| bana_k | hi, what does admin_state_up for LB means ? | 00:09 |
|---|---|---|
| johnsom | Enabled | 00:09 |
| bana_k | I have updated the admin_state_up as False and I am still able to load balance the traffic | 00:09 |
| johnsom | admin_state_up = False means it is disabled. It is still configured, but will not pass traffic | 00:09 |
| johnsom | That should not be the case. Which version of the code? | 00:10 |
| johnsom | Older releases had some bugs around that | 00:10 |
| bana_k | let me check | 00:10 |
| bana_k | Dec 16 | 00:10 |
| rm_work | somehow doing devstack on my local machine in a VM is not a lot faster >_> | 00:10 |
| bana_k | commit id 34873f99bc14ff69b54ea6960a90cdbc9b0c2275 | 00:10 |
| johnsom | Hmm, so that is mitaka. Looking | 00:12 |
| bana_k | ok. | 00:12 |
| johnsom | Octavia driver? | 00:13 |
| bana_k | yes | 00:13 |
| johnsom | Yeah, I don't think that fix made it back to Mitaka | 00:15 |
| johnsom | The fix was here: https://review.openstack.org/#/c/364707/ | 00:15 |
| johnsom | You can search your git history for that patch | 00:15 |
| bana_k | thanks johnsom, ll look | 00:15 |
| rm_work | yeah i pity people running on liberty :/ | 00:16 |
| johnsom | Yeah, liberty is EOL, so hopefully folks have moved on | 00:17 |
| rm_work | I wonder <_< | 00:18 |
| *** ducttape_ has joined #openstack-lbaas | 00:20 | |
| rm_work | so our py3x tests are broken because of nova, right? | 00:23 |
| johnsom | Yeah, last time I looked through the logs | 00:23 |
| rm_work | oh maybe it isn't | 00:26 |
| rm_work | i assumed it was at first but | 00:26 |
| rm_work | it might be our bad | 00:27 |
| rm_work | it's actually here: http://logs.openstack.org/47/382147/16/check/gate-neutron-lbaasv2-dsvm-py3x-api-ubuntu-xenial-nv/40de556/logs/screen-o-cw.txt.gz#_2017-01-19_22_10_40_391 | 00:27 |
| rm_work | which is on our side... gonna take a look | 00:28 |
| *** ducttape_ has quit IRC | 00:31 | |
| rm_work | yeah ok | 00:33 |
| rm_work | this is borked | 00:36 |
| *** gongysh has joined #openstack-lbaas | 00:37 | |
| rm_work | johnsom: opinion | 00:41 |
| rm_work | johnsom: a file uses both load_balancer and loadbalancer | 00:41 |
| rm_work | which one should I normalize to | 00:41 |
| rm_work | I prefer loadbalancer | 00:41 |
| rm_work | but I am not sure as a project | 00:41 |
| johnsom | Yeah, we haven't exactly been careful about that. I'm fine with loadbalancer | 00:42 |
| rm_work | k | 00:43 |
| rm_work | ugh yeah our VIP datamodel uses load_balancer | 00:43 |
| rm_work | whelp | 00:43 |
| rm_work | so i have to leave the keyword alone T_T | 00:43 |
| rm_work | anyway this might fix our py3x | 00:44 |
| rm_work | gonna push this up and grab dinner while i wait for zuul | 00:44 |
| rm_work | it's a bug in the noop driver <_< | 00:44 |
| johnsom | Ok, yeah, that looks farther than it got last time I looked. | 00:46 |
| johnsom | I would like to get some of these off non-voting, but I think we need to fix that too long bug first. | 00:46 |
| openstackgerrit | Adam Harwell proposed openstack/octavia: Cleanup noop network driver to fix py3x https://review.openstack.org/422942 | 00:48 |
| *** fnaval has quit IRC | 00:48 | |
| *** yamamoto_ has joined #openstack-lbaas | 00:51 | |
| johnsom | Ugh, the anchor doc doesn't render well. I'm going to fix that in the doc patch as well. | 00:51 |
| rm_work | ok | 00:51 |
| xgerman | nice | 00:51 |
| *** gongysh has quit IRC | 00:52 | |
| *** bradjones has quit IRC | 00:55 | |
| rm_work | johnsom: ok so, "marker" is supposed to be a number... like ... from the "Nth" entry | 01:06 |
| rm_work | though ... thinking about it that's ... | 01:06 |
| rm_work | kinda add | 01:06 |
| rm_work | *odd | 01:06 |
| rm_work | is that how neutron does it??? | 01:06 |
| johnsom | No, marker is the id | 01:06 |
| rm_work | uhh | 01:06 |
| rm_work | telling you quite assuredly | 01:06 |
| rm_work | from the code | 01:06 |
| rm_work | it isn't | 01:06 |
| rm_work | it's a number | 01:07 |
| rm_work | it slices the resultset | 01:07 |
| johnsom | I did in fact us the neutron client to get my test calls | 01:07 |
| rm_work | if that's not how neutron does it... then it was implemented wrong | 01:07 |
| rm_work | our implementation uses integer offset | 01:07 |
| johnsom | Yeah, that isn't right | 01:07 |
| rm_work | so if that's not what is supposed to happenb | 01:07 |
| rm_work | this whole thing needs to be reworked | 01:08 |
| rm_work | well #@$% | 01:08 |
| rm_work | alright | 01:08 |
| rm_work | I guess I can poke at this tomorrow <_< | 01:09 |
| rm_work | wish I could ask carlos why he did it this way | 01:11 |
| rm_work | but i don't think he's been on IRC in a while | 01:11 |
| johnsom | http://developer.openstack.org/api-ref/networking/v2/?expanded=list-subnets-detail#pagination | 01:11 |
| johnsom | "The marker parameter is the ID of the last item in the previous list. " | 01:12 |
| johnsom | That is what I was going on to test | 01:12 |
| rm_work | yeah this is not how he did it at all | 01:12 |
| *** amotoki has joined #openstack-lbaas | 01:15 | |
| rm_work | ok so I *thought* carlos was just pulling in the sorting/paging code from neutron | 01:21 |
| rm_work | like | 01:21 |
| rm_work | copy/paste | 01:21 |
| rm_work | but the neutron paging/sorting code looks like it's architected ENTIRELY differently <_< | 01:21 |
| johnsom | I actually pointed him at glance since it had cleaner code, but I don't think that is what happened. | 01:21 |
| rm_work | ah k | 01:22 |
| rm_work | i'll look at glance | 01:22 |
| johnsom | glance implemented both the old way (neutron) and the new api-wg way for sorting | 01:22 |
| rm_work | hmm k | 01:22 |
| rm_work | maybe it's the new way | 01:22 |
| johnsom | Well, pagination is um, yeah, openstack style: https://specs.openstack.org/openstack/api-wg/guidelines/pagination_filter_sort.html#pagination | 01:23 |
| johnsom | So... I don't know that there is a new way for that | 01:23 |
| rm_work | Pagination is just ... TODO | 01:24 |
| rm_work | lol | 01:24 |
| rm_work | yeah this looks ... different too? | 01:26 |
| rm_work | er | 01:26 |
| rm_work | ok i'm going to figure it out tomorrow | 01:26 |
| rm_work | maybe just start over | 01:26 |
| rm_work | you're correct, sorting doesn't actually do anything | 01:27 |
| rm_work | and pagination is ... <_< | 01:27 |
| johnsom | Well, on sorting, lubosz pointed out that I was using the "old" sorting syntax and not the new one listed on that page I linked above. | 01:28 |
| rm_work | I still love: | 01:28 |
| johnsom | So, it might be working ok | 01:28 |
| rm_work | 2017-01-20 00:51:06.087254 | Setting up the workspace | 01:28 |
| rm_work | 2017-01-20 00:51:06.087374 | ... this takes 3 - 5 minutes (logs at logs/devstack-gate-setup-workspace-new.txt.gz) | 01:28 |
| rm_work | so much lies | 01:28 |
| rm_work | I kinda want to submit a patch to devstack to change that to "30-45 minutes" | 01:28 |
| rm_work | OH yeah you are, lol | 01:28 |
| rm_work | didn't even look, just copied your queries | 01:29 |
| johnsom | Yeah, it is definitely 30 minutes. Though, in fairness, 10-15 of that is probably our image building | 01:29 |
| rm_work | yep sorting works | 01:29 |
| johnsom | Yeah, I just grabbed them from neutron client | 01:29 |
| rm_work | uhh but | 01:29 |
| rm_work | doesn't that mean we need to support the old way too | 01:29 |
| johnsom | I.e. trying to verify compat | 01:29 |
| rm_work | to be non-breaking? | 01:29 |
| rm_work | right | 01:29 |
| rm_work | .... | 01:29 |
| *** bana_k has quit IRC | 01:35 | |
| *** kevo has quit IRC | 01:35 | |
| *** yamamoto_ has quit IRC | 01:38 | |
| *** woodster_ has quit IRC | 01:45 | |
| *** ducttape_ has joined #openstack-lbaas | 01:55 | |
| *** ducttape_ has quit IRC | 01:59 | |
| *** gongysh has joined #openstack-lbaas | 02:11 | |
| *** ducttape_ has joined #openstack-lbaas | 02:13 | |
| *** reedip_ has joined #openstack-lbaas | 02:19 | |
| openstackgerrit | Nir Magnezi proposed openstack/octavia: WIP - Fix the amphora-agent support for RH based Linux flavors https://review.openstack.org/331841 | 02:21 |
| *** yamamoto_ has joined #openstack-lbaas | 02:26 | |
| *** beardedeagle has joined #openstack-lbaas | 02:28 | |
| *** reedip_ has quit IRC | 02:32 | |
| *** links has joined #openstack-lbaas | 02:53 | |
| *** ducttape_ has quit IRC | 02:54 | |
| *** bana_k has joined #openstack-lbaas | 03:33 | |
| *** armax has joined #openstack-lbaas | 03:40 | |
| *** amotoki has quit IRC | 03:49 | |
| *** links has quit IRC | 04:16 | |
| *** links has joined #openstack-lbaas | 04:18 | |
| *** amotoki has joined #openstack-lbaas | 04:19 | |
| *** amotoki_ has joined #openstack-lbaas | 04:25 | |
| *** bana_k has quit IRC | 04:26 | |
| *** amotoki has quit IRC | 04:27 | |
| *** beardedeagle has quit IRC | 04:30 | |
| *** bana_k has joined #openstack-lbaas | 04:52 | |
| *** catinthe_ has quit IRC | 04:54 | |
| *** catintheroof has joined #openstack-lbaas | 05:02 | |
| *** jerrygb has quit IRC | 05:05 | |
| *** bana_k has quit IRC | 05:11 | |
| *** bana_k has joined #openstack-lbaas | 05:11 | |
| rm_work | johnsom: this is port merge? http://logs.openstack.org/42/422942/1/check/gate-octavia-v1-dsvm-scenario-ubuntu-xenial-nv/9aa9599/console.html#_2017-01-20_01_52_38_287026 | 05:12 |
| rm_work | http://logs.openstack.org/42/422942/1/check/gate-octavia-v1-dsvm-scenario-ubuntu-xenial-nv/9aa9599/logs/testrepository.0.gz | 05:12 |
| johnsom | Sweet, I will work on it tomorrow | 05:13 |
| rm_work | lol didn't expect you to be around | 05:13 |
| rm_work | i was gonna take a glance now | 05:13 |
| *** catintheroof has quit IRC | 05:14 | |
| rm_work | also, i guess i only fixed one of the py3x tests | 05:14 |
| rm_work | so, one passes, but the rest don | 05:14 |
| rm_work | *don't (still nova i think) | 05:14 |
| rm_work | progress I guess | 05:14 |
| rm_work | https://review.openstack.org/#/c/422942/ | 05:14 |
| rm_work | i'll keep poking | 05:14 |
| rm_work | it seems the VM spun up but SSH doesn't get through right | 05:14 |
| johnsom | Ok, let me know what you find and if I should pick it up in the morning | 05:15 |
| rm_work | kk | 05:15 |
| rm_work | johnsom: it seems like just an insane amount of | 05:17 |
| rm_work | 2017-01-20 01:49:11,322 3281 INFO [octavia.tests.tempest.v1.scenario.base] Got IOError in check connection: <urlopen error [Errno 111] Connection refused> | 05:17 |
| rm_work | which is basically the VMs taking forever to spin up right? | 05:17 |
| rm_work | which is ... arbitrary kinda, so it explains why it's just *sometimes*.... if it's particularly slow, it goes over the threshold | 05:17 |
| rm_work | ah and it's trying the connection once every like ......... | 05:18 |
| rm_work | 0.001 seconds | 05:18 |
| rm_work | so ever 0.001 seconds it spits a log line | 05:18 |
| johnsom | Yeah, not sure. I think I added that log message to track down some of the flask issues | 05:18 |
| rm_work | 1000 lines per second | 05:18 |
| rm_work | that's why it gets so hufe | 05:18 |
| rm_work | *huge | 05:18 |
| johnsom | Hmm, that is odd | 05:18 |
| rm_work | there's like 60,000 lines of this | 05:18 |
| rm_work | more | 05:19 |
| rm_work | it's like 95% of the log | 05:19 |
| johnsom | I know we dropped that wait time, but umm | 05:19 |
| rm_work | first: | 05:19 |
| johnsom | Well, that should make it easier to fix | 05:19 |
| rm_work | 2017-01-20 01:47:27,361 3281 INFO [octavia.tests.tempest.v1.scenario.base] checking connection to ip: 172.24.5.13 port: 80 | 05:19 |
| rm_work | last: | 05:19 |
| rm_work | 2017-01-20 01:51:48,823 3279 INFO [octavia.tests.tempest.v1.scenario.base] Got Error in sending request: <urlopen error timed out> | 05:19 |
| rm_work | is from like line 700 to line 81000 | 05:20 |
| rm_work | so. yeah. | 05:20 |
| rm_work | that should be an easy fix | 05:20 |
| rm_work | was there a problem trying once per second? lolol | 05:20 |
| johnsom | I think that was your change, to not wait durning tests any longer than needed... grin | 05:21 |
| rm_work | lol down to 0.001?!? | 05:22 |
| johnsom | I don't know. We may not need to log that either... | 05:22 |
| rm_work | gonna look for where it is | 05:22 |
| rm_work | for example BTW: http://paste.openstack.org/show/595721/ | 05:22 |
| johnsom | Hmm, what I am thinking of was a mocked out sleep, so it would be odd that would impact the scenario tests | 05:23 |
| rm_work | ah found it | 05:23 |
| rm_work | there's no wait *at all* | 05:23 |
| rm_work | and since it bails instantly because it's an actual IOError not a timeout | 05:23 |
| rm_work | it's as past as python can send connections | 05:23 |
| rm_work | *as fast as | 05:23 |
| rm_work | https://github.com/openstack/octavia/blob/master/octavia/tests/tempest/v1/scenario/base.py#L615-L630 | 05:24 |
| rm_work | oh wait, what calls this | 05:24 |
| rm_work | sorry it's here: | 05:24 |
| rm_work | err | 05:25 |
| rm_work | like 5 lines below | 05:25 |
| rm_work | same thing, no sleep at all | 05:25 |
| rm_work | I might add a sleep(0.5)? maybe? | 05:25 |
| rm_work | is that kosher? | 05:25 |
| johnsom | Yeah, go for it. That is code I am working on for octavia v2 as well, so I will pull that in there as well. | 05:26 |
| johnsom | Heck, 1 second is probably just fine too | 05:27 |
| rm_work | or I can have it try as fast as it wants but only print the message every 1s or so | 05:27 |
| rm_work | i dunno | 05:27 |
| rm_work | eh just doing a 1s sleep is going to be fine | 05:27 |
| * rm_work watches it explode 6 months later | 05:27 | |
| johnsom | Just have it sleep, it is easy and polite to the other vms | 05:27 |
| rm_work | did we have a bug filed? | 05:29 |
| johnsom | Yes, I opened on | 05:29 |
| johnsom | One | 05:29 |
| rm_work | looking | 05:29 |
| rm_work | nothing matches subunit... | 05:30 |
| *** cody-somerville has joined #openstack-lbaas | 05:31 | |
| johnsom | Hmm, I think it closed with my first archive patch | 05:31 |
| rm_work | lol ok whatever | 05:31 |
| openstackgerrit | Adam Harwell proposed openstack/octavia: Fix connection timeout spam in subunit logs https://review.openstack.org/423041 | 05:31 |
| rm_work | I guess tomorrow I'll keep looking at py3x and also try to untangle this sorting/paging mess | 05:32 |
| johnsom | Ok, nope. I am too tired. No bug | 05:32 |
| rm_work | night :) | 05:33 |
| *** csomerville has quit IRC | 05:33 | |
| johnsom | I will nit on that patch that we should sleep after checking for the timeout, but it is a nit | 05:34 |
| rm_work | ehhh | 05:34 |
| rm_work | so i debated | 05:34 |
| rm_work | it runs the connection attempt as the condition for the while | 05:35 |
| rm_work | so by the time you get to the sleep, it's done at least one try | 05:35 |
| rm_work | oh | 05:36 |
| rm_work | it should be in an else | 05:36 |
| rm_work | i read the second bit wrong | 05:36 |
| rm_work | then we can technically go over | 05:36 |
| rm_work | but | 05:36 |
| rm_work | it's kinda ... meh | 05:36 |
| johnsom | Yeah | 05:36 |
| rm_work | unless we subtract the delay from the timeout check | 05:36 |
| rm_work | but that's iffy too | 05:36 |
| rm_work | or in this case it'd be | 05:37 |
| johnsom | Over complicates it | 05:37 |
| rm_work | if (time.time() - start + delay) > timeout | 05:37 |
| rm_work | I mean... it'd work | 05:37 |
| rm_work | would you rather be conservative in one direction or the other? | 05:37 |
| rm_work | ah and technically it doesn't need to be an`else` I guess, since the `if` raises | 05:38 |
| johnsom | Right | 05:39 |
| *** jerrygb has joined #openstack-lbaas | 05:39 | |
| rm_work | k | 05:40 |
| openstackgerrit | Adam Harwell proposed openstack/octavia: Fix connection timeout spam in subunit logs https://review.openstack.org/423041 | 05:40 |
| rm_work | should be fine | 05:40 |
| rm_work | made it 1s too | 05:40 |
| rm_work | i think you're right | 05:40 |
| rm_work | oh i'll hit the v2 as well | 05:40 |
| *** anilvenkata has joined #openstack-lbaas | 05:40 | |
| openstackgerrit | Adam Harwell proposed openstack/octavia: Fix connection timeout spam in subunit logs https://review.openstack.org/423041 | 05:41 |
| johnsom | Ok. I am going to go back to reading the news. Catch you tomorrow | 05:41 |
| *** catintheroof has joined #openstack-lbaas | 05:46 | |
| *** saju_m has joined #openstack-lbaas | 06:09 | |
| *** sticker has quit IRC | 06:12 | |
| *** armax has quit IRC | 06:19 | |
| *** armax has joined #openstack-lbaas | 06:20 | |
| *** armax has quit IRC | 06:20 | |
| *** saju_m has quit IRC | 06:59 | |
| *** kevo has joined #openstack-lbaas | 07:35 | |
| *** pcaruana has joined #openstack-lbaas | 07:43 | |
| *** afranc has joined #openstack-lbaas | 07:44 | |
| *** Aju has quit IRC | 07:51 | |
| *** bana_k has quit IRC | 07:57 | |
| *** tesseract has joined #openstack-lbaas | 08:05 | |
| *** gcheresh_ has joined #openstack-lbaas | 08:10 | |
| *** eezhova has joined #openstack-lbaas | 08:17 | |
| *** cody-somerville has quit IRC | 08:20 | |
| *** cody-somerville has joined #openstack-lbaas | 08:21 | |
| *** cody-somerville has quit IRC | 08:21 | |
| *** cody-somerville has joined #openstack-lbaas | 08:21 | |
| *** openstackgerrit has quit IRC | 09:02 | |
| *** saju_m has joined #openstack-lbaas | 09:11 | |
| *** saju_m has quit IRC | 09:24 | |
| *** saju_m has joined #openstack-lbaas | 09:36 | |
| *** jerrygb_ has joined #openstack-lbaas | 09:41 | |
| *** jerrygb has quit IRC | 09:43 | |
| *** jerrygb has joined #openstack-lbaas | 09:46 | |
| *** jerrygb_ has quit IRC | 09:48 | |
| reedip | anyone here? | 10:05 |
| *** gcheresh_ has quit IRC | 10:06 | |
| *** jerrygb has quit IRC | 10:13 | |
| *** jerrygb has joined #openstack-lbaas | 10:14 | |
| *** gongysh has quit IRC | 10:22 | |
| *** gongysh has joined #openstack-lbaas | 10:28 | |
| *** amotoki_ has quit IRC | 10:28 | |
| *** gongysh has quit IRC | 10:37 | |
| *** amotoki has joined #openstack-lbaas | 10:38 | |
| *** kevo has quit IRC | 10:42 | |
| *** yamamoto_ has quit IRC | 10:50 | |
| *** ipsecguy_ has quit IRC | 10:51 | |
| *** ducttape_ has joined #openstack-lbaas | 11:29 | |
| *** ducttape_ has quit IRC | 11:31 | |
| *** ipsecguy has joined #openstack-lbaas | 11:32 | |
| *** yamamoto has joined #openstack-lbaas | 11:40 | |
| *** amotoki has quit IRC | 11:44 | |
| *** pcaruana has quit IRC | 11:59 | |
| *** yamamoto has quit IRC | 12:02 | |
| *** yamamoto has joined #openstack-lbaas | 12:03 | |
| *** pcaruana has joined #openstack-lbaas | 12:06 | |
| *** yamamoto has quit IRC | 12:08 | |
| *** yamamoto has joined #openstack-lbaas | 12:10 | |
| *** gcheresh_ has joined #openstack-lbaas | 12:15 | |
| *** catinthe_ has joined #openstack-lbaas | 12:32 | |
| *** catintheroof has quit IRC | 12:33 | |
| *** yamamoto has quit IRC | 12:44 | |
| *** yamamoto has joined #openstack-lbaas | 12:44 | |
| *** jerrygb_ has joined #openstack-lbaas | 12:46 | |
| *** yamamoto has quit IRC | 12:48 | |
| *** jerrygb has quit IRC | 12:49 | |
| *** yamamoto has joined #openstack-lbaas | 12:51 | |
| *** gcheresh_ has quit IRC | 12:52 | |
| *** gcheresh has joined #openstack-lbaas | 12:52 | |
| *** yamamoto has quit IRC | 12:52 | |
| *** links has quit IRC | 12:55 | |
| *** amotoki has joined #openstack-lbaas | 12:56 | |
| *** ducttape_ has joined #openstack-lbaas | 13:11 | |
| *** ducttape_ has quit IRC | 13:21 | |
| *** ducttape_ has joined #openstack-lbaas | 13:24 | |
| *** ducttape_ has quit IRC | 13:34 | |
| *** bradjones has joined #openstack-lbaas | 13:38 | |
| *** bradjones has quit IRC | 13:38 | |
| *** bradjones has joined #openstack-lbaas | 13:38 | |
| *** ipsecguy has quit IRC | 13:44 | |
| *** ipsecguy has joined #openstack-lbaas | 13:45 | |
| *** ipsecguy has quit IRC | 13:46 | |
| *** ipsecguy has joined #openstack-lbaas | 13:47 | |
| *** yamamoto has joined #openstack-lbaas | 13:52 | |
| *** yamamoto has quit IRC | 14:02 | |
| *** matt-borland has joined #openstack-lbaas | 14:06 | |
| *** beardedeagle has joined #openstack-lbaas | 14:16 | |
| *** woodster_ has joined #openstack-lbaas | 14:27 | |
| *** catintheroof has joined #openstack-lbaas | 14:33 | |
| *** catinthe_ has quit IRC | 14:36 | |
| *** yamamoto has joined #openstack-lbaas | 14:47 | |
| *** yamamoto has quit IRC | 14:53 | |
| *** ducttape_ has joined #openstack-lbaas | 14:53 | |
| johnsom | reedip: Hi | 15:02 |
| *** ducttape_ has quit IRC | 15:02 | |
| *** gcheresh has quit IRC | 15:17 | |
| *** fnaval has joined #openstack-lbaas | 15:24 | |
| *** amotoki has quit IRC | 15:24 | |
| *** ankur-gupta-f2 has joined #openstack-lbaas | 15:25 | |
| *** ducttape_ has joined #openstack-lbaas | 15:29 | |
| *** yamamoto has joined #openstack-lbaas | 15:54 | |
| *** anilvenkata has quit IRC | 15:55 | |
| *** pcaruana has quit IRC | 16:03 | |
| *** yamamoto has quit IRC | 16:03 | |
| *** amotoki has joined #openstack-lbaas | 16:05 | |
| *** amotoki has quit IRC | 16:06 | |
| *** armax has joined #openstack-lbaas | 16:14 | |
| *** openstackgerrit has joined #openstack-lbaas | 16:20 | |
| openstackgerrit | Michael Johnson proposed openstack/octavia: Fix the docs page title https://review.openstack.org/422884 | 16:20 |
| *** _ducttape_ has joined #openstack-lbaas | 16:59 | |
| *** ducttape_ has quit IRC | 17:02 | |
| *** bana_k has joined #openstack-lbaas | 17:09 | |
| *** reedip_ has joined #openstack-lbaas | 17:14 | |
| *** csomerville has joined #openstack-lbaas | 17:31 | |
| *** cody-somerville has quit IRC | 17:34 | |
| *** _ducttape_ has quit IRC | 17:35 | |
| *** ducttape_ has joined #openstack-lbaas | 17:36 | |
| *** kevo has joined #openstack-lbaas | 17:37 | |
| *** bana_k has quit IRC | 17:39 | |
| *** eezhova has quit IRC | 17:45 | |
| *** eezhova has joined #openstack-lbaas | 17:47 | |
| openstackgerrit | Nakul Dahiwade proposed openstack/octavia: Align Octavia v2 API for Members https://review.openstack.org/423393 | 17:51 |
| *** eezhova has quit IRC | 18:12 | |
| *** bana_k has joined #openstack-lbaas | 18:26 | |
| *** reedip_ has quit IRC | 18:40 | |
| *** tesseract has quit IRC | 19:32 | |
| *** catinthe_ has joined #openstack-lbaas | 19:50 | |
| blogan | johnsom: the octavia v2 api is upposed to have the /lbaas prefix correct? | 19:52 |
| blogan | like ip:port/v2.0/lbaas/loadbalancers? | 19:52 |
| blogan | and do we also need /v2.0? or /v2? | 19:52 |
| *** catintheroof has quit IRC | 19:53 | |
| rm_work | wait what | 20:11 |
| rm_work | hold on, what does neutron use right now | 20:11 |
| rm_work | is v2 BEFORE lbaas? | 20:11 |
| *** gcheresh has joined #openstack-lbaas | 20:23 | |
| blogan | 2.0 | 20:27 |
| blogan | v2.0 | 20:27 |
| rm_work | hmm | 20:27 |
| rm_work | i kinda thought we weren't going to deal with that cruft, and would recommend deployers do l7 routing to invisibly make things go to the right place | 20:29 |
| rm_work | but maybe not | 20:29 |
| sindhu | blogan: johnsom : shouldn't octavia v2 have: /v2.0/loadbalancers , /v2.0/listeners ... etc? | 20:29 |
| rm_work | ^^ so, this | 20:29 |
| blogan | well if our goal is to have it just work by changing the keystone catalog, then we'd have to probably maintain the /v2.0/lbaas/ portions | 20:30 |
| rm_work | we can't change the keystone catalog, can we? | 20:30 |
| rm_work | lbaas is part of neutron | 20:30 |
| blogan | but l7 routing would work too, but not as seamless | 20:30 |
| rm_work | until we have our own keystone type | 20:30 |
| rm_work | isn't it? | 20:30 |
| blogan | i mean we can always just ahve /v2.0/loadbalancers and /lbaas/v2.0/loadbalancers at the same times, thats just adding one line | 20:31 |
| rm_work | catalog has to continue to point to neutron | 20:31 |
| rm_work | and the deployer has to l7 route /lbaas to us | 20:31 |
| rm_work | AFAIU | 20:31 |
| rm_work | there's no "lbaas" keystone catalog entry | 20:31 |
| blogan | there's not, but eventually there will be | 20:32 |
| rm_work | and at that point we DEFINITELY don't want /lbaas in our URL do we? | 20:32 |
| blogan | i dont | 20:32 |
| blogan | i also don't want v2.0, just v2 | 20:32 |
| rm_work | any change to the clients we'd need to make to use a different endpoint name, would also just include the endpoint URL changing too, right? | 20:33 |
| rm_work | hmm, that's the way keystone and nova went with v3 right? | 20:33 |
| rm_work | not v3.0 | 20:33 |
| blogan | pretty sure | 20:33 |
| rm_work | so i'd say that's the future direction of things | 20:33 |
| rm_work | and while we're depending on l7 routing, we may as well just fix that too | 20:33 |
| blogan | we'll have to see what glorius dear leader johnsom says | 20:34 |
| rm_work | so yes, i'd be behind ip:port/v2/loadbalancers ip:port/v2/pools etc | 20:34 |
| rm_work | hopefully it's clear that almost all of my "questions" above were rhetorical >_> | 20:35 |
| *** catinthe_ has quit IRC | 21:02 | |
| *** catintheroof has joined #openstack-lbaas | 21:03 | |
| *** beardedeagle has quit IRC | 21:06 | |
| *** beardedeagle has joined #openstack-lbaas | 21:06 | |
| *** catintheroof has quit IRC | 21:07 | |
| *** beardedeagle has quit IRC | 21:18 | |
| *** beardedeagle has joined #openstack-lbaas | 21:18 | |
| johnsom | Just got back from lunch, reading the scroll back... | 21:23 |
| *** eezhova has joined #openstack-lbaas | 21:24 | |
| johnsom | So, a few comments here: | 21:24 |
| johnsom | 1. octavia will be a new service in keystone (it's private I think at the moment) | 21:25 |
| johnsom | 2. We need to maintain /v2.0/lbaas/loadbalancers as neutron-lbaas has it for some time. This is so apps can just switch the service/endpoint and still enjoy the neutron-lbaas API | 21:26 |
| johnsom | 3. We can also, optionally, alias /v2.0/lbaas/loadbalancers to /v2.0/loadbalancers and have that as a go forward path. Though, I think there is some value to having the lbaas in there for folks that want to do L7 games with their API servers, endpoints, or API security stuffs (i.e. Apigee, etc.) | 21:27 |
| *** eezhova has quit IRC | 21:28 | |
| johnsom | rm_work, blogan, sindhu: thoughts, comments? | 21:30 |
| blogan | we can easily have both with pecan | 21:31 |
| blogan | /lbaas/loadbalancers and /loadbalancers | 21:31 |
| rm_work | yeah that's fine | 21:32 |
| blogan | i guess we can also have both v2 and v2.0 | 21:32 |
| blogan | but that'd be a bit weird | 21:32 |
| johnsom | Yeah, that was kind of my thought on the issue. That doesn't break the API | 21:32 |
| blogan | well if we did /v2/loadbalancers and /v2.0/lbaas/loadbalancers that'd be fine by me | 21:32 |
| johnsom | We could have v2 and v2.0, but all I really care about is v2.0 | 21:32 |
| blogan | as long as we have a deprecation plan for the /v2.0/lbaas | 21:32 |
| johnsom | Aren't all the projects v2.0 vs v2? | 21:33 |
| rm_work | only because of timing | 21:33 |
| johnsom | Actually it looks 50/50 | 21:34 |
| blogan | keystone is v3 | 21:34 |
| johnsom | With microversioning dropping it all together right? | 21:34 |
| blogan | i think so | 21:34 |
| sindhu | /v2/loadbalancers and /v2.0/lbaas/loadbalancers: +1 | 21:35 |
| johnsom | Yeah, so, for compatibility, we need v2.0, v2 is optional | 21:35 |
| johnsom | So, current discussion is: | 21:36 |
| johnsom | /v2/loadbalancers | 21:37 |
| johnsom | /v2/lbaas/loadbalancers | 21:38 |
| johnsom | /v2.0/loadbalancers | 21:38 |
| johnsom | */v2.0/lbaas/loadbalnacers | 21:38 |
| rm_work | yes | 21:38 |
| johnsom | Where * is required for compatibility | 21:38 |
| johnsom | Seems like a lot of paths, but I'm fine with it. | 21:38 |
| *** gcheresh has quit IRC | 21:39 | |
| johnsom | MVP is the * one, the others are optional/nice to have | 21:39 |
| *** jerrygb_ has quit IRC | 21:41 | |
| johnsom | I should capture this somewhere. Where would be most helpful for folks? Edit the spec, add it to a bug, or create an etherpad? | 21:42 |
| rm_work | why is MVP always the worst one >_> | 21:42 |
| johnsom | Because it's the old one we have to be compatible with.... Grin | 21:43 |
| openstackgerrit | Michael Johnson proposed openstack/octavia: Fix the docs page title https://review.openstack.org/422884 | 21:43 |
| sindhu | johnsom: https://etherpad.openstack.org/p/octavia-ocata-merge-review-priority why not here? | 21:44 |
| johnsom | rm_work Your -1 nit.... Grin | 21:44 |
| rm_work | yeah sorry about that <_< | 21:44 |
| rm_work | it irked me | 21:44 |
| johnsom | sindhu Ok | 21:44 |
| sshank_ | q | 21:45 |
| *** _ducttape_ has joined #openstack-lbaas | 21:48 | |
| johnsom | Ok, added to the etherpad, let me know if you have questions | 21:49 |
| *** _ducttape_ has quit IRC | 21:49 | |
| *** _ducttape_ has joined #openstack-lbaas | 21:50 | |
| sshank_ | So pools or others will be of the format: /v2.0/lbaas/pools ? | 21:50 |
| *** ducttape_ has quit IRC | 21:51 | |
| *** ndahiwade has joined #openstack-lbaas | 21:55 | |
| ndahiwade | sshank_: yes | 21:56 |
| *** matt-borland has quit IRC | 22:23 | |
| *** armax has quit IRC | 22:26 | |
| *** jerrygb has joined #openstack-lbaas | 22:40 | |
| *** armax has joined #openstack-lbaas | 22:42 | |
| *** beardedeagle has quit IRC | 22:43 | |
| openstackgerrit | Nakul Dahiwade proposed openstack/octavia: Align Octavia v2 API for Members https://review.openstack.org/423393 | 22:47 |
| openstackgerrit | Brandon Logan proposed openstack/octavia: Add common base type for v1 and v2 https://review.openstack.org/423542 | 22:52 |
| openstackgerrit | Brandon Logan proposed openstack/octavia: Add v2 load balancer type and controllers https://review.openstack.org/423543 | 22:52 |
| openstackgerrit | Brandon Logan proposed openstack/octavia: Add v2 load balancer type and controllers https://review.openstack.org/423543 | 22:54 |
| blogan | johnsom, sindhu, sshank_, ndahiwade: ^^ thats a first pass at getting it working. API requests work just like neutron-lbaas v2 as far as i tested. | 22:55 |
| johnsom | Cool! Thanks! | 22:55 |
| blogan | i realized, that decorator isn't really needed as we don't need to change anything about the data models | 22:55 |
| blogan | we just need to do it all with the types | 22:55 |
| sshank_ | blogan, Thank you. | 22:55 |
| blogan | i need to add tests to these rviews too and probably clean up some more, maybe split it out into one more rview, not sure yet | 22:56 |
| ndahiwade | blogan, thanks:) | 22:57 |
| blogan | if every controller and type is changed for this, it's not too bad of a work load | 22:57 |
| sindhu | blogan: thank you :) | 22:57 |
| rm_work | blogan: want to merge https://review.openstack.org/#/c/423041/ ? :P | 22:58 |
| rm_work | stop getting annoying false-failures on our scenarios | 22:58 |
| blogan | rm_work: i usually charge johnsom one cookie for these requests, but for you...2 COOKIES! | 22:58 |
| rm_work | oh god | 22:59 |
| rm_work | that's pretty greedy blogan | 22:59 |
| rm_work | but if that's the price | 22:59 |
| blogan | i'm a capitalist pig | 22:59 |
| *** eezhova has joined #openstack-lbaas | 22:59 | |
| johnsom | rm_work Just send him one of these like I did: Set-Cookie: RackSID=20170117111850; path=/; domain=.rackspace.com | 23:00 |
| rm_work | lol | 23:00 |
| blogan | my internal validation has been updated to reject those cookies | 23:01 |
| openstackgerrit | Brandon Logan proposed openstack/octavia: Add v2 load balancer type and controllers https://review.openstack.org/423543 | 23:02 |
| rm_work | ah blogan this one too: https://review.openstack.org/#/c/422942/ | 23:03 |
| rm_work | that should be free though because it's fixing your fail :P | 23:03 |
| blogan | rm_work: is it a fail if i didnt know python 3 existed? | 23:04 |
| rm_work | it should fail even on py2, not sure why we don't see that | 23:04 |
| rm_work | i guess py2 somehow figures out how to hash our object? | 23:05 |
| blogan | rm_work: its a fail if you add into that patch another change that renamed load_balancer to loadbalancer | 23:05 |
| rm_work | which doesn't really make sense | 23:05 |
| rm_work | lolol | 23:05 |
| rm_work | consistency | 23:05 |
| *** kevo has quit IRC | 23:05 | |
| rm_work | I'm actually really curious how py2 passes using this | 23:06 |
| rm_work | does py2 like ... id() the LB object or something? >_> | 23:07 |
| blogan | i actually thought we had to do that to make this work | 23:09 |
| blogan | actually i thought i had changed this to just use the id | 23:09 |
| blogan | like | 23:09 |
| blogan | id attribute | 23:09 |
| rm_work | hmmm | 23:09 |
| blogan | oh well | 23:09 |
| rm_work | did that ... never merge? :P | 23:09 |
| rm_work | is there another patch out there that already does this? lol | 23:09 |
| blogan | who knows, it could have been antoher noop driver too | 23:09 |
| blogan | maybe py2 uses the __repr__ method to hash it | 23:10 |
| rm_work | mayhaps | 23:12 |
| * rm_work is listening to josh convince people to use taskflow | 23:12 | |
| * rm_work is being swayed | 23:12 | |
| blogan | rm_work: don't give into it | 23:12 |
| rm_work | if we used jobboard this would be pretty sweet | 23:13 |
| blogan | well johnsom has always planned that | 23:13 |
| rm_work | yeah | 23:13 |
| blogan | and it is a good thing | 23:13 |
| rm_work | it is | 23:13 |
| blogan | my issues with taskflow haven't been with its usefulness | 23:13 |
| rm_work | yeah it's just painful for me to track what's happening, and sometimes it's really counterintuitive what happens with the objects stored, it seems | 23:14 |
| * johnsom Won't comment on that.... | 23:14 | |
| blogan | johnsom will comment on that | 23:14 |
| rm_work | didn't we have to have a task specifically just to mangle the dependency objects and refresh fields? | 23:14 |
| johnsom | blogan I was going to say something like your disposition for a certain coding style..... | 23:14 |
| blogan | johnsom: refresh my memory | 23:15 |
| blogan | workflow engines in general? | 23:15 |
| johnsom | "relaxed" coding style | 23:15 |
| rm_work | lol | 23:15 |
| johnsom | rm_work I think that was more about our data model situation than taskflow | 23:15 |
| rm_work | ah, mayhaps | 23:15 |
| rm_work | might be improperly blaming taskflow | 23:16 |
| johnsom | I.e. getting the data model objects updated from the DB at the right times | 23:16 |
| blogan | i'm very relaxed when i'm coding :) | 23:16 |
| johnsom | It's a polite way of saying "slapping it together" | 23:16 |
| johnsom | grin | 23:16 |
| blogan | thast the best way! | 23:16 |
| rm_work | gotta slap your troubles away | 23:16 |
| blogan | slap-a-da-bass | 23:17 |
| johnsom | So rm_work, now you have Josh's ear, you guys can work together on implementing jobboard for us.... | 23:17 |
| rm_work | maybe :) | 23:17 |
| rm_work | he'll have more time since he's not running for oslo PTL again :P | 23:17 |
| johnsom | Exactly... | 23:17 |
| blogan | that sounds like a smart move on his part | 23:17 |
| blogan | i'd hate to be a poor sucker who runs for ptl | 23:18 |
| blogan | amiright johnsom? | 23:18 |
| *** bana_k has quit IRC | 23:18 | |
| johnsom | Sigh, yeah..... https://review.openstack.org/#/c/423381/ | 23:18 |
| rm_work | :P | 23:18 |
| rm_work | interesting, we do this via gerrit? | 23:19 |
| blogan | -1 | 23:19 |
| rm_work | -1 for "continue to help this incredible team continue" | 23:19 |
| rm_work | nit | 23:19 |
| johnsom | Your nomination is via gerrit, votes are via the same old system... | 23:19 |
| johnsom | All formal and stuff now that we are a grown up project | 23:19 |
| rm_work | and "neturon stadium" | 23:20 |
| johnsom | Too much politicking? | 23:20 |
| rm_work | too little spellchecking :P | 23:20 |
| johnsom | Sigh, I even had someone read it over.... | 23:20 |
| rm_work | your mistake was not having ME read it over ^_^ | 23:21 |
| *** bana_k has joined #openstack-lbaas | 23:21 | |
| rm_work | I'm a proud member of Cosa Nostra Grammatica :P | 23:22 |
| *** _ducttape_ has quit IRC | 23:22 | |
| johnsom | There you go | 23:22 |
| rm_work | https://d3gqasl9vmjfd8.cloudfront.net/7371b292-687b-407f-8cdb-d6167eb020ac.png | 23:22 |
| johnsom | It only had one +2 anyway | 23:22 |
| rm_work | nice | 23:23 |
| rm_work | haha | 23:23 |
| rm_work | yeah surprised you cared enough to fix it | 23:23 |
| johnsom | The reviewer was a journalism major too. Though in all fairness I think she was trying to go to sleep.... | 23:23 |
| * johnsom Thinks it's good his spouse is not on IRC | 23:24 | |
| *** kevo has joined #openstack-lbaas | 23:24 | |
| johnsom | Hey, I had two -1 from constitutes, so I had to fix it.... | 23:25 |
| rm_work | heh | 23:26 |
| johnsom | Ok, back to fixing our sonar CI records from still pointing to non-existent HPE e-mail addresses | 23:26 |
| *** armax has quit IRC | 23:41 | |
| johnsom | rm_work are you going to the PTG? | 23:42 |
| rm_work | johnsom: lol i guess my email is the recovery email for octavia.sonar@gmail | 23:42 |
| rm_work | getting spammed about stuff noe | 23:42 |
| rm_work | *now | 23:42 |
| rm_work | no :( | 23:42 |
| johnsom | Yeah, I just changed it back to my e-mail address. | 23:42 |
| rm_work | kk | 23:42 |
| johnsom | I also updated the third party CI record address to be a forwarder at octavia.cloud instead of an old hp alias | 23:43 |
| johnsom | In case anyone needs to track us down | 23:43 |
| rm_work | cool | 23:44 |
| johnsom | I was just finishing up fixing the third party CI voting stuff after the spin out. we now have octavia-ci which has the third party gates in it for voting like they had under neutron | 23:45 |
| rm_work | how many people are going to the PTG from octavia I wonder? | 23:46 |
| johnsom | I know of 6-7 | 23:46 |
| rm_work | we still have 6-7 people? ^_^ | 23:46 |
| johnsom | Ha, funny | 23:47 |
| blogan | if you count me as a person | 23:47 |
| blogan | im subhuman though | 23:47 |
| johnsom | Well, yes, I did, so 5-6 | 23:47 |
| rm_work | maybe you're the "6th-7th" | 23:47 |
| rm_work | ah | 23:47 |
| rm_work | lol | 23:47 |
| johnsom | I didn't say gentlemen... | 23:47 |
| blogan | alright i've had enough of your insults | 23:47 |
| blogan | im done for the day...i think | 23:47 |
| openstackgerrit | Merged openstack/octavia: Fix connection timeout spam in subunit logs https://review.openstack.org/423041 | 23:47 |
| blogan | i may work on some more of the octavia stuff | 23:48 |
| blogan | WHILE RELAXING | 23:48 |
| *** eezhova has quit IRC | 23:48 | |
| johnsom | So, stackalytics says 50 people have done at least one review on "octavia group" projects | 23:48 |
| johnsom | blogan Enjoy! Thanks again for your work on the API stuff | 23:48 |
| rm_work | how many have commits? :/ | 23:51 |
| rm_work | though I guess reviews help too | 23:51 |
| *** kevo has quit IRC | 23:53 | |
| *** fnaval has quit IRC | 23:55 | |
| johnsom | Patch sets is also 50, but commits is hmm, also 50. Interesting. That includes octavia, the dashboard, and lbaas | 23:57 |
| rm_work | hmm | 23:58 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
