| *** bana_k has quit IRC | 00:03 | |
| *** bharath has left #openstack-lbaas | 00:04 | |
| *** ganeshna has joined #openstack-lbaas | 00:19 | |
| *** bharathm has joined #openstack-lbaas | 00:19 | |
| *** ganeshna has quit IRC | 00:23 | |
| *** bharathm has quit IRC | 00:44 | |
| *** amotoki has joined #openstack-lbaas | 00:47 | |
| *** amotoki has quit IRC | 00:50 | |
| *** amotoki has joined #openstack-lbaas | 00:56 | |
| *** vjay11 has joined #openstack-lbaas | 01:43 | |
| *** bharathm has joined #openstack-lbaas | 01:44 | |
| *** bharathm has quit IRC | 01:49 | |
| *** KunalGan_ has joined #openstack-lbaas | 01:54 | |
| *** vjay11 has quit IRC | 01:56 | |
| *** KunalGandhi has quit IRC | 01:57 | |
| *** ganeshna has joined #openstack-lbaas | 02:20 | |
| *** ganeshna has quit IRC | 02:25 | |
| *** bharathm has joined #openstack-lbaas | 02:47 | |
| *** bharathm has quit IRC | 02:52 | |
| *** ganeshna has joined #openstack-lbaas | 02:58 | |
| *** ganeshna has quit IRC | 03:05 | |
| *** Aish has joined #openstack-lbaas | 03:06 | |
| *** KunalGan_ has quit IRC | 03:17 | |
| *** Aish has left #openstack-lbaas | 03:22 | |
| *** vivek-ebay has joined #openstack-lbaas | 03:29 | |
| *** openstackgerrit has quit IRC | 03:31 | |
| *** openstackgerrit has joined #openstack-lbaas | 03:32 | |
| *** diogogmt has quit IRC | 04:03 | |
| *** vivek-ebay has quit IRC | 04:09 | |
| *** vivek-ebay has joined #openstack-lbaas | 04:17 | |
| *** bharathm has joined #openstack-lbaas | 04:36 | |
| *** ajmiller has quit IRC | 04:38 | |
| *** bharathm has quit IRC | 04:40 | |
| *** Aish has joined #openstack-lbaas | 04:42 | |
| *** Aish has left #openstack-lbaas | 05:09 | |
| *** kiran-r has joined #openstack-lbaas | 05:10 | |
| *** numan has joined #openstack-lbaas | 05:21 | |
| *** Alex_Stef has joined #openstack-lbaas | 05:23 | |
| *** vivek-ebay has quit IRC | 05:24 | |
| *** numan has quit IRC | 05:26 | |
| *** ganeshna has joined #openstack-lbaas | 05:32 | |
| *** numan has joined #openstack-lbaas | 05:38 | |
| *** bharathm has joined #openstack-lbaas | 05:50 | |
| *** reedip has joined #openstack-lbaas | 06:24 | |
| *** _kiran_ has joined #openstack-lbaas | 06:25 | |
| *** kiran-r has quit IRC | 06:27 | |
| *** vjay11 has joined #openstack-lbaas | 06:49 | |
| *** nmagnezi has joined #openstack-lbaas | 06:56 | |
| *** bharathm has quit IRC | 06:57 | |
| *** bana_k has joined #openstack-lbaas | 07:03 | |
| *** evgenyf has joined #openstack-lbaas | 07:05 | |
| *** haigang has joined #openstack-lbaas | 07:05 | |
| *** haigang has quit IRC | 07:06 | |
| *** I has joined #openstack-lbaas | 07:06 | |
| *** I is now known as Guest63057 | 07:07 | |
| *** Guest63057 has quit IRC | 07:07 | |
| *** haigang has joined #openstack-lbaas | 07:07 | |
| *** haigang has quit IRC | 07:09 | |
| *** haigang has joined #openstack-lbaas | 07:10 | |
| *** _kiran_ has quit IRC | 07:17 | |
| *** vjay11 has quit IRC | 07:21 | |
| *** apuimedo has joined #openstack-lbaas | 07:21 | |
| *** bana_k has quit IRC | 07:25 | |
| *** ganeshna has quit IRC | 07:33 | |
| *** jschwarz has joined #openstack-lbaas | 07:35 | |
| *** numan has quit IRC | 07:52 | |
| *** nmagnezi has quit IRC | 07:56 | |
| *** haigang has quit IRC | 08:06 | |
| *** ganeshna has joined #openstack-lbaas | 08:09 | |
| *** nmagnezi has joined #openstack-lbaas | 08:11 | |
| openstackgerrit | Adam Harwell proposed openstack/octavia: Refactor BarbicanAuth to allow for configurable auth method https://review.openstack.org/216140 | 08:15 |
|---|---|---|
| *** numan has joined #openstack-lbaas | 08:20 | |
| *** numan has quit IRC | 09:27 | |
| *** numan has joined #openstack-lbaas | 09:49 | |
| *** ctracey has quit IRC | 10:55 | |
| *** ctracey has joined #openstack-lbaas | 10:58 | |
| *** xgerman has quit IRC | 10:58 | |
| *** xgerman has joined #openstack-lbaas | 11:00 | |
| *** evgenyf has quit IRC | 11:01 | |
| *** amotoki has quit IRC | 11:02 | |
| *** ganeshna has quit IRC | 11:12 | |
| *** evgenyf has joined #openstack-lbaas | 11:31 | |
| *** nmagnezi has quit IRC | 11:37 | |
| *** diogogmt has joined #openstack-lbaas | 11:41 | |
| *** numan has quit IRC | 12:02 | |
| *** diogogmt has quit IRC | 12:05 | |
| *** numan has joined #openstack-lbaas | 12:13 | |
| *** nmagnezi has joined #openstack-lbaas | 12:25 | |
| *** numan has quit IRC | 12:26 | |
| *** nmagnezi has quit IRC | 12:33 | |
| *** nmagnezi has joined #openstack-lbaas | 12:41 | |
| *** numan has joined #openstack-lbaas | 12:41 | |
| *** clev-away is now known as clev | 13:13 | |
| *** clev is now known as clev-away | 13:13 | |
| *** clev-away has quit IRC | 13:22 | |
| openstackgerrit | Evgeny Fedoruk proposed openstack/neutron-lbaas: L7 capability extension implementation for lbaas v2 https://review.openstack.org/148232 | 13:26 |
| *** clev has joined #openstack-lbaas | 13:37 | |
| *** amotoki has joined #openstack-lbaas | 14:00 | |
| *** Alex_Stef has quit IRC | 14:29 | |
| *** ajmiller has joined #openstack-lbaas | 14:33 | |
| *** alejandrito has joined #openstack-lbaas | 14:37 | |
| *** vivek-ebay has joined #openstack-lbaas | 14:48 | |
| *** diogogmt has joined #openstack-lbaas | 14:49 | |
| *** numan has quit IRC | 14:59 | |
| *** mlavalle has joined #openstack-lbaas | 15:00 | |
| *** numan has joined #openstack-lbaas | 15:02 | |
| *** vivek-ebay has quit IRC | 15:04 | |
| *** Aish has joined #openstack-lbaas | 15:10 | |
| *** amotoki has quit IRC | 15:16 | |
| *** ganeshna has joined #openstack-lbaas | 15:19 | |
| *** TrevorV has joined #openstack-lbaas | 15:30 | |
| *** vivek-ebay has joined #openstack-lbaas | 15:36 | |
| *** vivek-eb_ has joined #openstack-lbaas | 15:38 | |
| *** vivek-ebay has quit IRC | 15:40 | |
| *** numan has quit IRC | 15:47 | |
| *** evgenyf has quit IRC | 15:47 | |
| ptoohill | Can we get some eyes on this. Would like/need to get the rest of tls stuff merged across projects as sson as possible. Please give reviews if you can. Thank you | 15:50 |
| ptoohill | https://review.openstack.org/#/c/188703/ | 15:50 |
| *** Aish has quit IRC | 15:51 | |
| TrevorV | I'll hook you up with a super pedantic pair of eyes later ptoohill :P | 15:52 |
| ptoohill | sounds good | 15:53 |
| ajmiller | ptoohill I will look at it today | 15:53 |
| ptoohill | Thank you | 15:53 |
| *** alejandrito has quit IRC | 15:53 | |
| *** jorgem has joined #openstack-lbaas | 16:08 | |
| *** Aish has joined #openstack-lbaas | 16:09 | |
| *** jschwarz has quit IRC | 16:10 | |
| *** vivek-ebay has joined #openstack-lbaas | 16:10 | |
| *** vivek-eb_ has quit IRC | 16:13 | |
| *** ganeshna has quit IRC | 16:17 | |
| *** Aish has quit IRC | 16:20 | |
| ptoohill | madhu_k ping? | 16:21 |
| *** jorgem has quit IRC | 16:21 | |
| johnsom | ptoohill he isn't in the office yet | 16:21 |
| ptoohill | Thank you. Wanted to ask where the gate file was. I was going to look into getting barbican installed for gate checks and couldnt find it off hand | 16:22 |
| ptoohill | He told me once, but not in scroll back anymore. Ill find it. | 16:24 |
| johnsom | ptoohill This one: https://github.com/openstack/neutron-lbaas/blob/master/neutron_lbaas/tests/contrib/gate_hook.sh | 16:26 |
| ptoohill | ah, looks like it, Thank you! | 16:27 |
| ptoohill | Not sure why barbican need so much extra set up then every other service though :/ | 16:27 |
| *** vivek-ebay has quit IRC | 16:32 | |
| *** ganeshna has joined #openstack-lbaas | 16:33 | |
| *** clev is now known as clev-away | 16:34 | |
| *** sbalukoff has quit IRC | 16:39 | |
| *** bana_k has joined #openstack-lbaas | 16:44 | |
| *** vivek-ebay has joined #openstack-lbaas | 16:51 | |
| *** Alex_Stef has joined #openstack-lbaas | 16:52 | |
| *** Aish has joined #openstack-lbaas | 16:56 | |
| *** minwang2 has joined #openstack-lbaas | 16:58 | |
| johnsom | mestery Advice on escalating this infra update? https://review.openstack.org/#/c/211319/ | 17:02 |
| * mestery looks | 17:02 | |
| mestery | johnsom: Lets let dougwig handle that, no infra escalations without his support | 17:03 |
| johnsom | Ok. Sounds good. | 17:03 |
| *** abdelwas has joined #openstack-lbaas | 17:04 | |
| *** Aish has quit IRC | 17:12 | |
| *** nmagnezi has quit IRC | 17:13 | |
| abdelwas | Hello blogan TrevorV | 17:18 |
| TrevorV | abdelwas, hello! | 17:19 |
| abdelwas | I was updating my local master branch and it looks broken | 17:19 |
| *** madhu_ak has joined #openstack-lbaas | 17:20 | |
| abdelwas | This patch removed getter functions from allow_address_pair, that are still in use by other part of the code: https://review.openstack.org/#/c/209210/ | 17:20 |
| abdelwas | (get_subnet, and get_network) function | 17:21 |
| xgerman | yes, I thought bogan wanted to fix... | 17:21 |
| xgerman | blogan | 17:21 |
| abdelwas | Yeah, johnsom pointed that https://review.openstack.org/#/c/202336/ is fixing this | 17:22 |
| bana_k | I was also having issues with LB going in error state | 17:22 |
| johnsom | Actually doesn't look like it | 17:23 |
| abdelwas | Ok | 17:23 |
| bana_k | I am looking into it. rest driver was not able to configure it after ssh connection test. | 17:23 |
| bana_k | ssh connection was successful | 17:23 |
| abdelwas | This is probably in LB creation, | 17:24 |
| abdelwas | bana_k but other procedures that require computing network Deltas will fail | 17:24 |
| abdelwas | (even with SSH) | 17:25 |
| bana_k | hmm I see thats why may be LB is going into error state. the error msges were not that clear | 17:25 |
| TrevorV | My failover review shouldn't be fixing broken master... | 17:32 |
| TrevorV | How did tests not catch stuff like that? | 17:32 |
| TrevorV | Sorry I'm late, was grabbing lunch stuff. | 17:33 |
| TrevorV | abdelwas, do you have a dependent review that is now affected or is just master failing to create load balancers? | 17:33 |
| abdelwas | Yes | 17:33 |
| TrevorV | which yes? Dependent review? | 17:33 |
| abdelwas | My active/standby (https://review.openstack.org/#/c/206252/) depends on that | 17:34 |
| TrevorV | Ah gotcha. Well from what blogan was showing me the other day, there are a number of existing getter methods that should still be available, just in the class inherited by allowed_address_pairs | 17:35 |
| TrevorV | Did that patch not merge/ | 17:35 |
| TrevorV | merge? | 17:35 |
| TrevorV | *** | 17:35 |
| TrevorV | I had been confused about that as well. | 17:35 |
| TrevorV | It is entirely possible the method you used to use isn't a thing anymore, but something equivalent should be available. | 17:36 |
| *** Alex_Stef has quit IRC | 17:36 | |
| abdelwas | test_get_network, test_get_subnet, and test_get_port were removed from the tests | 17:36 |
| abdelwas | the method is get_subnet | 17:37 |
| abdelwas | its definition were removed from the master tree | 17:37 |
| TrevorV | I'm looking into it real quick | 17:37 |
| abdelwas | but it is still in use in allow_address_pair | 17:37 |
| TrevorV | My devstack VM is a little slow, so give me a minute :D | 17:37 |
| abdelwas | This is just master, not anything related to my review | 17:38 |
| abdelwas | OK | 17:38 |
| TrevorV | I understand :) | 17:38 |
| johnsom | pep8 should have caught it, but I think since it's a self. it missed it | 17:38 |
| abdelwas | I have a reasoning for that | 17:40 |
| TrevorV | abdelwas, as I had said, if you are looking at allowed_address_pairs.py | 17:40 |
| abdelwas | get_subnet implementation was removed from allow_address_pair but not from the base class | 17:40 |
| TrevorV | The class definition has "neutron_bas.BaseNeutronDriver" as an inherited class. | 17:40 |
| abdelwas | So network deltas tests, think that the method is just there | 17:40 |
| abdelwas | but it is actually not | 17:41 |
| TrevorV | When you use AllowedAddressPairs it should be available through inheritance. | 17:41 |
| johnsom | Yeah, it is inherited in. abdelwas what is the error you are seeing? | 17:42 |
| abdelwas | The tox error I got after rebasing my patch | 17:42 |
| TrevorV | Did you have merge conflicts? | 17:42 |
| abdelwas | I used to have | 17:43 |
| TrevorV | Its possible you merged improperly. I had that issue the other day and broke johnsom :( | 17:43 |
| *** bharathm has joined #openstack-lbaas | 17:43 | |
| TrevorV | blogan helped me out. | 17:43 |
| TrevorV | I don't think blogan will be online today (he's got a lot of time to take off from Rax) but I can text him to see if he can help shed some light | 17:44 |
| abdelwas | the get_subnet method can not be left without implementation, as network_tasks.py uses it to CalculateDelta (around line 87) | 17:44 |
| abdelwas | Am I missing something ? | 17:44 |
| TrevorV | Its not unimplemented. | 17:44 |
| TrevorV | Its implemented in the inherited class. | 17:44 |
| *** ganeshna has quit IRC | 17:44 | |
| johnsom | TrevorV before you bother blogan let us have a look internally | 17:45 |
| TrevorV | Alright, sure thing johnsom | 17:45 |
| TrevorV | Just let me know if you want me to get him | 17:45 |
| *** clev-away is now known as clev | 17:46 | |
| rm_work | I think blogan said he was in office today? | 17:52 |
| rm_work | TrevorV: ^^ is that incorrect? | 17:52 |
| TrevorV | rm_work, he is | 17:52 |
| TrevorV | I texted him, he'll be back after lunch | 17:52 |
| rm_work | ah | 17:52 |
| TrevorV | I wasn't sure because I wasn't aware of the time :D | 17:52 |
| rm_work | heading in myself shortly | 17:56 |
| *** rm_work is now known as rm_work|away | 18:00 | |
| *** sbalukoff has joined #openstack-lbaas | 18:00 | |
| blogan | I'm here now | 18:02 |
| blogan | abdelwas, johnsom, TrevorV: yeah the get_subnet, get_port, get_network are now in a Neutron base class | 18:02 |
| blogan | which allowed address pairs inherits from | 18:02 |
| blogan | and the tests for those got moved into an equivalent test module | 18:02 |
| abdelwas | Yeah, got that part | 18:02 |
| blogan | abdelwas: the errors you are seeing are in your review? | 18:03 |
| abdelwas | It seems TrevorV was right about merge problem in my review | 18:03 |
| abdelwas | Yes it is | 18:03 |
| blogan | did you push up your mege fixes? | 18:03 |
| blogan | merge | 18:03 |
| abdelwas | I am working on the merge fixes | 18:03 |
| blogan | abdelwas: i can take a look if you would like | 18:03 |
| abdelwas | Don't bother it seems that I get an outdated version of base.py for some reason (it only had abstract methods) | 18:04 |
| abdelwas | I will fix this and test again and let you know if I am still stuck with it | 18:04 |
| abdelwas | Thank you very much blogan TrevorV. Appreciate it | 18:05 |
| TrevorV | Not a problem, like I said I did the same thing last week ha ha | 18:05 |
| blogan | abdelwas: there are actually 2 base.py's, one under the network package which is the abstract class for all network drivers, the other is the base.py under the network/neutron package, which is the partially implemented neutron base network driver | 18:06 |
| blogan | abdelwas: might have just had the wrong import statement | 18:06 |
| abdelwas | Yeah | 18:09 |
| abdelwas | That second file was not there in my local branch for a mysterious reason. I should have rebase daily for the last month :) | 18:10 |
| abdelwas | blogan thanks | 18:11 |
| bana_k | so this the log of o-cw | 18:11 |
| bana_k | https://gist.github.com/banveerad/d904ea7c4337e113c1c8 | 18:11 |
| bana_k | it doesn't show any error but LB goes into error state | 18:11 |
| bana_k | n if I try to delete that I am not able to delete it | 18:12 |
| *** jorgem has joined #openstack-lbaas | 18:13 | |
| blogan | bana_k: if you do a nova list under the account that is supposed to provision these, does it show the instance in ERROR? | 18:18 |
| blogan | bana_k: wait that git is saying the flow was successful | 18:19 |
| blogan | bana_k: and the LB got updated to ACTIVE | 18:19 |
| blogan | bana_k: look at the q-svc logs, i'm willing to bet it took too long to go active, so the octavia driver tossed it into ERROR bc ittook what looks like 6 minutes to go active | 18:20 |
| blogan | bana_k: over 6 mins | 18:20 |
| blogan | bana_k: and i believe the default timeout is 100 seconds in the octavia driver, some arbitrary number i pulled out of my ass, we might want to increase that default timeout to higher if this is a typical case | 18:21 |
| openstackgerrit | Sherif Abdelwahab proposed openstack/octavia: Amphora Flows and Service Drivers for Active Standby https://review.openstack.org/206252 | 18:22 |
| *** jorgem has quit IRC | 18:33 | |
| bana_k | nova instance is fine. I am able ssh to it. | 18:36 |
| *** johnsom has quit IRC | 18:36 | |
| *** johnsom_ has joined #openstack-lbaas | 18:36 | |
| bana_k | blogan: I saw the q-svc it doesnt have any error msgs. | 18:37 |
| openstackgerrit | Phillip Toohill proposed openstack/neutron-lbaas: Adding barbican to gate hook https://review.openstack.org/216374 | 18:38 |
| blogan | bana_k: it wouldn't be an error message, it'd be a debug message, though it probably should be an error message now that hindsight has kicked in | 18:45 |
| blogan | bana_k: grep for this string "Timeout has expired for load balancer" | 18:46 |
| blogan | you might find it | 18:46 |
| bana_k | blogan : Oh ok. q-svc has so many log msgs I might have missed it . | 18:46 |
| blogan | bana_k: yeah there's a lot of noise for sure | 18:47 |
| bana_k | blogan : yes ll do | 18:47 |
| *** bharathm has quit IRC | 18:48 | |
| bana_k | blogan: Yes i see that in the q-svc | 18:48 |
| *** rm_work|away is now known as rm_work | 18:48 | |
| *** bharathm has joined #openstack-lbaas | 18:48 | |
| blogan | bana_k: then yeah thats the problem, there si a config option you can set to increase the timeout to whatever you want | 18:49 |
| bana_k | I looked for it request_poll_interval/cfg.CONF.octavia.request_poll_timeout | 18:49 |
| blogan | bana_k: yeah you can add an [octavia] section in your neutron_lbaas.conf and put that option in | 18:50 |
| bana_k | blogan: ok ll do that | 18:50 |
| blogan | really it can go into any config you want, as long as its passed in when you start neutron-server (neutron_lbaas.conf gets automatically loaded though bc of magic) | 18:50 |
| bana_k | blogan : I think i ll put it in neutron.conf as of now | 18:51 |
| bana_k | blogan : thanks | 18:51 |
| blogan | bana_k: okay, wherever your heart desires | 18:51 |
| bana_k | but delete is not working when it goes into error state , I think I should look into that first | 18:52 |
| bana_k | before fixing this one. | 18:52 |
| blogan | bana_k: i just put it in neutron_lbaas.conf bc it is a neutron_lbaas option and its automatically loaded | 18:52 |
| blogan | bana_k: yeah i figured that would work | 18:52 |
| bana_k | oh ok . | 18:52 |
| bana_k | will do that | 18:52 |
| blogan | bana_k: if you could troubleshoot why that's not working when LB is in error state that'd be great, then again i could probably just reproduce it too | 18:53 |
| bana_k | sure ll do that | 18:53 |
| *** minwang2 has quit IRC | 18:54 | |
| *** numan has joined #openstack-lbaas | 18:55 | |
| *** numan has quit IRC | 18:56 | |
| *** numan has joined #openstack-lbaas | 18:57 | |
| rm_work | ptoohill: https://review.openstack.org/#/c/167885/ | 19:06 |
| ptoohill | Would be quite nice to get that in. Thank you for pointing this out | 19:06 |
| *** numan is now known as numans | 19:10 | |
| *** fnaval has joined #openstack-lbaas | 19:12 | |
| openstackgerrit | Phillip Toohill proposed openstack/neutron-lbaas: Adding barbican to gate hook https://review.openstack.org/216374 | 19:13 |
| blogan | bana_k: tested it out and it did delete the lb when it was an ERROR state, but i just updated the db to ERROR, didnt make the code do ti | 19:15 |
| bana_k | good news is that delete LB is working fine with latest stack | 19:15 |
| bana_k | yes I was about to say that | 19:15 |
| bana_k | n its now deleting the VM also | 19:16 |
| blogan | bana_k: good news | 19:21 |
| rm_work | ptoohill: https://review.openstack.org/#/c/216140/ | 19:29 |
| blogan | question for anyone paying attention | 19:40 |
| blogan | neturon-lbaas currently has default_tls_container_id for tls certificates | 19:41 |
| blogan | when using barbican, the user would supply the uri to their container, which makes the _id field a bit inappropriate | 19:42 |
| *** minwang2 has joined #openstack-lbaas | 19:42 | |
| blogan | so i want to change that to ref before v2 gets out of experimental mode | 19:42 |
| blogan | so default_tls_container_ref instead | 19:43 |
| blogan | i'd also argue dropping the default_ as well, which then leads to dropping the default_ off pool_id, but that could be biting off more than we can chew | 19:43 |
| blogan | anyone of thoughts? | 19:43 |
| blogan | johnsom_, xgerman, dougwig, ajmiller, ptoohill, rm_work, rm_you: ^^ | 19:44 |
| ptoohill | default pool makes sense, when we have more pools. default tls just doesnt as we will have the one certificate and then sni. If we plan on doing something different we can change it then | 19:45 |
| ptoohill | but pools already has plans | 19:45 |
| *** numans has quit IRC | 19:46 | |
| ajmiller | Are there any other precidents for "id" not being a uuid-derived thing? It seems like that is a pretty well established convention, breaking it seems like a bad idea, so if we change it, sooner is better. | 19:49 |
| sbalukoff | I like keeping it named 'default_pool' since that will be less confusing when people are using L7 switching. | 19:53 |
| sbalukoff | Also: +1 to ajmiller | 19:53 |
| sbalukoff | And I would agree that it probably makes sense to change it to a default_tls_container_ref instead of _id. | 19:54 |
| xgerman | URL is goodness | 19:55 |
| xgerman | ref is fine with me | 19:55 |
| blogan | okay yeah default_pool_id can stay, i'm convinced on that | 19:59 |
| blogan | default_ on the tls_container thing though? | 19:59 |
| blogan | and ajmiller to answer your qeustion, i do not know of any others | 20:00 |
| blogan | also, rm_work/rm_you brought up a point that "container" is a barbican term, and since its not supposed to be only barbican, it would make sense to change that to "bundle" | 20:05 |
| blogan | any thoughts on that? | 20:05 |
| xgerman | nah, just keep container | 20:06 |
| ptoohill | I still think container is a fairly generic term and not sure why everyone is against generic terms, but do like the way bundle sounds | 20:06 |
| xgerman | I think as long as we don’t call it amphora we will be fine | 20:06 |
| blogan | alright, going to call it server | 20:07 |
| blogan | tls_server_ref | 20:07 |
| ptoohill | contenedor | 20:07 |
| xgerman | yeah, I like container as well - it’s being used a lot those days | 20:08 |
| blogan | xgerman: lol for different reasons :) (docker) | 20:08 |
| blogan | xgerman: but really cahnging id to ref is the main issue i have | 20:09 |
| blogan | so keepign container i'd be fine with | 20:09 |
| xgerman | well, people will start looking for the docker instance | 20:09 |
| rm_work | heh | 20:09 |
| blogan | but since i was making one change, figured i'd see if anyone was open to other ones in the same patch | 20:09 |
| rm_work | i suggested bundle, so i might be biased, but i think it makes a lot of sense | 20:09 |
| rm_work | but wouldn't object to keeping it as container | 20:10 |
| rm_work | as long as id->reg | 20:10 |
| rm_work | * id->ref | 20:10 |
| johnsom_ | Thing Which Holds the Bits TWHB | 20:11 |
| xgerman | arca | 20:12 |
| xgerman | latin fox box? | 20:12 |
| johnsom_ | There you go | 20:13 |
| xgerman | yeah, in the amphora spirit... | 20:13 |
| ptoohill | tls_arca_ref | 20:16 |
| blogan | i woudl have voted for arca | 20:16 |
| johnsom_ | Rolls off the tongue | 20:16 |
| blogan | well over amphora | 20:16 |
| blogan | still new people coming in wouldn't know what the hell it is | 20:17 |
| rm_work | oooo | 20:22 |
| *** vivek-ebay has quit IRC | 20:25 | |
| *** TrevorV has quit IRC | 20:32 | |
| *** jerrygb has joined #openstack-lbaas | 20:33 | |
| *** jerrygb is now known as test0rz | 20:34 | |
| *** test0rz is now known as asd112z | 20:34 | |
| *** woodster_ has joined #openstack-lbaas | 20:37 | |
| sbalukoff | As long as we're obfuscating it from newbs, I'm happy. | 20:40 |
| sbalukoff | ;) | 20:40 |
| johnsom_ | And those that leave the project for a while.... | 20:40 |
| sbalukoff | So I'm still in favor of keeping the 'default_' in front of the tls container ref thingy. This is because if someone is using SNI and the cert they ask for isn't found, the web service must default to using something. | 20:41 |
| sbalukoff | johnsom: Aaw! You did miss me! | 20:41 |
| sbalukoff | Like a hole in the head, right? | 20:41 |
| *** johnsom_ is now known as johnsom | 20:44 | |
| *** vivek-ebay has joined #openstack-lbaas | 20:48 | |
| openstackgerrit | min wang proposed openstack/octavia: Add cert tracking and rotating in Housekeeping https://review.openstack.org/215359 | 20:50 |
| openstackgerrit | Phillip Toohill proposed openstack/neutron-lbaas: Adding barbican to gate hook https://review.openstack.org/216374 | 20:50 |
| rm_work | ptoohill: so if i get this fixed will it simplify your gate stuff? | 20:50 |
| ptoohill | a tad | 20:51 |
| ptoohill | rm_work: | 20:51 |
| rm_work | it's pretty trivial on the barbican side, really the complication is going to be on the infra side | 20:51 |
| ptoohill | 3 lines or so | 20:51 |
| rm_work | heh | 20:51 |
| rm_work | woo | 20:51 |
| ptoohill | :P | 20:51 |
| blogan | sbalukoff: but i guess is that not implied by sni being on the listener as well? | 20:53 |
| sbalukoff | Eh... it's about the same thing with 'default_pool' in the presence / absence of L7 policies referencing other pools. | 20:54 |
| sbalukoff | The 'default' I guess is a reminder that it might not actually be the one used if overridden by SNI / L7 policy. | 20:55 |
| blogan | sbalukoff: meaning sni and tls_container are on the listener, if sni is being used the default will be tls_container_ref, but tls_container_ref is also being used to terminate, which seems odd to have default_ in front | 20:55 |
| blogan | sbalukoff: kind of the same, but sni is just used to route traffic, not to terminate (i could have this wrong too), which means the default_tls_container_ref will ahve dual roles, but terminating traffic is probably the more common role | 20:56 |
| blogan | and it being on the same object as sni gives it implied defaultness if people already know what sni is | 20:57 |
| * blogan really deosn't like a long variable | 20:57 | |
| sbalukoff | No, SNI refers to the technology of being able to choose which cert is used to terminate a connection. So yes, in that sense SNI does terminate the traffic. | 20:57 |
| sbalukoff | Oh I agree that the variable name is probably too long. | 20:57 |
| sbalukoff | Let's just call it george. | 20:57 |
| sbalukoff | It doesn't make a huge difference either way. | 20:58 |
| blogan | ok misunderstanding of sni on my part (i'm dumb) | 20:59 |
| blogan | welp then | 20:59 |
| sbalukoff | Er... did you mean to say that L7 is used just to route traffic (not terminate it)? Because that is correct. | 20:59 |
| blogan | no i meant sni, and i was wrong | 20:59 |
| sbalukoff | Aah, ok. | 20:59 |
| sbalukoff | Ok, meeting time for me... BBIAB. | 20:59 |
| *** vivek-ebay has quit IRC | 21:02 | |
| blogan | okay gonna keep default_ since we'll be keeping default_ on pool | 21:02 |
| blogan | major thing is id to ref anyway | 21:02 |
| xgerman | ok, are we good with that ;-) | 21:28 |
| bharathm | In the command "neutron lbaas-member-create --subnet sub --address 10.0.0.5 --protocol- port 80 pool",, I believe the "sub" is the subnet that the member belongs to. Is there any reason for explicitly providing it rather than automatically picking it up from neutron db ? | 21:34 |
| bharathm | Or the subnet here just serves the purpose of routing to reach the backend ? | 21:38 |
| *** madhu_ak_ has joined #openstack-lbaas | 21:46 | |
| *** madhu_ak has quit IRC | 21:50 | |
| *** madhu_ak_ is now known as madhu_ak | 22:01 | |
| *** vivek-ebay has joined #openstack-lbaas | 22:02 | |
| *** vivek-ebay has quit IRC | 22:07 | |
| blogan | bharathm: how would one pick up the subnet from the neutron db? | 22:10 |
| bharathm | blogan: Sorry.. I meant if the subnet could be somehow retrieved from the neutron port information because the IP is assigned during the port creation from a specific subnet id | 22:11 |
| blogan | bharathm: assuming the ip is from a neutron server | 22:13 |
| blogan | i mean nova server, or some kind of device taht creates a neutron port | 22:13 |
| blogan | but im sure some smart logic could be put in to determine whether it is or isnt | 22:13 |
| *** apuimedo_ has joined #openstack-lbaas | 22:13 | |
| blogan | but woudl still could run into an issue where the IP is meant to be on the public internet but there's a subnet with a port on it with the same IP | 22:14 |
| blogan | but like i said some smart logic could be used to handle that. still think there would be some assumptions made though | 22:14 |
| bharathm | Understood.. I have been working my way around to figure if this subnet is somehow/somewhere used by the haproxy on the data plane.. | 22:16 |
| blogan | bharathm: well in the namespace driver it is not, but in Octavia it is | 22:17 |
| blogan | bharathm: at least I dont remember that it is because its a requirement to have all the subnets connected via an L3 router | 22:18 |
| blogan | well the vip subnet connected to the member subnets via L3 router | 22:18 |
| blogan | relic of the v1 namespace drvier that I carried over for lack of motivation to make it better :) | 22:19 |
| bharathm | He he.. Gotcha... I need to dive into the network implementation in Octavia.. Mainly the plug_network part.. | 22:20 |
| bharathm | That's my initial understanding too. As long as there's a route between vip subnet and member subnet, LB should work and I didn't see any particular requirement of passing the subnet information for the members while adding them to pool.. | 22:22 |
| blogan | bharathm: yeah that will connect the amphora (the thing running haproxy) to the subnet for each member | 22:22 |
| bharathm | Got it now.. Thanks blogan | 22:22 |
| blogan | bharathm: np, its really up to the backend to use it or not | 22:23 |
| blogan | err driver/backend | 22:23 |
| bharathm | :-) makes sense | 22:24 |
| blogan | however that does bring up an interesting point, if one driver requires all the subnets to be connected to the router and others do not, thats kind of leaking details about the driver and making the workflow inconsistent for different drivers | 22:24 |
| *** asd112z has quit IRC | 22:25 | |
| ptoohill | So, I cant seem to clone within the hook according to errors(maybe someone has other information). Adam is working, or was, on a patch that will solve this. | 22:28 |
| ptoohill | so, no tls ever | 22:29 |
| ptoohill | ever, ever | 22:30 |
| *** sbalukoff has quit IRC | 22:34 | |
| bana_k | blogan: adjusting those values worked fine. But q-svc is not able to read those values from neutron_lbaas.conf so had put them in neutron.conf | 22:36 |
| blogan | bana_k: hmm, perhaps that has changed | 22:40 |
| blogan | bana_k: the magic loading | 22:40 |
| blogan | well no, bc the service providers wouldn't be loading | 22:40 |
| bana_k | oh ok. | 22:42 |
| bana_k | create n delete listener taking a lot of time on my virtualbox based VM devstack setup | 22:47 |
| bana_k | has anyone had the same experience ? | 22:47 |
| blogan | bana_k: that shouldn't take too long, its really just doing an update of the haproxy config, well it might also be updating neutron security groups as well | 22:47 |
| blogan | bana_k: can you look at the o-cw logs and see if that gives any clues? | 22:48 |
| bana_k | k ll do that. | 22:49 |
| *** apuimedo_ has quit IRC | 22:51 | |
| bana_k | started create listener at 17:49:14.780 and LB went from active to update_pending status and went back to active state at 17:51:45.086. | 22:54 |
| *** mlavalle has quit IRC | 22:55 | |
| bana_k | blogan: in o-cw everything happens fast. But octavia driver 2015-08-24 17:51:45.086 DEBUG neutron_lbaas.drivers.octavia.driver [-] Octavia reports load balancer 775dbdd5-e220-46ca-b138-e46fdb3019b3 has provisioning status of ACTIVE from (pid=15454) thread_op /opt/stack/neutron-lbaas/neutron_lbaas/drivers/octavia/driver.py:61 | 22:55 |
| blogan | bana_k: so it finsihed | 22:56 |
| bana_k | oh I think its because of the updated poll interval and timeout | 22:56 |
| bana_k | values | 22:56 |
| blogan | but took 2 minutes | 22:56 |
| blogan | bana_k: yeah dont update the interval | 22:56 |
| bana_k | oh ok. got it | 22:56 |
| blogan | bana_k: great | 22:57 |
| *** diogogmt has quit IRC | 23:03 | |
| *** Aish has joined #openstack-lbaas | 23:05 | |
| bana_k | blogan: Is this is what abdelwas was discussing in the morning ? lwas> This is probably in LB creation, | 23:14 |
| bana_k | <abdelwas> bana_k but other procedures that require computing network Deltas will fail | 23:14 |
| bana_k | <abdelwas> (even with SSH) | 23:14 |
| bana_k | <bana_k> hmm I see thats why may be LB is going into error state. the error msges were not that clear | 23:14 |
| bana_k | <TrevorV> My failover review shouldn't be fixing broken master... | 23:14 |
| bana_k | <TrevorV> How did tests not catch stuff like that? | 23:14 |
| bana_k | <TrevorV> Sorry I'm late, was grabbing lunch stuff. | 23:14 |
| bana_k | <TrevorV> abdelwas, do you have a dependent review that is now affected or is just master failing to create load balancers? | 23:14 |
| bana_k | <abdelwas> Yes | 23:14 |
| bana_k | <TrevorV> which yes? Dependent review? | 23:14 |
| bana_k | <abdelwas> My active/standby (https://review.openstack.org/#/c/206252/) depends on that | 23:15 |
| bana_k | <TrevorV> Ah gotcha. Well from what blogan was showing me the other day, there are a number of existing getter methods that should still be available, just in the class inherited by allowed_address_pairs | 23:15 |
| bana_k | <TrevorV> Did that patch not merge/ | 23:15 |
| bana_k | <TrevorV> merge? | 23:15 |
| bana_k | <TrevorV> *** | 23:15 |
| bana_k | <TrevorV> I had been confused about that as well. | 23:15 |
| bana_k | <TrevorV> It is entirely possible the method you used to use isn't a thing anymore, but something equivalent should be available. | 23:15 |
| bana_k | * Alex_Stef has quit (Ping timeout: 244 seconds) | 23:15 |
| bana_k | <abdelwas> test_get_network, test_get_subnet, and test_get_port were removed from the tests | 23:15 |
| bana_k | <abdelwas> the method is get_subnet | 23:15 |
| bana_k | <abdelwas> its definition were removed from the master tree | 23:15 |
| bana_k | <TrevorV> I'm looking into it real quick | 23:15 |
| bana_k | <abdelwas> but it is still in use in allow_address_pair | 23:15 |
| bana_k | <TrevorV> My devstack VM is a little slow, so give me a minute :D | 23:15 |
| bana_k | <abdelwas> This is just master, not anything related to my review | 23:15 |
| bana_k | <abdelwas> OK | 23:15 |
| bana_k | <TrevorV> I understand :) | 23:15 |
| bana_k | <johnsom> pep8 should have caught it, but I think since it's a self. it missed it | 23:15 |
| bana_k | <abdelwas> I have a reasoning for that | 23:15 |
| bana_k | <TrevorV> abdelwas, as I had said, if you are looking at allowed_address_pairs.py | 23:15 |
| bana_k | <abdelwas> get_subnet implementation was removed from allow_address_pair but not from the base class | 23:15 |
| bana_k | <TrevorV> The class definition has "neutron_bas.BaseNeutronDriver" as an inherited class. | 23:15 |
| bana_k | <abdelwas> So network de | 23:15 |
| bana_k | oh shit sorry | 23:15 |
| blogan | ah bana_k you flooded the channel! | 23:15 |
| bana_k | I mean this https://gist.github.com/banveerad/c61872cbc13bd9655297 | 23:15 |
| blogan | lol | 23:15 |
| bana_k | mybad :D | 23:16 |
| abdelwas | My name is mentioned like never before :D | 23:16 |
| blogan | lol | 23:16 |
| bana_k | sorry again ppl ! | 23:17 |
| bana_k | please take a look at this log | 23:17 |
| blogan | bana_k: you've done something no one has ever done | 23:17 |
| bana_k | lol | 23:17 |
| bana_k | I am glad :P | 23:18 |
| blogan | bana_k: are you testing out the abdelwas's active passive reveiw and got this log? | 23:18 |
| bana_k | no just add members | 23:18 |
| bana_k | and load balancer got stuck in update pending state | 23:19 |
| bana_k | n now i cant even delete that | 23:19 |
| *** vivek-ebay has joined #openstack-lbaas | 23:19 | |
| blogan | hmm | 23:19 |
| blogan | so straight off master? | 23:19 |
| bana_k | yes sir | 23:19 |
| bana_k | this time let me double check it | 23:20 |
| bana_k | double checked, n yes | 23:20 |
| blogan | bana_k: whats the commit you're on in octavia? | 23:20 |
| bana_k | blogan: dfeea9ba1351c70ddb199d404d45a760a1538c0a | 23:21 |
| blogan | bana_k: ah you're using the rest driver | 23:21 |
| bana_k | yes | 23:21 |
| blogan | bana_k: that one i dont know as much about as johnsom and xgerman | 23:21 |
| johnsom | You rang? | 23:22 |
| blogan | johnsom: https://gist.github.com/banveerad/c61872cbc13bd9655297 | 23:22 |
| blogan | looks like an error happening in the rest driver on member create | 23:22 |
| xgerman | hi | 23:22 |
| blogan | johnsom: im sure my code change messed this up too | 23:22 |
| blogan | johnsom: but thats just a guess based in no factual data | 23:23 |
| xgerman | blogan we will take your code change privileges away every time we are tow weeks from code freeze | 23:23 |
| blogan | xgerman: hey that review went up before 2 weeks :) | 23:23 |
| *** vivek-ebay has quit IRC | 23:23 | |
| johnsom | Yeah, and xgerman +2'd it | 23:24 |
| blogan | boom! | 23:24 |
| johnsom | You will note, I did not however.... | 23:24 |
| xgerman | I know I am guilty | 23:24 |
| blogan | lol well technically, it worked with everything in master, it has only messed up reviews unmerged | 23:24 |
| blogan | which i knew would happen | 23:24 |
| blogan | and this is failing in the amp driver which i dont believe that review touch, but its possible that amp driver is getting data it doesn't expect | 23:25 |
| bana_k | and I think at least we should get the LB back to error state or something else its a deadlock | 23:26 |
| johnsom | It's a pretty straight forward call. I suspect something was not happy on the amp side | 23:26 |
| bana_k | we cant do anything to the LB | 23:26 |
| blogan | well it shoulda broke the ssh driver too if it was a parameter contract change, so it may not be that cahnge that did this | 23:26 |
| johnsom | The status stuff is in the patchset I'm working on now. It's done, just needs unit test. | 23:27 |
| blogan | bana_k: you can update the status to ACTIVE with a db call and then you can do something with it :) | 23:27 |
| johnsom | bana_k is the amp still running, i.e. visible in nova list? | 23:27 |
| bana_k | hehehe sure ll do | 23:27 |
| bana_k | yes | 23:27 |
| bana_k | its still running | 23:27 |
| johnsom | Cool, ssh in "sudo ssh -i /etc/octavia/.ssh/octavia_ssh_key ubuntu@<mgmt ip>" | 23:28 |
| johnsom | Then look at the agent log /var/log/upstart/amphora-agent.log | 23:29 |
| johnsom | There might be some hints here | 23:29 |
| bana_k | oh ok sshing | 23:29 |
| johnsom | REST on devstack works for the most part for me. I've been doing a bunch of testing with it for the health manager stuff | 23:30 |
| openstackgerrit | min wang proposed openstack/octavia: Add cert tracking and rotating in Housekeeping https://review.openstack.org/215359 | 23:32 |
| johnsom | My guess is the interface didn't get plugged into the amp. There is likely an error farther up the log | 23:32 |
| bana_k | /var/log/upstart/amphora-agent.log is empty :/ | 23:33 |
| johnsom | Something bad happened earlier in the process then, or this amp is running ssh driver. | 23:35 |
| johnsom | I don't think I have ever seen it empty with the rest driver. | 23:35 |
| johnsom | Did you create the lb with ssh driver, then switch to rest driver? | 23:35 |
| *** vivek-ebay has joined #openstack-lbaas | 23:36 | |
| bana_k | johnsom : no | 23:37 |
| *** kev0 has joined #openstack-lbaas | 23:37 | |
| bana_k | johnsom : https://gist.github.com/banveerad/e1df110b6dc6337b16e3 | 23:37 |
| johnsom | Hmmm | 23:37 |
| bana_k | my octauva conf | 23:38 |
| *** woodster_ has quit IRC | 23:38 | |
| johnsom | That looks ok | 23:38 |
| *** woodster_ has joined #openstack-lbaas | 23:38 | |
| bana_k | and this works too curl -k --cert /etc/octavia/certs/client.pem https://192.168.0.8:8443/0.5/info | python -m json.tool | 23:39 |
| johnsom | bana_k oh, is the log file named octavia-agent.log? | 23:40 |
| *** diogogmt has joined #openstack-lbaas | 23:41 | |
| bana_k | lol, yes | 23:41 |
| bana_k | no error msgs on that | 23:41 |
| johnsom | sorry, I renamed it to clean up some mis-matches in this new patchset | 23:41 |
| johnsom | ok | 23:41 |
| johnsom | sudo ifconfig | 23:42 |
| johnsom | any .0 interfaces? | 23:42 |
| bana_k | yes eth1.0 | 23:42 |
| openstackgerrit | Brandon Logan proposed openstack/neutron-lbaas: Change tls and sni listener API attributes to ref https://review.openstack.org/216465 | 23:44 |
| johnsom | is it up or down? | 23:45 |
| *** ajmiller has quit IRC | 23:50 | |
| bana_k | ethtool says link detected yes | 23:50 |
| bana_k | n has ip address | 23:50 |
| johnsom | Can you give the output of "sudo ifconfig eth1.0" | 23:50 |
| bana_k | says up | 23:50 |
| johnsom | Ok. So, that is why you got the not found. | 23:51 |
| johnsom | Did you add a member to the same subnet as the VIP? | 23:51 |
| johnsom | (listener) | 23:51 |
| bana_k | https://gist.github.com/banveerad/d21ecc6c8c6f50506dc5 | 23:51 |
| bana_k | yes | 23:52 |
| bana_k | VIP is 10.0.0.13 | 23:52 |
| bana_k | n mems IP : 10.0.0.18 and 17 | 23:52 |
| johnsom | Yeah, this is blogan's bug. abdelwas reported this. The calculate delta is broken. Was a bug filed for it? | 23:53 |
| bana_k | I dont think so | 23:54 |
| johnsom | It's trying to bring up an interface on the amphora that is already up. | 23:54 |
| bana_k | should I file the bug? | 23:57 |
| *** apuimedo has quit IRC | 23:57 | |
| johnsom | Filing it now | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!