| *** madhu_ak has quit IRC | 00:14 | |
| bana_k | hi, What exactly needs to be done for tls support in the driver ? | 00:18 |
|---|---|---|
| bana_k | was reading up the https://review.openstack.org/#/c/130659/6/specs/version0.5/tls-data-security.rst | 00:18 |
| bana_k | may be the wrong dock to read? | 00:18 |
| *** bharath has quit IRC | 00:37 | |
| johnsom | FYI, I have tested Min's health manager code with the fail over flow. It built the amphora, but there are still some bugs. | 00:46 |
| johnsom | Still, cool to have a fail over trigger! | 00:46 |
| *** crc32 has joined #openstack-lbaas | 00:49 | |
| *** crc32 has quit IRC | 00:52 | |
| xgerman | nice | 00:59 |
| *** bana_k has quit IRC | 01:01 | |
| *** crc32 has joined #openstack-lbaas | 01:02 | |
| *** chlong has joined #openstack-lbaas | 01:03 | |
| *** fnaval has quit IRC | 01:24 | |
| *** fnaval has joined #openstack-lbaas | 01:37 | |
| *** vivek-ebay has quit IRC | 01:39 | |
| *** madhu_ak has joined #openstack-lbaas | 01:46 | |
| *** madhu_ak has quit IRC | 01:51 | |
| *** madhu_ak has joined #openstack-lbaas | 01:53 | |
| *** diogogmt has joined #openstack-lbaas | 01:54 | |
| *** vivek-ebay has joined #openstack-lbaas | 02:06 | |
| *** woodster_ has quit IRC | 02:10 | |
| *** vivek-ebay has quit IRC | 02:55 | |
| *** vivek-ebay has joined #openstack-lbaas | 02:56 | |
| *** KunalGandhi has quit IRC | 03:00 | |
| *** madhu_ak has quit IRC | 03:05 | |
| *** madhu_ak has joined #openstack-lbaas | 03:06 | |
| *** diogogmt has quit IRC | 03:20 | |
| *** bana_k has joined #openstack-lbaas | 03:24 | |
| *** bana_k has quit IRC | 03:34 | |
| *** bharath has joined #openstack-lbaas | 03:38 | |
| *** bharath has quit IRC | 03:42 | |
| *** madhu_ak has quit IRC | 03:43 | |
| *** madhu_ak has joined #openstack-lbaas | 03:43 | |
| openstackgerrit | Carlos Garza proposed openstack/octavia: Implement UDP heartbeat sender and receiver https://review.openstack.org/201882 | 03:51 |
| openstackgerrit | Carlos Garza proposed openstack/octavia: Implement UDP heartbeat sender and receiver https://review.openstack.org/201882 | 03:52 |
| *** enikanorov2 has quit IRC | 04:06 | |
| *** vivek-ebay has quit IRC | 04:12 | |
| *** madhu_ak has quit IRC | 04:14 | |
| *** madhu_ak has joined #openstack-lbaas | 04:14 | |
| *** enikanorov2 has joined #openstack-lbaas | 04:15 | |
| *** ganeshna has joined #openstack-lbaas | 04:15 | |
| *** ganeshna has quit IRC | 04:17 | |
| *** vivek-ebay has joined #openstack-lbaas | 04:21 | |
| *** madhu_ak has quit IRC | 04:22 | |
| openstackgerrit | Al Miller proposed openstack/neutron-lbaas: Filter get_pool_members to return members from the desired pool https://review.openstack.org/210968 | 04:34 |
| *** madhu_ak has joined #openstack-lbaas | 04:42 | |
| *** ganeshna has joined #openstack-lbaas | 04:49 | |
| *** ganeshna has quit IRC | 05:00 | |
| *** ganeshna has joined #openstack-lbaas | 05:00 | |
| *** vivek-ebay has quit IRC | 05:15 | |
| *** madhu_ak has quit IRC | 05:26 | |
| *** bana_k has joined #openstack-lbaas | 05:38 | |
| *** crc32 has quit IRC | 05:44 | |
| openstackgerrit | Doug Wiegley proposed openstack/neutron-lbaas: Remove default service provider from conf file https://review.openstack.org/211757 | 06:08 |
| *** bharath_ has joined #openstack-lbaas | 06:09 | |
| *** bharath_ has quit IRC | 06:14 | |
| *** apuimedo has quit IRC | 06:40 | |
| *** vivek-ebay has joined #openstack-lbaas | 06:58 | |
| *** vivek-ebay has quit IRC | 07:02 | |
| *** jschwarz has joined #openstack-lbaas | 07:21 | |
| *** bana_k has quit IRC | 07:28 | |
| rm_work | hmm, doubt anyone is up | 07:31 |
| rm_work | looking at diskimage builder stuff | 07:31 |
| rm_work | wanted to ask if there was anything else I was missing before i start poking at things :P | 07:32 |
| rm_work | does octavia build and install the disk image as part of devstack spinup? or is that manual after ./stack | 07:36 |
| *** ganeshna has quit IRC | 07:36 | |
| *** jschwarz has quit IRC | 07:41 | |
| *** openstackgerrit has quit IRC | 07:41 | |
| *** jschwarz has joined #openstack-lbaas | 07:42 | |
| *** openstackgerrit has joined #openstack-lbaas | 07:42 | |
| *** openstackgerrit has quit IRC | 07:45 | |
| *** jschwarz has quit IRC | 07:46 | |
| *** rm_work is now known as rm_work|away | 07:56 | |
| *** rm_work|away is now known as rm_work | 07:58 | |
| *** bharath has joined #openstack-lbaas | 07:58 | |
| *** jschwarz has joined #openstack-lbaas | 07:58 | |
| *** openstackgerrit has joined #openstack-lbaas | 08:00 | |
| *** rm_work is now known as rm_work|away | 08:00 | |
| *** bharath has quit IRC | 08:03 | |
| *** rm_work|away is now known as rm_work | 08:03 | |
| openstackgerrit | Evgeny Fedoruk proposed openstack/neutron-lbaas: Fixing Radware LBaaS v2 driver LB deletion https://review.openstack.org/212401 | 08:10 |
| *** chlong has quit IRC | 08:13 | |
| rm_work | man, this takes longer than I remember to spin up <_< | 08:19 |
| *** numan has joined #openstack-lbaas | 08:23 | |
| *** numan has quit IRC | 08:23 | |
| *** numan has joined #openstack-lbaas | 08:23 | |
| rm_work | oh what | 08:23 |
| rm_work | 2015-08-13 08:20:44.548 ERROR neutron.services.service_base [-] Error loading provider 'neutron_lbaas.drivers.octavia.driver.OctaviaDriver' for service LOADBALANCERV2 | 08:23 |
| rm_work | it's already automatically configured to try to use octavia? the driver isn't done yet though, right? | 08:23 |
| *** rm_work is now known as rm_work|away | 08:25 | |
| *** apuimedo has joined #openstack-lbaas | 08:25 | |
| *** rm_work|away is now known as rm_work | 08:25 | |
| *** rm_work is now known as rm_work|away | 08:28 | |
| *** rm_work|away is now known as rm_work | 08:29 | |
| rm_work | whatever, i don't need neutron-lbaas to test octavia | 08:29 |
| rm_work | AUGH | 08:32 |
| rm_work | except it's set up so i do >_> | 08:32 |
| rm_work | if ! is_service_enabled $Q_SVC || ! is_service_enabled $LBAAS_V2; then | 08:32 |
| rm_work | die "The neutron $Q-SVC and $LBAAS_V2 services must be enabled to use $OCTAVIA" | 08:32 |
| rm_work | really? :( | 08:32 |
| rm_work | but i can't enable LBAASV2 because it tries to use the nonexistent octavia driver? | 08:33 |
| rm_work | i guess i'll revisit this in the afternoon when people are on | 08:33 |
| rm_work | also there's a typo there i think | 08:34 |
| rm_work | $Q-SVC | 08:34 |
| rm_work | ah though i see now that it should build the amp image as part of initialization, cool | 08:36 |
| *** bharath__ has joined #openstack-lbaas | 09:47 | |
| *** bharath__ has quit IRC | 09:51 | |
| *** iwi has joined #openstack-lbaas | 10:26 | |
| iwi | hi there, is lbaas v2 supporting the "native" heat resources for LBaaS ? | 10:28 |
| *** chlong has joined #openstack-lbaas | 11:03 | |
| *** shunliz has joined #openstack-lbaas | 11:31 | |
| *** bharath has joined #openstack-lbaas | 11:35 | |
| *** bharath has quit IRC | 11:40 | |
| *** [1]mikeymeitbual has joined #openstack-lbaas | 12:01 | |
| *** jschwarz_ has joined #openstack-lbaas | 12:03 | |
| *** mikeymeitbual has quit IRC | 12:03 | |
| *** [1]mikeymeitbual is now known as mikeymeitbual | 12:03 | |
| *** jschwarz has quit IRC | 12:05 | |
| *** evgenyf has joined #openstack-lbaas | 12:11 | |
| *** jschwarz__ has joined #openstack-lbaas | 12:31 | |
| *** ajmiller has quit IRC | 12:33 | |
| *** jschwarz_ has quit IRC | 12:34 | |
| *** openstackgerrit has quit IRC | 13:01 | |
| *** openstackgerrit has joined #openstack-lbaas | 13:02 | |
| *** jschwarz__ is now known as jschwarz | 13:02 | |
| *** mestery has quit IRC | 13:03 | |
| *** mestery has joined #openstack-lbaas | 13:11 | |
| *** ajmiller has joined #openstack-lbaas | 13:14 | |
| *** rohara has joined #openstack-lbaas | 13:22 | |
| *** bharath has joined #openstack-lbaas | 13:24 | |
| *** diogogmt has joined #openstack-lbaas | 13:28 | |
| *** bharath has quit IRC | 13:28 | |
| *** ajmiller has quit IRC | 13:43 | |
| *** diogogmt has quit IRC | 13:49 | |
| *** evgenyf has quit IRC | 14:04 | |
| *** evgenyf has joined #openstack-lbaas | 14:06 | |
| *** iwi has quit IRC | 14:08 | |
| *** devlaps has quit IRC | 14:24 | |
| *** iwi has joined #openstack-lbaas | 14:25 | |
| *** minwang2 has joined #openstack-lbaas | 14:33 | |
| *** mlavalle has joined #openstack-lbaas | 14:35 | |
| *** sbalukoff has quit IRC | 14:37 | |
| *** iwi has quit IRC | 14:38 | |
| *** iwi has joined #openstack-lbaas | 14:39 | |
| *** fnaval has quit IRC | 14:43 | |
| *** jschwarz has quit IRC | 14:43 | |
| openstackgerrit | Pengtao Huang proposed openstack/neutron-lbaas: change if to elif https://review.openstack.org/212560 | 14:49 |
| *** vivek-ebay has joined #openstack-lbaas | 14:50 | |
| *** ajmiller has joined #openstack-lbaas | 14:53 | |
| *** iwi1 has joined #openstack-lbaas | 14:54 | |
| *** iwi has quit IRC | 14:54 | |
| *** diogogmt has joined #openstack-lbaas | 14:54 | |
| *** fnaval has joined #openstack-lbaas | 14:56 | |
| *** TrevorV has joined #openstack-lbaas | 14:58 | |
| *** vivek-ebay has quit IRC | 14:59 | |
| *** sbalukoff has joined #openstack-lbaas | 15:04 | |
| *** bharath has joined #openstack-lbaas | 15:13 | |
| *** bharath has quit IRC | 15:18 | |
| *** haigang has joined #openstack-lbaas | 15:24 | |
| *** haigang has quit IRC | 15:24 | |
| *** haigang has joined #openstack-lbaas | 15:25 | |
| *** shunliz has quit IRC | 15:29 | |
| *** vivek-ebay has joined #openstack-lbaas | 15:32 | |
| *** amotoki has joined #openstack-lbaas | 15:44 | |
| *** haigang has quit IRC | 15:49 | |
| *** haigang has joined #openstack-lbaas | 15:49 | |
| *** sbalukoff has quit IRC | 15:54 | |
| *** jschwarz has joined #openstack-lbaas | 15:56 | |
| *** woodster_ has joined #openstack-lbaas | 15:57 | |
| *** minwang2 has quit IRC | 16:02 | |
| *** haigang has quit IRC | 16:15 | |
| *** evgenyf has quit IRC | 16:17 | |
| *** alejandrito has joined #openstack-lbaas | 16:22 | |
| xgerman | dougwig, blogan: https://review.openstack.org/#/c/210968/4/neutron_lbaas/services/loadbalancer/plugin.py | 16:24 |
| xgerman | wonder if you an have a look and comment | 16:25 |
| ajmiller | also see https://bugs.launchpad.net/neutron/+bug/1483100 | 16:25 |
| openstack | Launchpad bug 1483100 in neutron "neutron-lbaas V2 lbaas-member-list returns all members, not just those for the specified pool" [Undecided,In progress] - Assigned to Al Miller (al-miller) | 16:25 |
| *** abdelwas has joined #openstack-lbaas | 16:29 | |
| *** diogogmt has quit IRC | 16:31 | |
| *** openstackgerrit has quit IRC | 16:31 | |
| *** diogogmt has joined #openstack-lbaas | 16:31 | |
| *** openstackgerrit has joined #openstack-lbaas | 16:32 | |
| *** crc32 has joined #openstack-lbaas | 16:33 | |
| jschwarz | dougwig, can I ask for an oh-so-old lbaas v1 patch review? https://review.openstack.org/#/c/174384/ :) | 16:43 |
| *** KunalGandhi has joined #openstack-lbaas | 16:45 | |
| *** enikanorov2 has quit IRC | 16:48 | |
| *** minwang2 has joined #openstack-lbaas | 16:52 | |
| *** apuimedo has quit IRC | 16:55 | |
| *** enikanorov2 has joined #openstack-lbaas | 16:57 | |
| *** jschwarz has quit IRC | 16:58 | |
| *** bharath has joined #openstack-lbaas | 17:02 | |
| *** bharath has quit IRC | 17:06 | |
| *** vivek-ebay has quit IRC | 17:08 | |
| *** sbalukoff has joined #openstack-lbaas | 17:17 | |
| *** bharath has joined #openstack-lbaas | 17:23 | |
| *** amotoki has quit IRC | 17:25 | |
| *** vivek-ebay has joined #openstack-lbaas | 17:32 | |
| *** vivek-ebay has quit IRC | 17:32 | |
| *** bana_k has joined #openstack-lbaas | 17:34 | |
| *** bana_k has quit IRC | 17:46 | |
| *** madhu_ak has joined #openstack-lbaas | 17:49 | |
| *** SumitNaiksatam has joined #openstack-lbaas | 17:52 | |
| rm_work | xgerman / ajmiller / johnsom: any of you around to explain how i get devstack working? | 17:53 |
| ajmiller | rm_work. I'm around, but having devstack problems myself. | 17:53 |
| *** vivek-ebay has joined #openstack-lbaas | 17:53 | |
| johnsom | ajmiller is the expert. | 17:53 |
| ajmiller | What is your issue? | 17:53 |
| rm_work | is it the think with ERROR: openstack Missing parameter(s): | 17:53 |
| rm_work | if so, i know the fix for THAT | 17:53 |
| xgerman | I haven’t installed in a while but in scroll back I saw you had stock LBaaS — you need dougwig’s patch | 17:53 |
| johnsom | BTW, the devstack scripts do create and load the amp image into glance and configure octavia for it | 17:54 |
| *** vivek-ebay has quit IRC | 17:54 | |
| rm_work | my issue is that neutron-lbaas won't start because it can't find the octavia driver | 17:54 |
| rm_work | ahhh yeah that is what i figured | 17:54 |
| rm_work | so which patch do i need | 17:54 |
| rm_work | ajmiller: what problem are you having? | 17:54 |
| johnsom | I cherry pick this one: https://review.openstack.org/#/c/174114/ | 17:54 |
| rm_work | is that "enable_plugin neutron-lbaas https://review.openstack.org/openstack/neutron-lbaas https://review.openstack.org/#/c/174114/" ? | 17:55 |
| ajmiller | rm_work. I don't have routability from the devstack host OS into my devstack's private network, using my default "sample" scripts. | 17:55 |
| rm_work | ajmiller: ah, ok, then nothing i can help with yet :P | 17:56 |
| rm_work | trying to remember how to tell it to pick a certain review | 17:56 |
| ajmiller | It used to "just work". Now it doesn't. Something has changed in the environment, like an iptables rule or somesuch. | 17:56 |
| *** sc68cal has quit IRC | 17:56 | |
| rm_work | ah no it's "enable_plugin neutron-lbaas https://review.openstack.org/openstack/neutron-lbaas refs/changes/14/174114/20" | 17:57 |
| rm_work | kk | 17:57 |
| ajmiller | rm_work -- if you see "no route to host" errors, then it might be what I'm seeing. | 17:58 |
| * ajmiller wonders how I did that. | 17:58 | |
| *** rm_work is now known as rm_work|away | 17:58 | |
| *** rm_work|away is now known as rm_work | 18:00 | |
| *** sc68cal has joined #openstack-lbaas | 18:02 | |
| *** sc68cal has quit IRC | 18:02 | |
| *** sc68cal has joined #openstack-lbaas | 18:02 | |
| *** bana_k has joined #openstack-lbaas | 18:26 | |
| TrevorV | Hey all, general question: can more than one review be dependent on a single review? | 18:45 |
| TrevorV | Example: bana_k needs doug's octavia driver review in lbaas, but I do as well, but don't necessarily need bana_k 's review... so can we BOTH depend on dougwig's review? | 18:45 |
| iwi1 | hi all, is there any documentation on how to enable lbas v2 on non-devstack kilo ? | 18:47 |
| iwi1 | and second question - can V2 be used from heat (in kilo) ? | 18:47 |
| rm_work | TrevorV: yeah no problem with that | 18:54 |
| rm_work | it's a tree | 18:55 |
| blogan | iwi1: not much documentation for that, though it'd be very similar to v1 | 18:57 |
| blogan | iwi1: and i do not believe heat has been updated for v2 | 18:58 |
| TrevorV | Yeah, blogan answered for me rm_work thanks | 18:58 |
| *** jorgem has joined #openstack-lbaas | 18:58 | |
| iwi1 | blogan: thanks | 18:59 |
| openstackgerrit | Trevor Vardeman proposed openstack/neutron-lbaas: Fixing Octavia Driver delete operations https://review.openstack.org/212702 | 19:00 |
| *** bharath has quit IRC | 19:03 | |
| *** bharath has joined #openstack-lbaas | 19:04 | |
| *** bharath has quit IRC | 19:08 | |
| TrevorV | dougwig, you present my man? | 19:20 |
| *** fnaval has quit IRC | 19:23 | |
| *** kfox1111 has joined #openstack-lbaas | 19:25 | |
| kfox1111 | I'm on kilo, but thinking ahead towards liberty where we will start playing with octavia... | 19:26 |
| kfox1111 | we've got some grid services that are https, but do client cert based auth. | 19:26 |
| kfox1111 | will that still work with the octavia ssl termination stuff, or does octavia have to just do tcp lb? | 19:27 |
| rm_work | hey where's the API docs for neutron-lbaas v2? | 19:27 |
| rm_work | xgerman / blogan / ajmiller / johnsom ^^ | 19:27 |
| *** fnaval has joined #openstack-lbaas | 19:28 | |
| ajmiller | rm_work http://developer.openstack.org/api-ref-networking-v2-ext.html | 19:30 |
| rm_work | k | 19:31 |
| rm_work | that is what i found too, but no TLS info in there yet | 19:31 |
| *** bharath has joined #openstack-lbaas | 19:32 | |
| *** numan has quit IRC | 19:33 | |
| fnaval | kicking off usage tests | 19:34 |
| fnaval | damn - wrong window | 19:34 |
| *** alejandrito has quit IRC | 19:44 | |
| rm_work | fnaval: ok go for it | 19:50 |
| TrevorV | johnsom, you around? | 19:57 |
| ajmiller | TrevorV He's out to lunch.. | 19:59 |
| TrevorV | ajmiller, any idea how long ago he left so I can talk to him when he gets back? | 19:59 |
| ajmiller | Just under an hour ago. | 20:00 |
| TrevorV | Alright, so relatively soon I should be able to talk with him? | 20:00 |
| ajmiller | I would think so. | 20:00 |
| rm_work | :P | 20:01 |
| rm_work | TrevorV: your impatience rivals even my own :P | 20:01 |
| fnaval | k thanks rm_work | 20:01 |
| TrevorV | ajmiller, I only ask because sometimes people have to go to lunch and run errands or have appointments outside of work, or meetings, etc etc that colleagues know about, so I was trying to find an ETA. Sorry for bothering ya! | 20:02 |
| TrevorV | rm_work, its not impatience, its just we have a meeting in 30 | 20:02 |
| TrevorV | And I don't want to miss it. | 20:02 |
| ajmiller | NP. | 20:02 |
| rm_work | oh right | 20:03 |
| rm_work | that think | 20:03 |
| rm_work | *that thing | 20:03 |
| johnsom | TrevorV just got back from lunch | 20:21 |
| rm_work | ok cool, script seems to build octavia now | 20:23 |
| rm_work | err, build devstack with octavia | 20:23 |
| rm_work | so now i just need to test the new thing | 20:24 |
| TrevorV | johnsom, I'll PM you | 20:24 |
| *** alejandrito has joined #openstack-lbaas | 20:28 | |
| openstackgerrit | Merged openstack/neutron-lbaas: Remove ipv6_pd_enabled from subnet data model https://review.openstack.org/211615 | 20:38 |
| *** sbalukoff has quit IRC | 20:46 | |
| *** woodster_ has quit IRC | 20:50 | |
| *** woodster_ has joined #openstack-lbaas | 20:54 | |
| *** sbalukoff has joined #openstack-lbaas | 20:58 | |
| xgerman | kfix1111 Octavia will have SSL termination build in | 21:01 |
| xgerman | kfox1111 | 21:01 |
| xgerman | it’s on our list for Liberty: https://etherpad.openstack.org/p/YVR-neutron-octavia | 21:02 |
| xgerman | see topic Prioritized List | 21:02 |
| johnsom | kfox1111 Actually, I don't think client certs for authentication is in scope for Liberty Octavia. Just one way TLS offload | 21:10 |
| xgerman | you are right — at least we only tested what you are describing ;-) | 21:12 |
| johnsom | You can of course still pass the TLS through Octavia as a TCP flow, but we don't have TLS client auth in there yet. | 21:13 |
| johnsom | kfox1111 It would be great for us if you could write up your use case in a bug in our launchpad: https://bugs.launchpad.net/octavia/+filebug | 21:26 |
| xgerman | well, RfE would be great indeed... | 21:27 |
| openstackgerrit | Adam Harwell proposed openstack/octavia: Updating cert_parser for cert loading https://review.openstack.org/208035 | 21:27 |
| rm_work | had to rebase for testing | 21:28 |
| xgerman | ok | 21:28 |
| johnsom | rm_work Did you come up with any ideas on https://review.openstack.org/#/c/211371/ | 21:30 |
| johnsom | ? | 21:30 |
| rm_work | that was also stalled on me actually getting devstack set up | 21:32 |
| rm_work | which is now working | 21:32 |
| rm_work | and I wanted to ask you what actually comes back there | 21:33 |
| rm_work | what does the json request look like | 21:33 |
| rm_work | i have never seen decode("string-escape") | 21:33 |
| rm_work | but need to know what an example string looks like | 21:33 |
| johnsom | For what ever reason, it comes back with the "'s \"'d | 21:34 |
| rm_work | what is the base string tho | 21:34 |
| rm_work | prior to json decoding at all | 21:34 |
| johnsom | The other folks impacted on google said to use the string-escape codec. | 21:35 |
| johnsom | Which blows chunks on the py34 unit test | 21:35 |
| rm_work | i need to figure out how to debug this | 21:35 |
| rm_work | which if i had a sample string would be easier | 21:36 |
| TrevorV | You're a simple strin | 21:36 |
| TrevorV | string** | 21:36 |
| johnsom | I closed all of those windows. It will take me a few minutes to setup an image with the string logged for you | 21:37 |
| *** ajmiller has quit IRC | 21:37 | |
| rm_work | '{"a": "Hello world"}' | 21:37 |
| rm_work | ^^ that is what JSON is supposed to look like in string form | 21:37 |
| rm_work | and when it includes strings | 21:37 |
| TrevorV | johnsom, I have to leave for the day. GOt some stuff at home I have to hurry for. I'm going to save that etherpad link and then check it later/tomorrow when I can get to it again. okay>? | 21:37 |
| TrevorV | okay?**** | 21:37 |
| johnsom | Right. The stuff coming out of flask on py27 is like (from memory here) '{\"a\": \"Hello world\"}' | 21:38 |
| rm_work | so I wonder what flask is returning | 21:38 |
| rm_work | ahh hmm | 21:38 |
| openstackgerrit | Trevor Vardeman proposed openstack/octavia: Adding amphora failover flows https://review.openstack.org/202336 | 21:38 |
| johnsom | TrevorV Sounds good! | 21:38 |
| rm_work | I think flask has functions for decoding json internally, maybe those would be a better bet? | 21:39 |
| johnsom | rm_work give me a few minutes (it is pretty quick) and I will setup a test image | 21:39 |
| rm_work | kk | 21:39 |
| TrevorV | blogan, other interested parties: | 21:39 |
| TrevorV | https://etherpad.openstack.org/p/Octavia_Amphora_Failover_Steps | 21:39 |
| crc32 | johnsom: xgerman is suggesting we ditch the status_sender.json file and inject an oslo formated config. Did you ever write that code to do this. I was under the impression you did this at the mid ycle | 21:39 |
| *** TrevorV has quit IRC | 21:39 | |
| rm_work | flask.request.get_json() | 21:39 |
| johnsom | Yeah, they didn't help, tried them. get_json() and such. I also ran into some trouble with the flask test methods and the get_json stuff | 21:40 |
| rm_work | hmm | 21:40 |
| johnsom | xgerman already beat me up over that | 21:40 |
| *** ajmiller has joined #openstack-lbaas | 21:40 | |
| xgerman | crc32 the sender was conceived when it wasn’t clear that we would have the whole config file on the amphora via the config drive (if we use the REST agent) | 21:40 |
| crc32 | he's beating me up over it now. | 21:41 |
| johnsom | crc32 I was working on it. I saw those comment this morning and was going to sync with you. I will take that on and finish it, no problem. | 21:41 |
| *** alejandrito has quit IRC | 21:41 | |
| crc32 | ok. Then I'll write up some mock teses. | 21:41 |
| xgerman | johnsom thank you | 21:41 |
| crc32 | tests | 21:41 |
| *** madhu_ak has quit IRC | 21:41 | |
| xgerman | crc32 also didn’t want to beat — I just noticed the situation has changed with the wider acceptance of the REST agent :-) | 21:41 |
| johnsom | The xgerman beating was for not using get_json() | 21:42 |
| rm_work | I am trying to set up my own quick test to see if i can replicate it with a standalone flask app | 21:43 |
| *** madhu_ak has joined #openstack-lbaas | 21:43 | |
| xgerman | yeah, the whole thing is bizarre — I suggested to him the .json stuff as well | 21:43 |
| xgerman | and I made him switch from form parameters to json | 21:43 |
| johnsom | Yeah, man, the form worked just fine... grin | 21:45 |
| rm_work | heh | 21:45 |
| johnsom | rm_work "{\"subnet_cidr\": \"10.0.0.0/24\", \"gateway\": \"10.0.0.1\"}" | 21:55 |
| rm_work | yeah ok seeing similar | 21:55 |
| johnsom | I wasted a day trying to figure out why and a better way to deal with it. | 21:56 |
| rm_work | sent in this: | 21:56 |
| rm_work | {'a': 1} | 21:56 |
| rm_work | '{\\'a\\': 1}' | 21:56 |
| rm_work | get that | 21:56 |
| crc32 | like """{subnet_cidr":"10.0.0.0/24"}""" | 21:56 |
| crc32 | oh did you json dump the string twice? | 21:58 |
| rm_work | johnsom: hmm looks like if i use double-quotes on the sending-side | 21:58 |
| rm_work | johnsom: then it looks ok | 21:58 |
| rm_work | {"a": "hello"} -> '{"a": "hello"}' | 21:58 |
| rm_work | and in this case .get_json() works | 21:59 |
| johnsom | But I'm giving json.loads a dict.... https://review.openstack.org/#/c/208793/5/octavia/amphorae/drivers/haproxy/rest_api_driver.py | 22:00 |
| johnsom | opps, dumps | 22:00 |
| rm_work | so the "sending side" is directly from the result of json.dumps() ? | 22:00 |
| johnsom | Ye[ | 22:01 |
| johnsom | yep | 22:01 |
| rm_work | hmm | 22:01 |
| openstackgerrit | Sherif Abdelwahab proposed openstack/octavia: Amphora Flows and Service Drivers for Active Standby https://review.openstack.org/206252 | 22:01 |
| crc32 | json.dumps(json.dumps({"test":0})) -> '"{\\"test\\": 0}"' | 22:01 |
| rm_work | json.dumps({'a': "hello"}) -> '{"a": "hello"}' | 22:01 |
| rm_work | yeah | 22:02 |
| rm_work | that looks like a double-dump | 22:02 |
| * blogan snickers at double-dump | 22:02 | |
| rm_work | :P | 22:02 |
| crc32 | print json.dumps(json.dumps({'test':0}) | 22:02 |
| rm_work | crc32 is correct tho, it looks like maybe it's being encoded twice on the other end | 22:02 |
| rm_work | where is the "other end" of this | 22:03 |
| rm_work | ah you already linked it | 22:03 |
| johnsom | https://review.openstack.org/#/c/211371/1/octavia/amphorae/backends/agent/api_server/server.py | 22:03 |
| rm_work | ah | 22:04 |
| rm_work | uhh | 22:04 |
| rm_work | if that is using requests to post | 22:04 |
| rm_work | i think "json" is supposed to take a dict | 22:04 |
| rm_work | not a string | 22:04 |
| rm_work | r = requests.post(url, data=json.dumps(payload)) | 22:05 |
| rm_work | or | 22:05 |
| rm_work | r = requests.post(url, json=payload) | 22:05 |
| rm_work | i THINK | 22:05 |
| crc32 | json.dumps takes a string. | 22:06 |
| rm_work | err | 22:06 |
| crc32 | sorry takes a dict | 22:06 |
| rm_work | yes | 22:06 |
| rm_work | i think you either dumps() the dict and provide it as data | 22:06 |
| rm_work | or just provide the dict directly as "json" | 22:06 |
| rm_work | trying to find the approproate Requests doc | 22:07 |
| rm_work | *appropriate | 22:07 |
| rm_work | yeah can't find the doc | 22:08 |
| rm_work | but i tested | 22:08 |
| rm_work | http://stackoverflow.com/a/26344315/4655717 | 22:09 |
| rm_work | you're double-encoding on the other end | 22:09 |
| rm_work | https://review.openstack.org/#/c/208793/5/octavia/amphorae/drivers/haproxy/rest_api_driver.py line 308 | 22:10 |
| rm_work | johnsom: ^^ | 22:10 |
| rm_work | thanks for the catch crc32 | 22:10 |
| rm_work | johnsom: sorry I didn't catch it in the original review | 22:13 |
| johnsom | That's ok. The requests docs are so clear on this issue: json – (optional) json to send in the body of the Request. | 22:14 |
| johnsom | Which implies to me it's json encoded. | 22:14 |
| johnsom | So, riddle is why does it not fail on py34 | 22:14 |
| johnsom | I will give this a try | 22:14 |
| rm_work | yeah | 22:15 |
| rm_work | docs are bad here | 22:15 |
| rm_work | i only know because i use requests *a lot* | 22:15 |
| rm_work | for personal projects | 22:15 |
| rm_work | and also please switch at the same time to using get_json() | 22:15 |
| rm_work | it really is the best way to do this | 22:16 |
| rm_work | I am guessing the only reason it was failing was due to the double-encode | 22:16 |
| rm_work | because i can reproduce that failure here | 22:16 |
| rm_work | err, locally, in my test flask app | 22:16 |
| *** jorgem has quit IRC | 22:20 | |
| johnsom | rm_work works. Head slap. I will fix the patchset now | 22:22 |
| rm_work | kk | 22:23 |
| *** jorgem has joined #openstack-lbaas | 22:23 | |
| *** vivek-ebay has joined #openstack-lbaas | 22:26 | |
| rm_work | xgerman: did you want to +2 the current iteration of https://review.openstack.org/#/c/208035/ ? | 22:33 |
| rm_work | before I workflow it, assuming this test works | 22:34 |
| xgerman | sure, you still need to +A | 22:34 |
| johnsom | I haven't had a chance to look at that one yet | 22:35 |
| xgerman | it already has three +2 | 22:35 |
| xgerman | but the more the merrier | 22:35 |
| *** vivek-ebay has quit IRC | 22:46 | |
| *** vivek-ebay has joined #openstack-lbaas | 22:46 | |
| bana_k | I see couple of examples in lbaas driver where tls info is passed on as 'default_tls_container_id': listener.default_tls_container_id and sni as sni_container_ids in listener dict | 22:51 |
| bana_k | is that all is need for octaiva driver also? | 22:51 |
| bana_k | needed* | 22:52 |
| bana_k | xgerman blogan johnsom TrevorV | 22:56 |
| bana_k | ? | 22:56 |
| johnsom | bana_k I am here, but I don't have the details to answer your question. Sorry | 22:56 |
| bana_k | oh ok. NP | 22:57 |
| blogan | bana_k: are you working off this review: https://review.openstack.org/#/c/209682/ | 22:57 |
| bana_k | no | 22:58 |
| blogan | bana_k: i believe pothole was working on that and discovered octavia did not have sni exposed through the API so he was adding that | 22:58 |
| *** jorgem has quit IRC | 22:58 | |
| blogan | bana_k: https://review.openstack.org/#/c/209684/ | 22:58 |
| blogan | bana_k: but if you look there, octavia takes tls_certificate_id, instead of what neutron_lbaas's deafult_tls_container_id | 22:59 |
| bana_k | but plugin will be passing the container id isn't it? | 23:00 |
| *** woodster_ has quit IRC | 23:00 | |
| *** abdelwas has quit IRC | 23:01 | |
| openstackgerrit | Michael Johnson proposed openstack/octavia: Fixes REST driver json encoding https://review.openstack.org/211371 | 23:02 |
| blogan | bana_k: yes | 23:02 |
| johnsom | rm_work ^^ | 23:02 |
| blogan | bana_k: are you working on the octavia driver to pass tls info to octavia? | 23:02 |
| bana_k | yes | 23:02 |
| blogan | bana_k: so yeah it should just be as simple as passing what the plugin gives you, and giving it to octavia but octavia accepts it as that other name | 23:07 |
| rm_work | which again, pothole had already done in that review blogan linked earlier, no? | 23:07 |
| bana_k | oh k . Got it | 23:07 |
| blogan | rm_work: correct | 23:08 |
| blogan | it still needs tests | 23:08 |
| xgerman | tests... | 23:08 |
| rm_work | johnsom: LGTM, +2/+A | 23:09 |
| rm_work | assuming tests pass, it'll go right in | 23:09 |
| johnsom | Thanks! | 23:09 |
| johnsom | Yeah, I tox'd it up | 23:09 |
| rm_work | note that I am not afraid of +A pre-jenkins | 23:09 |
| rm_work | because if jenkins has a problem it'll override the +A anyway | 23:09 |
| johnsom | Haha | 23:09 |
| rm_work | people always question me about that <_< | 23:09 |
| rm_work | i just look at it as efficiency | 23:09 |
| johnsom | Question or shame? | 23:10 |
| rm_work | both :P | 23:10 |
| johnsom | I at least wait for the sonar job because then I know it has passed py27 | 23:10 |
| *** minwang2 has quit IRC | 23:12 | |
| *** minwang2 has joined #openstack-lbaas | 23:13 | |
| rm_work | erk, just noticed this stuff is incorrectly instantiating a BarbicanCertManager directly... https://review.openstack.org/#/c/208035/4/octavia/amphorae/drivers/haproxy/ssh_driver.py,cm | 23:14 |
| rm_work | i guess I'll make the fix for that as a followup | 23:14 |
| rm_work | rather than insist pothole fix code in that review that he didn't touch :P | 23:14 |
| johnsom | Or you could update this one, and the rest driver.... | 23:15 |
| rm_work | xgerman: i have a question on https://review.openstack.org/#/c/208035/4 which maybe pothole should be answering but possibly you can answer too? | 23:18 |
| rm_work | does the rest driver do that too johnsom ? | 23:18 |
| bana_k | yea, so I am kinda confused about what I am supposed to do :D | 23:18 |
| rm_work | bana_k: what was the actual task? | 23:18 |
| rm_work | "fix up the octavia driver for neutron-lbaas" | 23:19 |
| rm_work | ? | 23:19 |
| xgerman | mmh, let me check | 23:20 |
| bana_k | "there was some work to do in the LBaaSv2 Octavia driver to finish up TLS" | 23:20 |
| rm_work | ah | 23:20 |
| openstackgerrit | Merged openstack/octavia: Fixes REST driver json encoding https://review.openstack.org/211371 | 23:20 |
| rm_work | yeah so i think it might just be to finish testing pothole's CR | 23:20 |
| bana_k | hmm ok. | 23:21 |
| xgerman | bana_k looks like it | 23:21 |
| xgerman | rm_work no the rest driver hasn’t that code yet | 23:21 |
| johnsom | rm_work It looks like the REST driver borrowed the code from the old ssh driver.... | 23:22 |
| xgerman | I am sort of behind what i’s being put into the ssh driver | 23:22 |
| openstackgerrit | Adam Harwell proposed openstack/octavia: Fix a typo in the devstack plugin.sh which caused a truncated error https://review.openstack.org/212834 | 23:22 |
| rm_work | ^^ trivial | 23:22 |
| rm_work | just was annoying me last night | 23:22 |
| rm_work | I kept getting "ERROR: The neutron" | 23:23 |
| rm_work | and that was it T_T | 23:23 |
| johnsom | I will be bold on that one and +2 | 23:24 |
| rm_work | heh | 23:24 |
| rm_work | xgerman: fastest workflow ever, or FASTEST WORKFLOW EVER? :P | 23:24 |
| rm_work | one character change FTW | 23:24 |
| xgerman | done | 23:25 |
| rm_work | like magic :) | 23:25 |
| johnsom | rm_work People will think you are fishing for ATC.... | 23:25 |
| rm_work | hah | 23:25 |
| xgerman | I have seen a couple of those if -> elif | 23:27 |
| xgerman | there was even somebody who looked like Doug’s wife | 23:27 |
| rm_work | yeah i saw that :P | 23:27 |
| johnsom | Hahaha, I am not going to touch that comment | 23:28 |
| rm_work | Too many Wiegleys :P | 23:28 |
| xgerman | rm_work regarding the ssl cert stuff I think pothole also has a patch which refactors that so we can share code better between drivers | 23:29 |
| davidlenwell | hey.. do we have a digital image of the octavia sticker? | 23:30 |
| davidlenwell | I want to put it in a slide for a talk I am giving | 23:30 |
| xgerman | Julian did those — I will see if he is still around | 23:31 |
| davidlenwell | thanks xgerman | 23:31 |
| openstackgerrit | Merged openstack/octavia: Fix a typo in the devstack plugin.sh which caused a truncated error https://review.openstack.org/212834 | 23:36 |
| davidlenwell | thanks xgerman!! | 23:42 |
| xgerman | great — make us proud... | 23:43 |
| xgerman | though I am hoping sbalukoff is there and heckles :-) | 23:43 |
| davidlenwell | you will be a small part of my "state of advanced services" portion of my talk | 23:43 |
| sbalukoff | Oh, I plan on it. | 23:44 |
| davidlenwell | hopefully I don't get too much heckling from bluebox.. I might have to flip a table over and walk out ;) | 23:44 |
| davidlenwell | you'll be there tonight sbalukoff? | 23:44 |
| sbalukoff | Oh wait! That's tonight? Sorry, I actually won't be there tonight. | 23:44 |
| davidlenwell | then I can say anything I want about you ;) | 23:45 |
| rm_work | we usually do that already :P | 23:45 |
| * rm_work waves at sbalukoff | 23:45 | |
| sbalukoff | (Ended up working from home-- caring for a stupid liar (and sick) kitty) | 23:45 |
| sbalukoff | Haha! | 23:45 |
| rm_work | fixing this issue presently | 23:46 |
| rm_work | ^^ using the Barbican stuff directly | 23:46 |
| rm_work | I don't see it being used in the REST driver? | 23:46 |
| xgerman | nope | 23:47 |
| openstackgerrit | Bharath M proposed openstack/octavia: Add Housekeeping to manage spare amphora https://review.openstack.org/202829 | 23:47 |
| rm_work | hmm | 23:48 |
| rm_work | wtf | 23:48 |
| rm_work | cfg.StrOpt('cert_manager', | 23:48 |
| rm_work | default='barbican_cert_manager', | 23:48 |
| rm_work | help=_('Name of the cert manager to use')), | 23:48 |
| rm_work | ended up in "haproxy_amphora_opts" | 23:48 |
| rm_work | thus config group "haproxy_amphora" | 23:49 |
| rm_work | which is weird | 23:49 |
| rm_work | the certmanager itself registers opts already, and does so in the "certificates" group | 23:50 |
| *** diogogmt has quit IRC | 23:51 | |
| xgerman | mmh | 23:51 |
| rm_work | looks like you did that xgerman | 23:51 |
| xgerman | yep, I recall | 23:52 |
| rm_work | but i also don't really understand how this aliasing stuff works | 23:52 |
| rm_work | the default is "barbican_cert_manager" which i saw defined in some setup config | 23:52 |
| xgerman | that cert stuff is for generating the cert to talk to the amphora | 23:52 |
| rm_work | to alias the full path name | 23:52 |
| rm_work | hmm | 23:52 |
| rm_work | but it's for certManager | 23:52 |
| rm_work | not certGenerator | 23:52 |
| rm_work | O_o | 23:52 |
| rm_work | xgerman: can you collaborate with me on this patchset? I'll submit it with that removed and the fixes to the SSH driver | 23:53 |
| rm_work | and you can pull it down and make a patchset to fix REST to match? | 23:53 |
| xgerman | sure, can do | 23:53 |
| rm_work | and fix it to use the alias because i am not sure how you did that | 23:54 |
| * johnsom ducks | 23:54 | |
| rm_work | lol | 23:54 |
| xgerman | yeah, I have to try to remember as well | 23:54 |
| rm_work | johnsom: do you know how those config aliases work? :P | 23:54 |
| johnsom | I was ducking because I thought xgerman was going to volutell me to take on the REST part. That said, I have spent some time looking at how the oslo config stuff works. | 23:55 |
| xgerman | nah, I will do that myself :-) | 23:56 |
| xgerman | REST — oslo I need to try to remember | 23:56 |
| rm_work | ah | 23:57 |
| rm_work | it's that we half-converted it to using Stevedore | 23:57 |
| rm_work | but only half | 23:57 |
| rm_work | ok, I can finish that conversion I guess | 23:57 |
| johnsom | rm_work What is your question? it's aliased to "haproxy_amphora" below in the register_opts section | 23:58 |
| rm_work | johnsom: somehow "barbican_cert_manager" is aliased to the full class path | 23:58 |
| rm_work | for BarbicanCertManager() | 23:58 |
| rm_work | it's a stevedore thing | 23:58 |
| *** jorgem has joined #openstack-lbaas | 23:58 | |
| rm_work | i saw it in a config once | 23:58 |
| johnsom | rm_work Yeah, entry_points in the setup.cfg | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!