| blogan | grrr zuul is failing for the octavia tests, wish i could find the octavia and neutron logs on it | 00:00 |
|---|---|---|
| *** blogan_ has joined #openstack-lbaas | 00:01 | |
| johnsom | Yeah, it's not done yet | 00:02 |
| johnsom | I am wondering if some of the gate tweaks Min and I did need pulled over | 00:02 |
| blogan_ | its failing on all the setupClass methods, it can't creat an LB for some reason | 00:04 |
| johnsom | https://review.openstack.org/#/c/209675/48/neutron_lbaas/tests/contrib/gate_hook.sh | 00:04 |
| blogan_ | but cna't see why without neutron and octavia logs | 00:04 |
| johnsom | Could be the backing store issue | 00:04 |
| blogan_ | doesn't that pull octavia down? is it not pulling down octavia now? | 00:05 |
| blogan_ | without that review | 00:05 |
| johnsom | We are pulling in octavia, it's the backing store setting I'm wondering about. | 00:06 |
| blogan_ | so without this review, the enable_plugin for octavia is still being set? | 00:08 |
| blogan_ | also looks like neutron-lbaas plugin is being added twice | 00:09 |
| blogan_ | in that review | 00:09 |
| johnsom | Wow, the jobs are done, but it is still sitting on it | 00:15 |
| *** SumitNaiksatam has quit IRC | 00:15 | |
| johnsom | Ok, game on. | 00:18 |
| *** devlaps has quit IRC | 00:18 | |
| *** bradjones has quit IRC | 00:25 | |
| *** minwang2_ has joined #openstack-lbaas | 00:27 | |
| *** xgerman_ has joined #openstack-lbaas | 00:28 | |
| openstackgerrit | Stephen Balukoff proposed openstack/neutron-lbaas: Make pools independent of listeners https://review.openstack.org/218560 | 00:28 |
| *** manishg_ has quit IRC | 00:29 | |
| openstackgerrit | Michael Johnson proposed openstack/neutron-lbaas: Various fixes for tempest tests with octavia https://review.openstack.org/223345 | 00:30 |
| openstackgerrit | Michael Johnson proposed openstack/neutron-lbaas: Set Octavia as the reference LBaaSv2 driver https://review.openstack.org/223265 | 00:30 |
| johnsom | blogan_ Yeah, it looks like that gate code was changed since I had it working. I don't know why there is dup in there. | 00:33 |
| johnsom | I have pulled in the gate settings that I think we need | 00:33 |
| *** madhu_ak has quit IRC | 00:34 | |
| *** minwang2 has quit IRC | 00:35 | |
| *** akamyshnikova has quit IRC | 00:35 | |
| *** xgerman has quit IRC | 00:36 | |
| *** minwang2_ is now known as minwang2 | 00:36 | |
| *** diogogmt has quit IRC | 00:36 | |
| *** xgerman_ is now known as xgerman | 00:37 | |
| *** akamyshnikova has joined #openstack-lbaas | 00:37 | |
| *** ganeshna has joined #openstack-lbaas | 00:39 | |
| openstackgerrit | min wang proposed openstack/octavia: Add cert tracking and rotating in Housekeeping https://review.openstack.org/215359 | 00:39 |
| *** ganeshna has quit IRC | 00:48 | |
| *** vivek-ebay has joined #openstack-lbaas | 01:03 | |
| *** vivek-eb_ has joined #openstack-lbaas | 01:05 | |
| *** vivek-ebay has quit IRC | 01:09 | |
| blogan_ | johnsom: great, thats what i was wondering about, looked like octavia wasn't being setup | 01:13 |
| johnsom | Yeah, it's just a waiting game now | 01:13 |
| bana_k | Octaiva is not at all compatible with the LbaasV1 right ? | 01:18 |
| blogan_ | bana_k: nope, but a driver could be created for it! | 01:22 |
| blogan_ | bana_k: would it get merged? no | 01:22 |
| bana_k | oh ok cool. got it :) thanks | 01:23 |
| bana_k | blogan_ and heat support is going to come in M | 01:23 |
| bana_k | ? | 01:24 |
| sbalukoff | bana_k: If we add it, sure. :) | 01:28 |
| sbalukoff | bana_k: I don't know that we've defined exactly what that looks like yet. | 01:29 |
| johnsom | blogan Things are starting to pass | 01:29 |
| *** bana_k has quit IRC | 01:33 | |
| blogan_ | johnsom: yep see that | 01:34 |
| blogan_ | johnsom: running it locally too just for shits and giggles | 01:34 |
| blogan_ | sbalukoff: yeah not sure about heat support, if someone wants to add it great! | 01:35 |
| blogan_ | just need the someone and motivation to do it | 01:35 |
| *** manishg has joined #openstack-lbaas | 01:36 | |
| blogan_ | johnsom: welp one test just failed locally, stats call | 01:37 |
| blogan_ | johnsom: never failed for me earlier so hopefully its just a fluke | 01:37 |
| *** clev is now known as clev-away | 01:40 | |
| *** sballe has quit IRC | 01:46 | |
| *** manishg has quit IRC | 01:46 | |
| johnsom | So far, so good on the gate | 02:01 |
| *** vivek-eb_ has quit IRC | 02:04 | |
| *** manishg has joined #openstack-lbaas | 02:08 | |
| *** KunalGandhi has quit IRC | 02:10 | |
| *** bharathm has joined #openstack-lbaas | 02:42 | |
| *** bharathm has quit IRC | 02:47 | |
| *** ganeshna has joined #openstack-lbaas | 02:48 | |
| johnsom | Build timed out..... sigh | 02:53 |
| blogan_ | lol | 02:53 |
| blogan_ | i didnt know they had that | 02:53 |
| blogan_ | but now we kno 2 hours and 5 minutes | 02:53 |
| blogan_ | which was concerning anyway | 02:53 |
| johnsom | Yeah, Min and I hit that before too | 02:53 |
| blogan_ | is that a global setting? | 02:54 |
| johnsom | I took 30 minutes just to boot the darn devstack.... | 02:54 |
| blogan_ | hell it might have gone 4 hours | 02:54 |
| blogan_ | still had the rest of the load balancer tests, then members, then pools | 02:54 |
| johnsom | Yeah, I don't know how to fix that. It's Jenkins killing it. We got some time and memory back by taking horizon out of the build, but probably not enough. | 02:55 |
| blogan_ | well i think the real problem is the length of time it takes to boot VMs | 02:56 |
| johnsom | My comment was that we need to restructure the tests | 02:56 |
| blogan_ | well we could just run the smokes | 02:56 |
| blogan_ | but thats not very thorough | 02:56 |
| johnsom | Yeah. If they had vt-x turned on this would not be a problem... | 02:57 |
| blogan_ | how hard is it to get that turned on? impossible? | 02:57 |
| johnsom | I started an internal conversation about that, but it isn't going to happen any time soon | 02:58 |
| blogan_ | dang | 02:58 |
| johnsom | It's kvm and kernel settings | 02:58 |
| blogan_ | welp 3 option i see, 1) do smokes only 2) make a namespace/process based compute driver in octavia 3) dont make it ref | 02:59 |
| blogan_ | all are shitty options | 02:59 |
| blogan_ | though im sure there are others im not thinking | 02:59 |
| johnsom | Yeah, containers would change this too | 03:00 |
| blogan_ | yeah, but we'll need to get phil's code in for that | 03:00 |
| blogan_ | since hot plugging does not work | 03:00 |
| johnsom | So I'm not sure the process/namespace is worth the effort since we are going down the container path anyway | 03:00 |
| blogan_ | which i was planning on doing some testing on soon anyway | 03:00 |
| blogan_ | yeah i agree, and its not something any of us will use so its kinda pointless other than to give us a faux "good" environment | 03:01 |
| johnsom | Well, I think we should think about this and pick it up tomorrow. (at least that is what my wife is thinking in the other room) | 03:01 |
| blogan_ | lol mine too | 03:01 |
| johnsom | Ok, catch you tomorrow | 03:01 |
| blogan_ | see ya man | 03:01 |
| blogan_ | take care | 03:01 |
| *** blogan_ has quit IRC | 03:02 | |
| *** ganeshna has quit IRC | 03:05 | |
| openstackgerrit | Stephen Balukoff proposed openstack/octavia: Move roadmap to wiki https://review.openstack.org/208763 | 03:21 |
| *** vivek-ebay has joined #openstack-lbaas | 03:28 | |
| *** manishg has quit IRC | 03:28 | |
| *** manishg has joined #openstack-lbaas | 03:29 | |
| *** SumitNaiksatam has joined #openstack-lbaas | 03:31 | |
| *** vivek-eb_ has joined #openstack-lbaas | 03:31 | |
| *** vivek-ebay has quit IRC | 03:32 | |
| *** bana_k has joined #openstack-lbaas | 03:33 | |
| *** manishg has quit IRC | 03:52 | |
| *** amotoki has joined #openstack-lbaas | 03:56 | |
| *** minwang2 has quit IRC | 03:58 | |
| *** macjack has joined #openstack-lbaas | 03:58 | |
| *** Guest79093 has quit IRC | 04:02 | |
| *** Brian has joined #openstack-lbaas | 04:03 | |
| *** Brian is now known as Guest34756 | 04:04 | |
| *** ganeshna has joined #openstack-lbaas | 04:04 | |
| *** vjay14 has joined #openstack-lbaas | 04:06 | |
| *** Guest34756 has quit IRC | 04:09 | |
| *** Brian has joined #openstack-lbaas | 04:58 | |
| *** Brian is now known as Guest23087 | 04:58 | |
| *** amit213 has quit IRC | 04:59 | |
| *** amit213 has joined #openstack-lbaas | 04:59 | |
| *** bana_k has quit IRC | 05:06 | |
| openstackgerrit | Michael Johnson proposed openstack/octavia: Don't build amphora image if one is available https://review.openstack.org/223922 | 05:15 |
| openstackgerrit | Michael Johnson proposed openstack/neutron-lbaas: Set Octavia as the reference LBaaSv2 driver https://review.openstack.org/223265 | 05:22 |
| johnsom | This is an experiment to see if using a pre-built amphora image speeds up the gate enough to finish. | 05:28 |
| sbalukoff | I was wondering whether that's what that was about. | 05:30 |
| sbalukoff | Seems to make sense to me. | 05:30 |
| *** johnsom_ has joined #openstack-lbaas | 05:31 | |
| *** sbalukoff has quit IRC | 05:38 | |
| *** sbalukoff has joined #openstack-lbaas | 05:40 | |
| *** KunalGandhi has joined #openstack-lbaas | 05:46 | |
| *** KunalGan_ has joined #openstack-lbaas | 05:54 | |
| *** KunalGandhi has quit IRC | 05:55 | |
| *** vjay14 has quit IRC | 06:01 | |
| *** ganeshna has quit IRC | 06:14 | |
| *** numans has joined #openstack-lbaas | 06:15 | |
| *** Alex_Stef has joined #openstack-lbaas | 06:21 | |
| *** ganeshna has joined #openstack-lbaas | 06:23 | |
| openstackgerrit | Michael Johnson proposed openstack/octavia: Don't build amphora image if one is available https://review.openstack.org/223922 | 06:24 |
| *** bharathm has joined #openstack-lbaas | 06:34 | |
| openstackgerrit | ZhaoBo proposed openstack/neutron-lbaas: neutron lbaas should add the new appended timestamp field https://review.openstack.org/220804 | 06:37 |
| *** johnsom_ has quit IRC | 06:42 | |
| *** KunalGan_ has quit IRC | 06:46 | |
| *** vivek-ebay has joined #openstack-lbaas | 07:00 | |
| *** ganeshna has quit IRC | 07:03 | |
| *** vivek-eb_ has quit IRC | 07:04 | |
| *** ganeshna has joined #openstack-lbaas | 07:07 | |
| *** SumitNaiksatam has quit IRC | 07:11 | |
| *** SumitNaiksatam has joined #openstack-lbaas | 07:12 | |
| *** Guest23087 has quit IRC | 07:19 | |
| *** Guest23087 has joined #openstack-lbaas | 07:24 | |
| *** evgenyf has joined #openstack-lbaas | 07:31 | |
| *** [1]evgenyf has joined #openstack-lbaas | 07:35 | |
| *** evgenyf has quit IRC | 07:38 | |
| *** [1]evgenyf is now known as evgenyf | 07:38 | |
| *** vivek-ebay has quit IRC | 07:41 | |
| *** ganeshna has quit IRC | 07:55 | |
| *** nmagnezi has joined #openstack-lbaas | 08:00 | |
| *** nmagnezi has quit IRC | 08:08 | |
| *** kiran-r has joined #openstack-lbaas | 08:09 | |
| *** nmagnezi has joined #openstack-lbaas | 08:10 | |
| openstackgerrit | Evgeny Fedoruk proposed openstack/neutron-lbaas: L7 support for Radware LBaaS v2 driver https://review.openstack.org/202147 | 08:10 |
| openstackgerrit | Evgeny Fedoruk proposed openstack/neutron-lbaas: L7 capability extension implementation for lbaas v2 https://review.openstack.org/148232 | 08:14 |
| *** ganeshna has joined #openstack-lbaas | 08:22 | |
| openstackgerrit | Merged openstack/neutron-lbaas: Add device_id when a loadbalancer is created instead of at listener create https://review.openstack.org/221985 | 08:27 |
| *** kbyrne has joined #openstack-lbaas | 08:53 | |
| *** bharathm has quit IRC | 09:50 | |
| *** ganeshna has quit IRC | 10:29 | |
| *** Divya has quit IRC | 10:37 | |
| *** Guest23087 has quit IRC | 10:48 | |
| *** Brian has joined #openstack-lbaas | 10:48 | |
| *** Brian is now known as Guest35770 | 10:48 | |
| *** vjay14 has joined #openstack-lbaas | 11:02 | |
| *** itsuugo has joined #openstack-lbaas | 11:04 | |
| *** eranra has joined #openstack-lbaas | 11:06 | |
| *** itsuugo has quit IRC | 11:11 | |
| *** eranra_ has joined #openstack-lbaas | 11:11 | |
| *** eranra_ has quit IRC | 11:13 | |
| *** eranra has quit IRC | 11:13 | |
| *** eranra has joined #openstack-lbaas | 11:14 | |
| *** eranra has quit IRC | 11:16 | |
| *** Guest35770 has quit IRC | 11:19 | |
| *** itsuugo has joined #openstack-lbaas | 11:19 | |
| *** itsuugo has quit IRC | 11:26 | |
| *** vjay14 has quit IRC | 11:34 | |
| *** Brian has joined #openstack-lbaas | 11:44 | |
| *** Brian is now known as Guest52648 | 11:44 | |
| *** ganeshna has joined #openstack-lbaas | 11:58 | |
| *** ganeshna has quit IRC | 11:58 | |
| *** ganeshna has joined #openstack-lbaas | 11:59 | |
| *** evgenyf has quit IRC | 12:04 | |
| *** itsuugo has joined #openstack-lbaas | 12:05 | |
| *** vjay14 has joined #openstack-lbaas | 12:06 | |
| *** evgenyf has joined #openstack-lbaas | 12:09 | |
| *** vjay14 has quit IRC | 12:10 | |
| *** vjay14 has joined #openstack-lbaas | 12:17 | |
| *** ganeshna has quit IRC | 12:19 | |
| *** ganeshna has joined #openstack-lbaas | 12:19 | |
| *** itsuugo has quit IRC | 12:22 | |
| *** ganeshna has quit IRC | 12:23 | |
| *** vjay14 has quit IRC | 12:23 | |
| *** ganeshna has joined #openstack-lbaas | 12:24 | |
| *** macjack has quit IRC | 12:25 | |
| *** kiran-r has quit IRC | 12:25 | |
| *** vjay14 has joined #openstack-lbaas | 12:39 | |
| *** ganeshna has quit IRC | 12:40 | |
| *** kiran-r has joined #openstack-lbaas | 12:44 | |
| *** sbalukoff has quit IRC | 12:48 | |
| *** sbalukoff has joined #openstack-lbaas | 12:50 | |
| *** bharathm has joined #openstack-lbaas | 12:51 | |
| *** bharathm has quit IRC | 12:55 | |
| *** ganeshna has joined #openstack-lbaas | 13:03 | |
| *** Alex_Stef has quit IRC | 13:16 | |
| *** itsuugo has joined #openstack-lbaas | 13:17 | |
| *** ganeshna has quit IRC | 13:24 | |
| openstackgerrit | Ihar Hrachyshka proposed openstack/neutron-lbaas: tox.ini: check migrations for the subproject only https://review.openstack.org/224083 | 13:30 |
| *** kiran-r has quit IRC | 13:32 | |
| *** numans has quit IRC | 13:35 | |
| *** itsuugo has quit IRC | 13:39 | |
| *** amotoki has quit IRC | 13:42 | |
| *** eranra has joined #openstack-lbaas | 13:55 | |
| *** eranra has quit IRC | 13:56 | |
| *** minwang2 has joined #openstack-lbaas | 14:16 | |
| *** vjay14 has quit IRC | 14:21 | |
| *** openstackgerrit has quit IRC | 14:31 | |
| *** openstackgerrit has joined #openstack-lbaas | 14:31 | |
| *** vjay14 has joined #openstack-lbaas | 14:40 | |
| *** minwang2 has quit IRC | 14:40 | |
| *** minwang2 has joined #openstack-lbaas | 14:41 | |
| *** diogogmt has joined #openstack-lbaas | 14:42 | |
| *** ganeshna has joined #openstack-lbaas | 14:50 | |
| *** nmagnezi has quit IRC | 14:56 | |
| openstackgerrit | Phillip Toohill proposed openstack/octavia: Updates for containers functionality https://review.openstack.org/199954 | 14:56 |
| *** openstackgerrit has quit IRC | 15:01 | |
| *** openstackgerrit has joined #openstack-lbaas | 15:02 | |
| *** vivek-ebay has joined #openstack-lbaas | 15:04 | |
| *** ganeshna has quit IRC | 15:04 | |
| *** ganeshna has joined #openstack-lbaas | 15:07 | |
| *** manishg has joined #openstack-lbaas | 15:09 | |
| *** vivek-ebay has quit IRC | 15:09 | |
| *** manishg_ has joined #openstack-lbaas | 15:10 | |
| *** KunalGandhi has joined #openstack-lbaas | 15:13 | |
| *** manishg has quit IRC | 15:13 | |
| *** ajmiller has joined #openstack-lbaas | 15:21 | |
| *** KunalGandhi has quit IRC | 15:22 | |
| *** KunalGandhi has joined #openstack-lbaas | 15:30 | |
| *** bharathm has joined #openstack-lbaas | 15:35 | |
| *** TrevorV has joined #openstack-lbaas | 15:37 | |
| *** vjay15 has joined #openstack-lbaas | 15:37 | |
| *** vjay14 has quit IRC | 15:38 | |
| *** minwang2 has quit IRC | 15:39 | |
| openstackgerrit | Michael Johnson proposed openstack/octavia: Don't build amphora image if one is available https://review.openstack.org/223922 | 15:42 |
| rm_work | blogan / johnsom: ok, so the dsvm is running now, just timing out because of long builds? still not sure how it is taking so long given it took 4 minutes average on the RAX images i was testing | 15:45 |
| rm_work | blogan: johnsom: but do you have a link to some logs from one of the runs? | 15:45 |
| johnsom | rm_work This was the last run: http://logs.openstack.org/65/223265/8/check/gate-neutron-lbaasv2-dsvm-api/cc42f10/ | 15:46 |
| johnsom | I'm trying to have it download the amp image instead of doing the build. That should save a bunch of setup time. | 15:46 |
| *** minwang2 has joined #openstack-lbaas | 15:46 | |
| rm_work | yeah but ... only once | 15:47 |
| johnsom | Unfortunately my trick to inject it into the localrc during the gate doesn't seem to work, so I modified the devstack plugin.sh script to download it | 15:47 |
| johnsom | That is what is running now. I have it downloading the image from our CDN. My hope is this will drop the devstack setup time from 30 minutes and give the test enough time to run | 15:50 |
| *** devlaps has joined #openstack-lbaas | 15:54 | |
| rm_work | hmm k | 15:54 |
| johnsom | Here is a run from yesterday that was succeeding, but timed out: http://logs.openstack.org/65/223265/7/check/gate-neutron-lbaasv2-dsvm-api/897d0b6/console.html | 15:55 |
| *** minwang2 has quit IRC | 15:59 | |
| rm_work | thx | 16:00 |
| rm_work | johnsom: in that log, I see VM builds taking 4-7 minutes | 16:01 |
| *** evgenyf has quit IRC | 16:01 | |
| rm_work | that is much more like what i expected | 16:01 |
| johnsom | Devstack took 27 minutes to setup | 16:02 |
| rm_work | ah, yeah | 16:02 |
| rm_work | i thought you were saying the VMs were taking 30m to build | 16:03 |
| johnsom | No, the devstack setup | 16:03 |
| johnsom | That is what I'm trying to reduce | 16:03 |
| rm_work | hmm and that's mostly the amp image? got it | 16:03 |
| johnsom | Well, the amp image build can take some time, so giving it a shot | 16:03 |
| blogan | i dont think saving 30 minutes will help, i still think the job will still take 3 hours at a minimum | 16:05 |
| johnsom | Yeah, could be | 16:05 |
| *** sballe has joined #openstack-lbaas | 16:06 | |
| blogan | 4-7 minute build times for each load balancer is the killer, whereas the namespace driver just spun up a process and booted in like 3 seconds | 16:06 |
| johnsom | Yeah, it's less than a minute if you have the vt-x available | 16:07 |
| *** minwang2 has joined #openstack-lbaas | 16:07 | |
| blogan | that would cut it down by a lot for sure | 16:07 |
| johnsom | It's the VM inside a VM problem with qemu dropping to tcg | 16:07 |
| blogan | but you said it wouldn't be soon before they fixed that | 16:08 |
| *** mlavalle has joined #openstack-lbaas | 16:09 | |
| *** vivek-ebay has joined #openstack-lbaas | 16:09 | |
| *** vivek-ebay has quit IRC | 16:09 | |
| *** vivek-ebay has joined #openstack-lbaas | 16:10 | |
| johnsom | Yeah, I don't think so. We would need both your systems and ours to enable it | 16:10 |
| blogan | is that a decision that can be made by infra or is that like rackspace and hp enabling it in their public cloud? | 16:11 |
| blogan | i dont know how all that works | 16:11 |
| johnsom | I don't really either, but my guess is that our clouds would have to enable it | 16:11 |
| johnsom | http://kashyapc.com/2012/01/14/nested-virtualization-with-kvm-intel/ | 16:11 |
| johnsom | There is also this: http://docs.openstack.org/developer/devstack/guides/devstack-with-nested-kvm.html | 16:12 |
| *** vivek-ebay has quit IRC | 16:13 | |
| *** minwang2 has quit IRC | 16:17 | |
| rm_work | johnsom: ok, so i know we use quite a bit of disk now for some reason? | 16:29 |
| rm_work | got the advice that it is actually quite I/O bound more than anything, and reducing disk size of flavors helps a lot | 16:30 |
| johnsom | Sorry, can't quite parse that | 16:30 |
| rm_work | and IIRC we just increased the disk size a bit? | 16:30 |
| *** ganeshna has quit IRC | 16:30 | |
| *** ganeshna has joined #openstack-lbaas | 16:31 | |
| johnsom | I didn't increase the disk size, I told it that we could have up to 24G available to nova for the amps | 16:31 |
| rm_work | oh, k | 16:32 |
| rm_work | can we REDUCE the disk size? | 16:32 |
| rm_work | how much does it really need? | 16:32 |
| johnsom | I set VOLUME_BACKING_SIZE=24G | 16:32 |
| rm_work | k | 16:33 |
| johnsom | We give the root disk, via flavor, 5GB. I think we really only use 2GB of that. | 16:33 |
| johnsom | I don't really understand how that is going to make any difference though | 16:34 |
| johnsom | It's 1.6GB of used space actually | 16:35 |
| johnsom | If we want to drop that from 5 to 2 we can try it. | 16:35 |
| johnsom | I just want to see how we do on timing with this test | 16:35 |
| rm_work | yeah we should try that | 16:38 |
| rm_work | they said it's the I/O of creating the disk images and such is a lot | 16:39 |
| johnsom | Let me double check 2GB boots on my devstack while we wait for this | 16:39 |
| rm_work | k | 16:40 |
| rm_work | johnsom: do you see a flavor "42" in devstack? | 16:43 |
| rm_work | johnsom: and if so, what are the specs on it | 16:43 |
| johnsom | | 42 | m1.nano | 64 | 0 | 0 | | 1 | 1.0 | True | | 16:43 |
| rm_work | err so that's | 16:44 |
| rm_work | 64mb memory??? | 16:44 |
| rm_work | 1g disk? | 16:44 |
| rm_work | not sure of the solumns | 16:45 |
| rm_work | *columns | 16:45 |
| johnsom | 64MB RAM, 0GB disk | 16:45 |
| *** bharathm has quit IRC | 16:46 | |
| rm_work | lol | 16:46 |
| rm_work | well, maybe 5->2GB will help | 16:46 |
| johnsom | yeah, no idea | 16:46 |
| *** bharathm has joined #openstack-lbaas | 16:46 | |
| *** KunalGan_ has joined #openstack-lbaas | 16:50 | |
| *** KunalGan_ has quit IRC | 16:50 | |
| *** ganeshna has quit IRC | 16:52 | |
| *** KunalGandhi has quit IRC | 16:54 | |
| *** crc32 has joined #openstack-lbaas | 16:55 | |
| *** diogogmt_ has joined #openstack-lbaas | 16:56 | |
| *** diogogmt has quit IRC | 16:56 | |
| *** diogogmt_ is now known as diogogmt | 16:56 | |
| rm_work | johnsom: i also wonder if either increasing or reducing the RAM would help | 16:56 |
| *** minwang2 has joined #openstack-lbaas | 16:56 | |
| rm_work | depends on what is blocking i guess | 16:56 |
| johnsom | I had to reduce it with the custom flavor as it said there wasn't enough RAM available in the gate host to boot the amp | 16:57 |
| dougwig | would anyone be available to attempt to convert the lbaas tempest tests to this framework? http://docs.openstack.org/developer/tempest/plugin.html (xgerman?) | 16:57 |
| xgerman | in M? | 16:57 |
| xgerman | and we don’t want to go rally? | 16:57 |
| dougwig | tempest folks are at a code sprint, working on that interface right now, so it'd be interesting if we could give it a go and give them feedback (similar to what we did with the devstack plugin.) | 16:58 |
| johnsom | rm_work Ok, I built a 2GB image and booted it in a smaller flavor. After this test I will re-work the patches and give it a go. | 16:59 |
| johnsom | rm_work just don't understand this issue, since I can reproduce these speed issues with vt-x on and off | 16:59 |
| *** sbalukoff has quit IRC | 17:00 | |
| johnsom | rm_work It's a check box on my VMWare host. | 17:01 |
| rm_work | ah yeah, one would assume that if it doesn't slow the build with vt-x on, then... yeah | 17:01 |
| rm_work | hmm | 17:01 |
| johnsom | I've give it go though. I just want to see the speedup with the download on this run. | 17:02 |
| blogan | i assume trove has the same issues but they just run less tests? | 17:02 |
| rm_work | yeah i think their advice was like... do less tests | 17:02 |
| rm_work | lol | 17:02 |
| rm_work | or "use nova-lxc" | 17:02 |
| blogan | yeah | 17:02 |
| johnsom | Yeah, the trove and cue guys just boot one or two vms for all of the tests | 17:02 |
| blogan | which i think is going to end up being our best option | 17:02 |
| johnsom | We talked to them about it | 17:02 |
| openstackgerrit | Carlos Garza proposed openstack/octavia: Implementing EventStreamer https://review.openstack.org/218735 | 17:02 |
| rm_work | but would nova-lxc work with our existing nova/neutron compute/network drivers? | 17:03 |
| blogan | well we'd ahve to restructure the tests for that | 17:03 |
| johnsom | Yeah, I agree. Containers are probably the answer | 17:03 |
| blogan | rm_work: no bc im sure they dont support hot plugging, but if we got phils patches reviewed and merged they m ight work with that | 17:03 |
| johnsom | Wow, download went from 30 minutes for devstack build to ~4 | 17:03 |
| rm_work | but then, that also wouldn't test our default compute option <_< | 17:03 |
| blogan | then again i dont know how nova-lxc works, is that magnum or something nova extension? | 17:04 |
| blogan | well if we can default to containers i dont think anyone woudl complain | 17:04 |
| blogan | well i wouldnt | 17:04 |
| blogan | :) | 17:04 |
| johnsom | Oh, nevermind, devstack build failed | 17:04 |
| rm_work | heh | 17:04 |
| rm_work | but isn't HP planning on full VMs? | 17:04 |
| blogan | johnsom: i tried building the amphora image on a 2GB disk size flavor and it failed for me | 17:04 |
| rm_work | so we'd need to make sure that code path works <_< | 17:04 |
| blogan | i've heard whispers of using containers | 17:05 |
| crc32 | <psst>containers</psst> | 17:05 |
| johnsom | Did you rebuild the image down to a 2GB size? You have to do that | 17:05 |
| blogan | johnsom: ah probably not | 17:05 |
| johnsom | rm_work blogan Do you guys know how to get the devstacklog.txt for that last dsvm run? | 17:08 |
| blogan | i do not | 17:08 |
| johnsom | The console log says they were kept, but I can't find them via the link | 17:08 |
| rm_work | hmm | 17:14 |
| rm_work | johnsom: did the whole jenkins run finish? | 17:14 |
| rm_work | i am never able to find the log links until the entire zuul job finishes and jenkins votes | 17:14 |
| johnsom | It failed on devstack setup. | 17:14 |
| rm_work | then the vote includes the log links | 17:14 |
| *** madhu_ak has joined #openstack-lbaas | 17:18 | |
| *** SumitNaiksatam has quit IRC | 17:27 | |
| *** diogogmt has quit IRC | 17:34 | |
| xgerman | blogan/dougwig: https://wiki.openstack.org/wiki/Octavia/Weekly_Meeting_Agenda#Meeting_2015-09-16 | 17:39 |
| openstackgerrit | Michael Johnson proposed openstack/octavia: Don't build amphora image if one is available https://review.openstack.org/223922 | 17:42 |
| johnsom | rm_work Ok, running a 2GB image and flavor | 17:46 |
| rm_work | k, it might help SOME at least... maybe a minimal amount :/ | 17:47 |
| rm_work | but yeah given your earlier results i wouldn't imagine it's a gamechanger | 17:48 |
| *** diogogmt has joined #openstack-lbaas | 17:55 | |
| *** bana_k has joined #openstack-lbaas | 17:59 | |
| rm_work | johnsom: yeah watching a local devstack build and it really does take a while for that image builder <_< | 18:02 |
| rm_work | I will grab your patch and see how my local builds improve | 18:02 |
| johnsom | Yeah, it does | 18:02 |
| *** sbalukoff has joined #openstack-lbaas | 18:09 | |
| rm_work | so the only issue i see with this | 18:10 |
| rm_work | is that we then have issues testing changes that touch our image builder | 18:10 |
| rm_work | because we make a change... and it flies through the gate without actually testing it | 18:10 |
| rm_work | if we can solve that, then i'd vote to merge (something like) this | 18:11 |
| johnsom | Yeah, agree, I was exploring this as an experiment | 18:11 |
| rm_work | yeah | 18:11 |
| rm_work | should be solvable | 18:12 |
| rm_work | just have to think a bit | 18:12 |
| *** diogogmt has quit IRC | 18:16 | |
| *** openstackgerrit has quit IRC | 18:16 | |
| *** KunalGandhi has joined #openstack-lbaas | 18:16 | |
| *** openstackgerrit has joined #openstack-lbaas | 18:16 | |
| *** diogogmt has joined #openstack-lbaas | 18:17 | |
| *** numans has joined #openstack-lbaas | 18:22 | |
| openstackgerrit | Phillip Toohill proposed openstack/octavia: Updates for containers functionality https://review.openstack.org/199954 | 18:33 |
| johnsom | Hmmm, wondering if that last run picked up the right image | 18:34 |
| *** mlavalle has quit IRC | 18:34 | |
| rm_work | johnsom: stack.sh run in RAX public cloud went from 1918 to 1200 seconds | 18:35 |
| rm_work | in my one test | 18:35 |
| johnsom | Cool. I'm just wondering if it got the 2gb image or if the timing was bad and it downloaded the 5gb image | 18:35 |
| rm_work | also yeah, 5gb -> 2gb would reduce download time for the image, unless it is sparse? | 18:35 |
| rm_work | is it sparse? | 18:35 |
| rm_work | i actually hope it's sparse | 18:36 |
| crc32 | rm_work: the isinstance method on __eq__ is actually the pythonic way of testing for equality. Its the only safe way to make quality work in a SOLID pattern. | 18:36 |
| rm_work | hmm | 18:36 |
| crc32 | yea I originally didn't have it. | 18:36 |
| johnsom | Yeah, it's sparse so it really didn't change the image size at all. | 18:36 |
| rm_work | i have mostly heard differently, but maybe it is different for __eq__ | 18:36 |
| rm_work | might be an exception | 18:36 |
| johnsom | Just trying to address your request of a 2GB image and flavor vs. the default 5gb | 18:36 |
| rm_work | i will look around a bit | 18:36 |
| rm_work | yeah | 18:36 |
| rm_work | we'll see what happens to build times | 18:37 |
| rm_work | they MIGHT speed up some | 18:37 |
| rm_work | because it is true that it has to allocate the space | 18:37 |
| rm_work | might be like 20s each or so tho :/ | 18:37 |
| crc32 | If I don't use it and instead just call "==" directky for some reason differen't sub classes will look identicle even when they have different data. | 18:37 |
| rm_work | yeah, you def need to include __eq__ | 18:39 |
| rm_work | ah, i think i see | 18:39 |
| rm_work | yeah, so i guess without isinstance() in your __eq__ method, it's more like javascript's shitty == | 18:40 |
| rm_work | and WITH isinstance() it's more like === | 18:40 |
| rm_work | so yeah, i think you changed my mind, i wasn't really considering exactly where it is | 18:40 |
| rm_work | __eq__ seems like a good exception for the "avoid using isinstance()" rule I follow | 18:41 |
| crc32 | I don't understand it either but for some reasons subclasses with differen't values end up passing the equality test with out isinstance. I'm looking for the example | 18:44 |
| *** manishg_ has quit IRC | 19:00 | |
| *** bharathm has quit IRC | 19:00 | |
| *** manishg has joined #openstack-lbaas | 19:01 | |
| *** manishg has quit IRC | 19:01 | |
| *** manishg has joined #openstack-lbaas | 19:02 | |
| *** bana_k has quit IRC | 19:15 | |
| *** mlavalle has joined #openstack-lbaas | 19:38 | |
| *** bharathm has joined #openstack-lbaas | 19:44 | |
| *** blallau has joined #openstack-lbaas | 19:45 | |
| *** manishg has quit IRC | 19:46 | |
| *** vjay15 has quit IRC | 19:47 | |
| *** bharathm has quit IRC | 19:49 | |
| *** manishg_ has joined #openstack-lbaas | 19:50 | |
| *** vjay15 has joined #openstack-lbaas | 19:50 | |
| *** bana_k has joined #openstack-lbaas | 19:59 | |
| *** mestery has quit IRC | 20:02 | |
| *** mestery has joined #openstack-lbaas | 20:03 | |
| *** bharathm has joined #openstack-lbaas | 20:07 | |
| *** Aish has joined #openstack-lbaas | 20:07 | |
| openstackgerrit | min wang proposed openstack/octavia: Add cert tracking and rotating in Housekeeping https://review.openstack.org/215359 | 20:30 |
| *** ajmiller has quit IRC | 20:39 | |
| *** ajmiller has joined #openstack-lbaas | 20:40 | |
| *** blallau has quit IRC | 20:58 | |
| rm_work | dougwig: it looks like ALL jobs for neutron-lbaas pipe through a single template | 21:00 |
| rm_work | dougwig: do I need to add {subbranch-designator}? lol | 21:00 |
| rm_work | we run: gate-neutron-lbaasv2-octavia-dsvm-api | 21:01 |
| rm_work | err this doesn't make sense | 21:02 |
| rm_work | '{pipeline}-neutron-{lbaasversion}-dsvm-{lbaastest}{branch-designator}' | 21:02 |
| rm_work | gate-neutron-lbaasv2-octavia-dsvm-api | 21:02 |
| rm_work | that doesn't line up | 21:02 |
| rm_work | am I looking at the wrong thing? | 21:02 |
| johnsom | Yes, that is the experimental gate. | 21:03 |
| johnsom | Ignore that for now | 21:03 |
| dougwig | lbaasv2-octavia is the lbaasversion. | 21:03 |
| rm_work | rofl ok | 21:03 |
| dougwig | that field is a misnomer now. | 21:03 |
| dougwig | but yeah, we need to stretch that template a bit more. | 21:04 |
| rm_work | and so "api" is the {lbaastest} | 21:04 |
| dougwig | right | 21:04 |
| rm_work | and there is no {branch-designator}? | 21:04 |
| dougwig | that's a standard thing, we don't touch that | 21:04 |
| rm_work | k | 21:04 |
| rm_work | the value of the would be like | 21:04 |
| rm_work | "-liberty"? | 21:04 |
| rm_work | the parsing on this is really wonky | 21:05 |
| rm_work | this doesn't look like a sane regex, for example | 21:05 |
| rm_work | (ignoring the obvious "sane regex?" joke) | 21:06 |
| *** nmagnezi has joined #openstack-lbaas | 21:06 | |
| *** Alex_Stef has joined #openstack-lbaas | 21:07 | |
| dougwig | that's like sane perl. my brain won't even parse the words. | 21:09 |
| rm_work | how does this tell what {branch-designator} is? | 21:10 |
| rm_work | there's no split | 21:10 |
| rm_work | it looks to me like .*.* | 21:10 |
| *** nmagnezi has quit IRC | 21:10 | |
| dougwig | rm_work: you're going to hurt your brain, insisting on looking into every dark corner of that stuff. | 21:13 |
| rm_work | ugh | 21:13 |
| rm_work | i'll just throw random shit at this wall and see if it sticks | 21:13 |
| *** nmagnezi has joined #openstack-lbaas | 21:15 | |
| *** TrevorV has quit IRC | 21:17 | |
| openstackgerrit | Merged openstack/octavia: Move roadmap to wiki https://review.openstack.org/208763 | 21:18 |
| *** Aish_ has joined #openstack-lbaas | 21:23 | |
| *** Aish_ has quit IRC | 21:24 | |
| sbalukoff | Yay! | 21:24 |
| *** Aish has quit IRC | 21:25 | |
| rm_work | can someone who is better at BASH explain what this does: | 21:28 |
| rm_work | testenv=${2:-"apiv2"} | 21:28 |
| rm_work | is that like .... $2 if it exists or else "apiv2" ? | 21:28 |
| *** bana_k has quit IRC | 21:29 | |
| madhu_ak | rm_work yes. even if there is no $2, it sets $2 apiv2 as default | 21:29 |
| rm_work | kk | 21:29 |
| *** bana_k has joined #openstack-lbaas | 21:35 | |
| *** nmagnezi has quit IRC | 22:06 | |
| *** vjay15 has quit IRC | 22:14 | |
| *** mlavalle has quit IRC | 22:52 | |
| *** crc32 has quit IRC | 23:11 | |
| openstackgerrit | Adam Harwell proposed openstack/neutron-lbaas: Split up gate tests for LBaaSv2 https://review.openstack.org/224345 | 23:21 |
| rm_work | erm, I THINK that should do it | 23:21 |
| rm_work | johnsom: ^^ | 23:22 |
| rm_work | dougwig: ^^ | 23:22 |
| johnsom | Cool, I will take a look | 23:22 |
| rm_work | I added some comments | 23:26 |
| rm_work | dougwig: hoping that is what you had in mind | 23:31 |
| bana_k | johnsom: So for active stand we need to use rest driver and not the ssh driver ? | 23:36 |
| bana_k | standby* | 23:37 |
| johnsom | Correct. We are going to obsolete the ssh driver | 23:37 |
| bana_k | ok | 23:37 |
| johnsom | So the developer didn't implement ssh | 23:37 |
| bana_k | oh ok thought so :) | 23:37 |
| bana_k | thanks | 23:37 |
| *** chlong has quit IRC | 23:41 | |
| rm_work | personally I still like the SSH driver <_< | 23:45 |
| rm_work | but | 23:45 |
| rm_work | yeah | 23:45 |
| rm_work | johnsom: https://jenkins04.openstack.org/job/gate-neutron-lbaasv2-dsvm-api/1254/console | 23:46 |
| rm_work | appears to have worked | 23:46 |
| rm_work | I set the default to only do "load_balancers" tests | 23:46 |
| rm_work | johnsom: try depending on this for your change to update to octavia ref? | 23:46 |
| rm_work | johnsom: and it should be able to pass gate | 23:46 |
| * johnsom gives rm_work a thumbs up | 23:46 | |
| rm_work | we might want to pick a nice subset of tests later to be the "default" | 23:47 |
| johnsom | Ok, let me look at that. I feel like I should undo some of my hacks too | 23:47 |
| johnsom | Or at least make them better hacks | 23:47 |
| bana_k | johnsom : Error while creating the LB. https://gist.github.com/banveerad/90611a894b7775e86b89 | 23:50 |
| johnsom | Hmmm, I wonder if there is a missing dependency | 23:51 |
| johnsom | bana_k The first one is because the amphora image was built with the MASTER code and not the code in this patch. | 23:57 |
| johnsom | The second one uncovered an old bug | 23:57 |
| johnsom | Cool! | 23:57 |
| bana_k | johnsom : Oh I had applied the patch | 23:58 |
| rm_you | woo old bugs | 23:58 |
| bana_k | n I can see the new config file for the octavia | 23:58 |
| johnsom | Yeah, a revert in code in master is updating status instead of provisioning/operating status | 23:58 |
| bana_k | n in octavia db load_balancer does not have the status column | 23:59 |
| bana_k | oh ok so I will take the new code from master | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!