta-f4 | https://review.openstack.org/#/c/446223/ | 16:37 |
| rm_work | why do you have both that AND a patch for the skeleton? | 16:41 |
| rm_work | I think you ONLY need one or the other? | 16:41 |
| *** aojea has joined #openstack-lbaas | 16:41 | |
| ankur-gupta-f4 | Will look later afternoon | 16:41 |
| rm_work | yeah, priority IMO is the API patches | 16:42 |
| johnsom | Hmm, so nlbaas we don't let them PUT the protocol or protocol_port.... | 16:49 |
| johnsom | I am trying to think of the ramifications if we allow that now.... | 16:50 |
| ankur-gupta-f4 | running into a fun functional error for 'create_with_one_listener' . Since listener now requires loadbalancer_id. | 16:53 |
| ankur-gupta-f4 | johnsom: | 16:53 |
| rm_work | johnsom: right, didn't i remove that? | 16:53 |
| ankur-gupta-f4 | and rm_work: | 16:53 |
| rm_work | ankur-gupta-f4: are you trying to un-skip the graph tests? | 16:54 |
| johnsom | rm_work no, it's still there in patch 30 | 16:54 |
| rm_work | I don't think those will work yet | 16:54 |
| rm_work | we haven't added the graph-create stuff | 16:54 |
| ankur-gupta-f4 | k undoing skip. will just address johnsoms questions and push up | 16:55 |
| ankur-gupta-f4 | comments* | 16:55 |
| johnsom | rm_work If you have a few minutes for review, I would welcome feedback on https://review.openstack.org/438757 | 16:59 |
| rm_work | yeah I am doing it now | 16:59 |
| rm_work | it is taking me a while because i am being incredibly nitpicky | 17:00 |
| johnsom | Ha! | 17:00 |
| rm_work | lots of: nit: add "the" | 17:00 |
| johnsom | Totally cool with that | 17:00 |
| *** eezhova has quit IRC | 17:00 | |
| johnsom | It's a lot of content spread over a lot of distractions.... | 17:01 |
| rm_work | wtf is "flavor" | 17:01 |
| rm_work | did we merge that? | 17:01 |
| rm_work | I thought we didn't have flavors yet | 17:01 |
| johnsom | We do have it, it did merge, but it's current form is pretty much useless | 17:02 |
| rm_work | T_T | 17:02 |
| *** catintheroof has joined #openstack-lbaas | 17:07 | |
| openstackgerrit | Ankur proposed openstack/octavia master: Octavia v2 API for listeners https://review.openstack.org/424744 | 17:12 |
| ankur-gupta-f4 | just pushed up. passing tests locally. lets see what happens on Jenkins | 17:12 |
| ankur-gupta-f4 | I really want to unskip more tests | 17:13 |
| ankur-gupta-f4 | but i guess we cant till graph | 17:13 |
| *** amotoki has quit IRC | 17:14 | |
| *** pcaruana has quit IRC | 17:14 | |
| *** eezhova has joined #openstack-lbaas | 17:16 | |
| *** kobis1 has quit IRC | 17:17 | |
| *** kobis has joined #openstack-lbaas | 17:19 | |
| *** eezhova has quit IRC | 17:24 | |
| rm_work | johnsom: ok i am about to post for at least the first few sections | 17:27 |
| johnsom | Ok | 17:27 |
| rm_work | posted | 17:29 |
| rm_work | lots of nits, have fun :P | 17:30 |
| johnsom | Grin | 17:30 |
| rm_work | for the status codes, I don't actually know what's correct anymore since we're changing a bunch of stuff to be purposefully incorrect per HTTP spec | 17:32 |
| rm_work | soooo | 17:32 |
| rm_work | I am just assuming you put the right ones :/ | 17:32 |
| johnsom | I am matching neutron-lbaas and adding any that I know Octavia may return for errors. | 17:33 |
| rm_work | k | 17:33 |
| rm_work | hmm PUT/DELETE seem fine actually | 17:35 |
| johnsom | Feel free to put in comments about questionable status codes. I don't mind double checking them | 17:35 |
| johnsom | Ha, that is all.... I put more comments than that on German's operator guide.... Grin | 17:36 |
| xgerman | ha | 17:36 |
| rm_work | lol | 17:37 |
| rm_work | oh, cool | 17:37 |
| rm_work | this fixes that dumb bug in neutron-lbaas where the status call doesn't actually show members' operating_status | 17:37 |
| rm_work | which i almost went in and fixed | 17:37 |
| rm_work | but | 17:37 |
| rm_work | i don't care because neutron-lbaas needs to die | 17:37 |
| openstackgerrit | Ankur proposed openstack/octavia master: [WIP] LB/Listener Stats https://review.openstack.org/449243 | 17:40 |
| openstackgerrit | Ankur proposed openstack/octavia master: [WIP] Add loadbalancer status tree call https://review.openstack.org/433982 | 17:42 |
| johnsom | Oh yeah, the quotes thing. I forgot that I started fixing those and didn't finish... | 17:43 |
| ankur-gupta-f4 | purging env rebuilding for testing purposes | 17:44 |
| rm_work | ankur-gupta-f4: gonna rebase the next two again | 17:47 |
| rm_work | after i make sure they actually pass | 17:48 |
| ankur-gupta-f4 | the next two? | 17:49 |
| ankur-gupta-f4 | which two | 17:49 |
| rm_work | pools/members | 17:49 |
| rm_work | ah they fail functionals :/ | 17:49 |
| rm_work | fff | 17:49 |
| ankur-gupta-f4 | just double check with ndahiwade and sshank in case they are working on it currently | 17:50 |
| rm_work | wait so | 17:51 |
| rm_work | ALL creates are 201??? | 17:51 |
| rm_work | not 202? | 17:51 |
| rm_work | ugh | 17:51 |
| *** links has joined #openstack-lbaas | 17:52 | |
| rm_work | ndahiwade / sshank ? | 17:56 |
| rm_work | i can push up rebases, it won't matter if they want to override them with their own patches later | 17:56 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Introduce Octavia v2 API for Members https://review.openstack.org/405932 | 17:56 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Introduce Octavia v2 API for pools https://review.openstack.org/405922 | 17:56 |
| sshank | rm_work, we were out for lunch. | 17:57 |
| rm_work | no worries | 17:57 |
| rm_work | was just seeing if it was safe to push some changes | 17:57 |
| rm_work | if you were out for lunch, i assume the answer was "yes" :{ | 17:57 |
| rm_work | :P | 17:57 |
| sshank | yes. | 17:57 |
| sshank | :D | 17:57 |
| rm_work | speaking of eating... i should go to sleep soon, and pop on in the morning to check on status | 17:58 |
| rm_work | err, actually, that doesn't have anything to do with eating <_< | 17:58 |
| openstackgerrit | German Eichberger proposed openstack/octavia master: Initial version of the operator guide https://review.openstack.org/448759 | 17:58 |
| *** links has quit IRC | 17:58 | |
| xgerman | pep8 couldn’t access something on my box… so rolling the dice | 17:59 |
| *** SumitNaiksatam has joined #openstack-lbaas | 18:00 | |
| *** gcheresh_ has joined #openstack-lbaas | 18:03 | |
| rm_work | <_< | 18:05 |
| rm_work | xgerman: fails for me too | 18:06 |
| rm_work | ah yeah lots of pep8 issues | 18:06 |
| xgerman | time to start drinking… | 18:06 |
| rm_work | https://gist.github.com/rm-you/1231fd50bfdc501c039f92d1d3eeeaf0 | 18:07 |
| rm_work | not too bad i guess | 18:07 |
| *** catintheroof has quit IRC | 18:07 | |
| xgerman | ok, better fix that | 18:07 |
| *** nakul_d has joined #openstack-lbaas | 18:11 | |
| *** eezhova has joined #openstack-lbaas | 18:25 | |
| *** kobis has quit IRC | 18:25 | |
| *** m-greene has joined #openstack-lbaas | 18:48 | |
| *** m-greene_ has joined #openstack-lbaas | 18:49 | |
| openstackgerrit | German Eichberger proposed openstack/octavia master: Initial version of the operator guide https://review.openstack.org/448759 | 18:51 |
| xgerman | https://www.irccloud.com/pastebin/CGK8JfsO/ | 18:51 |
| johnsom | That is neat | 18:58 |
| johnsom | That is some scripts we pulled in from neutron | 18:58 |
| johnsom | Oh, I think I have seen that before with something. It's an issue with the version of xargs. You are running that on a mac native aren't you.... | 19:00 |
| johnsom | I think the solution is to go get the xargs from the "fix macOS" repository | 19:01 |
| *** gcheresh_ has quit IRC | 19:04 | |
| *** kobis has joined #openstack-lbaas | 19:08 | |
| KeithMnemonic | is that the guide, i said i would walk though? or was it a differnt one. finally up in SUSE so i can now work on this | 19:14 |
| *** eezhova has quit IRC | 19:19 | |
| *** eezhova has joined #openstack-lbaas | 19:23 | |
| *** tesseract has quit IRC | 19:28 | |
| *** dougwig has joined #openstack-lbaas | 19:39 | |
| johnsom | KeithMnemonic That is a different guide. German is working on a high level operator guide (how to do some maintenance). You signed up for a setup/install guide | 19:45 |
| johnsom | If I remember correctly | 19:45 |
| KeithMnemonic | ok thanks for the reminder ;-) | 19:45 |
| KeithMnemonic | it took us a while to get settled into SUSE | 19:45 |
| johnsom | I hope it is going well so far | 19:46 |
| johnsom | FYI: https://docs.openstack.org/contributor-guide/project-install-guide.html | 19:46 |
| johnsom | That is kind of the starting point for the install guide work | 19:46 |
| johnsom | Our notes from the PTG are here:https://etherpad.openstack.org/p/octavia-ptg-pike | 19:47 |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Adds v2 load balancer API section https://review.openstack.org/438757 | 19:58 |
| KeithMnemonic | going well, you know "drinking from the firehose" | 20:02 |
| *** KeithMnemonic has quit IRC | 20:29 | |
| *** blogan_ has joined #openstack-lbaas | 20:54 | |
| *** blogan has quit IRC | 20:55 | |
| openstackgerrit | Ankur proposed openstack/octavia master: Introduce TESTING.rst for quick developer reference https://review.openstack.org/403938 | 21:09 |
| *** blogan_ is now known as blogan | 21:10 | |
| *** eezhova has quit IRC | 21:11 | |
| reedip | o/ | 22:14 |
| *** aojea has quit IRC | 22:18 | |
| johnsom | o/ | 22:26 |
| *** fnaval has quit IRC | 22:41 | |
| *** fnaval has joined #openstack-lbaas | 22:50 | |
| *** catintheroof has joined #openstack-lbaas | 23:13 | |
| *** catintheroof has quit IRC | 23:20 | |
| *** catintheroof has joined #openstack-lbaas | 23:21 | |
| *** catintheroof has quit IRC | 23:33 | |
| *** catintheroof has joined #openstack-lbaas | 23:38 | |
| *** catintheroof has quit IRC | 23:45 | |
| *** armax has quit IRC | 23:55 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!