Friday, 2017-04-21

« Thursday, 2017-04-20 Index Saturday, 2017-04-22 »
*** JudeC has joined #openstack-lbaas00:17
openstackgerritMerged openstack/octavia master: Removing dependency on eventlet and oslo.service  https://review.openstack.org/45487300:22
JudeCrm_work: You were right I think00:23
JudeCPatch in Merge Conflict00:23
*** blogan_ has joined #openstack-lbaas00:26
*** blogan_ has quit IRC00:27
*** blogan_ has joined #openstack-lbaas00:28
*** blogan_ has quit IRC00:29
*** blogan has quit IRC00:29
*** blogan_ has joined #openstack-lbaas00:35
*** blogan_ has quit IRC00:36
*** blogan_ has joined #openstack-lbaas00:37
*** blogan_ has quit IRC00:38
*** blogan_ has joined #openstack-lbaas00:39
*** jerrygb has joined #openstack-lbaas00:45
*** jerrygb has quit IRC00:52
*** aojea has joined #openstack-lbaas01:02
*** jerrygb has joined #openstack-lbaas01:04
*** aojea has quit IRC01:06
*** bzhao_ has quit IRC01:09
*** JudeC has quit IRC01:30
*** JudeC has joined #openstack-lbaas01:52
*** JudeC has quit IRC02:09
*** fnaval has quit IRC02:15
*** gongysh has joined #openstack-lbaas02:22
*** amotoki has joined #openstack-lbaas02:58
*** JudeC has joined #openstack-lbaas02:59
*** amotoki has quit IRC03:03
*** sanfern has joined #openstack-lbaas03:05
*** reedip has quit IRC03:08
*** sanfern has quit IRC03:13
*** sanfern has joined #openstack-lbaas03:13
*** JudeC has quit IRC03:18
xgermanAs expected last place http://civs.cs.cornell.edu/cgi-bin/results.pl?id=E_072c4cd7ff0673b503:26
*** JudeC has joined #openstack-lbaas03:26
*** JudeC has quit IRC03:27
*** amotoki has joined #openstack-lbaas04:09
*** gongysh has quit IRC04:14
*** JudeC has joined #openstack-lbaas04:34
rm_workaugh I forgot to vote T_T04:41
rm_workgotta do more publicity I think :P04:41
*** gongysh has joined #openstack-lbaas04:43
*** cpuga has joined #openstack-lbaas04:44
openstackgerritJude Cross proposed openstack/octavia master: Fix health_manager to exit without waiting  https://review.openstack.org/45642004:46
JudeCHope I did that right :)04:46
rm_worknope :P04:46
rm_workyou added eventlet back in T_T04:47
rm_workbut thanks for helping with my review percentage :P04:47
openstackgerritJude Cross proposed openstack/octavia master: Fix health_manager to exit without waiting  https://review.openstack.org/45642004:49
rm_workeandersson: your stackalytics affiliation isn't filled, you should do that ^_^04:49
*** armax has quit IRC04:50
*** gongysh has quit IRC05:09
*** blogan_ has quit IRC05:17
*** reedip has joined #openstack-lbaas05:20
*** gongysh has joined #openstack-lbaas05:59
*** vlaza has joined #openstack-lbaas06:09
*** rcernin has joined #openstack-lbaas06:15
*** pcaruana has joined #openstack-lbaas06:18
*** voelzmo has joined #openstack-lbaas06:34
*** gcheresh_ has joined #openstack-lbaas06:53
*** chlong has quit IRC06:56
*** tesseract has joined #openstack-lbaas07:11
*** aojea has joined #openstack-lbaas07:17
*** aojea_ has joined #openstack-lbaas07:18
*** gongysh has quit IRC07:18
*** aojea has quit IRC07:21
openstackgerritcheng proposed openstack/octavia master: Add proxy-protocol-header opts for listening insert-header  https://review.openstack.org/45875507:51
*** JudeC has quit IRC07:52
*** cpuga has quit IRC08:02
*** gongysh has joined #openstack-lbaas08:09
*** jerrygb_ has joined #openstack-lbaas08:11
*** jerrygb has quit IRC08:11
openstackgerritZhaoBo proposed openstack/octavia master: [WIP]Extend api to accept qos_policy_id  https://review.openstack.org/45830808:13
*** rcernin has quit IRC08:13
*** jerrygb_ has quit IRC08:21
*** jerrygb has joined #openstack-lbaas08:22
*** rcernin has joined #openstack-lbaas08:27
*** gongysh has quit IRC08:51
*** amotoki has quit IRC08:57
*** amotoki has joined #openstack-lbaas09:08
*** amotoki has quit IRC09:30
*** amotoki has joined #openstack-lbaas09:33
openstackgerritZhaoBo proposed openstack/octavia master: [WIP]Extend api to accept qos_policy_id  https://review.openstack.org/45830809:40
*** sanfern has quit IRC09:47
*** gcheresh_ has quit IRC10:02
*** amotoki has quit IRC10:05
*** yamamoto has quit IRC10:25
*** reedip has quit IRC10:41
*** gcheresh_ has joined #openstack-lbaas11:03
openstackgerritAdam Harwell proposed openstack/octavia master: Fix import in v2 pool types  https://review.openstack.org/45372111:06
rm_worktaking over and updating more patches11:06
openstackgerritAdam Harwell proposed openstack/octavia master: Implement sorting and pagination for octavia  https://review.openstack.org/38214711:09
rm_workjohnsom: rebased that, it's ready for review11:09
rm_worki'll probably end up rebasing it on top of quotas I guess <_<11:10
rm_workor, does quotas actually need pagination?11:10
*** sanfern has joined #openstack-lbaas11:13
*** sanfern has quit IRC11:24
*** sanfern has joined #openstack-lbaas11:29
*** yamamoto has joined #openstack-lbaas11:33
*** yamamoto has quit IRC11:35
*** yamamoto has joined #openstack-lbaas11:39
*** yamamoto has quit IRC11:52
*** ajo has quit IRC12:25
*** sanfern has quit IRC12:28
*** ajo has joined #openstack-lbaas12:33
openstackgerritReedip proposed openstack/octavia master: [WIP]Extend api to accept qos_policy_id  https://review.openstack.org/45830812:35
*** reedip has joined #openstack-lbaas12:47
*** vlaza has quit IRC12:48
openstackgerritReedip proposed openstack/octavia master: [WIP]Extend api to accept qos_policy_id  https://review.openstack.org/45830812:52
*** yamamoto has joined #openstack-lbaas12:57
*** yamamoto has quit IRC12:58
*** yamamoto has joined #openstack-lbaas12:58
*** yamamoto has quit IRC13:15
*** yamamoto has joined #openstack-lbaas13:16
*** yamamoto has quit IRC13:16
*** yamamoto has joined #openstack-lbaas13:16
*** vlaza has joined #openstack-lbaas13:24
*** yamamoto has quit IRC13:25
*** yamamoto has joined #openstack-lbaas13:26
*** yamamoto has quit IRC13:30
*** vlaza has left #openstack-lbaas13:32
*** cpuga has joined #openstack-lbaas13:43
*** chlong has joined #openstack-lbaas13:44
*** cpuga has quit IRC13:47
*** cpuga has joined #openstack-lbaas13:49
*** catintheroof has joined #openstack-lbaas13:59
*** catintheroof has quit IRC14:10
*** catintheroof has joined #openstack-lbaas14:10
*** rcernin has quit IRC14:24
xgermanrm_work yt?14:26
*** yamamoto has joined #openstack-lbaas14:27
*** amotoki has joined #openstack-lbaas14:33
*** fnaval has joined #openstack-lbaas14:33
*** cpuga has quit IRC14:34
*** fnaval has quit IRC14:35
*** fnaval has joined #openstack-lbaas14:35
*** amotoki has quit IRC14:36
*** cpuga has joined #openstack-lbaas14:36
*** amotoki has joined #openstack-lbaas14:39
*** cpuga_ has joined #openstack-lbaas14:40
*** gongysh has joined #openstack-lbaas14:41
*** cpuga has quit IRC14:43
*** armax has joined #openstack-lbaas14:46
*** yamamoto has quit IRC14:53
*** chlong_ has joined #openstack-lbaas14:54
*** chlong_ has quit IRC14:54
*** blogan has joined #openstack-lbaas14:57
*** krypto has joined #openstack-lbaas14:58
*** krypto has joined #openstack-lbaas14:58
*** voelzmo has quit IRC15:05
*** voelzmo has joined #openstack-lbaas15:05
*** amotoki has quit IRC15:09
*** voelzmo has quit IRC15:10
*** chlong has quit IRC15:13
johnsomo/15:14
*** aojea_ has quit IRC15:16
*** amotoki has joined #openstack-lbaas15:26
*** amotoki has quit IRC15:30
*** amotoki has joined #openstack-lbaas15:32
*** gongysh has quit IRC15:33
*** amotoki has quit IRC15:35
*** reedip has quit IRC15:36
*** blogan_ has joined #openstack-lbaas15:42
*** blogan has quit IRC15:45
*** amotoki has joined #openstack-lbaas15:48
*** yamamoto has joined #openstack-lbaas15:51
*** gcheresh_ has quit IRC16:10
*** blogan_ is now known as blogan16:16
*** fnaval has quit IRC16:25
*** voelzmo has joined #openstack-lbaas16:27
*** blogan_ has joined #openstack-lbaas16:27
*** blogan has quit IRC16:27
*** blogan_ is now known as blogan16:28
*** voelzmo has quit IRC16:33
rm_workxgerman: here now16:38
*** amotoki has quit IRC16:39
rm_workxgerman: you've got a number of outstanding notes on https://review.openstack.org/#/c/448759/16:41
rm_workbut it's close16:41
xgermanyep16:41
xgermanso it seems devstack does not survive reboot16:42
xgermanso how did we set up stuff last time?16:42
*** nmagnezi has quit IRC16:52
*** ajo has quit IRC16:54
*** yamamoto has quit IRC16:55
*** cpuga_ has quit IRC16:57
*** ajo has joined #openstack-lbaas17:00
*** nmagnezi has joined #openstack-lbaas17:02
*** krypto has quit IRC17:15
rm_workxgerman: right, it does not17:15
rm_workxgerman: on the morning of the event we need to spin up all the VMs and then install devstack17:16
xgermanyeah, I like to start the people off with some stuff. So you scripted something ti run stack or should I make the user do that + start up their vms?17:16
rm_workthat is why I sent an annoyed email about not getting the resources earlier17:16
xgermangotcha17:16
rm_workand they were like "just make an image!" and I said "yeah but no"17:16
xgermanI recall17:17
openstackgerritMerged openstack/octavia master: Fix import in v2 pool types  https://review.openstack.org/45372117:19
*** tesseract has quit IRC17:19
johnsomUgh, why in the world did we let the insert_headers be on listeners instead of pools???17:20
rm_workjohnsom: oh uhh... on listeners, should "default_tls_container_ref" return as None in the response, or not be there17:21
rm_workif it isn't filled17:21
rm_workjust made some changes to make things actually ... correct, on the listener return. posting as soon as I know the answer to that17:21
johnsom        "default_tls_container_ref": null,17:23
johnsom"sni_container_refs": [],17:23
rm_worki mean17:23
rm_workis that from neutron-lbaas?17:23
*** ig0r_ has quit IRC17:24
rm_workkk17:24
openstackgerritAdam Harwell proposed openstack/octavia master: Correct some v2 Listener response entries  https://review.openstack.org/45892017:24
rm_workwe had some ... issues17:24
johnsomYeah, that was a nlbaas show details17:24
johnsomDoes that close one of the bugs I opened?17:25
johnsomhttps://bugs.launchpad.net/octavia/+bug/168156417:25
openstackLaunchpad bug 1681564 in octavia "Octavia v2 API listener response is missing fields" [Medium,Triaged]17:25
johnsomhttps://bugs.launchpad.net/octavia/+bug/168156517:25
openstackLaunchpad bug 1681565 in octavia "Octavia v2 API listener has unnecessary response field" [Medium,Triaged]17:25
rm_worknot sure17:25
rm_workI don't see insert_headers in the *neutron-lbaas* response17:27
johnsomI think that only got implemented in octavia17:28
johnsomLet me check17:28
johnsomYeah, that was an octavia only feature17:29
*** ig0r_ has joined #openstack-lbaas17:29
johnsomGeez, because of that I am so temped to move it17:29
rm_workwhelp17:30
johnsomLeave it in listener deprecated but expose it on pool going forward.17:30
openstackgerritAdam Harwell proposed openstack/octavia master: Correct some v2 Listener response entries  https://review.openstack.org/45892017:31
rm_workadded the bug numbers17:31
johnsomI know we discussed this, but I don't remember why we would put it on listener.  It isn't a per-listener setting, it's per pool.17:36
rm_workerm17:36
rm_worklol17:36
rm_workis it?17:37
rm_workI feel like you could make a case for either maybe17:38
johnsomlisteners are front end, all about the request from the client.  pools address backend.  The headers are only inserted into requests to the backend17:41
*** cpuga has joined #openstack-lbaas17:41
rm_workah so in the haproxy config, that line gets dropped in the `backend` section?17:41
rm_workI guess that makes sense17:42
johnsomIt's even coded that way: https://github.com/openstack/octavia/blob/master/octavia/common/jinja/haproxy/templates/macros.j2#L20617:42
rm_workhmmm that's coded very statically for those two things17:42
rm_workwe need to fix that to actually do insert_headers properly17:42
johnsomAs it is, if you set one listener one way, and the other another way, it's going to be global last one wins I think17:42
openstackgerritGerman Eichberger proposed openstack/octavia master: Initial version of the operator guide  https://review.openstack.org/44875917:43
rm_workerm well17:43
rm_workthe config is per listener17:43
rm_worklike each listener gets its own haproxy config17:44
johnsomIn fact, which listener does it pick now that pools aren't tied to listeners17:44
rm_workthe one for the config it's generating :P17:44
rm_workeach haproxy config has only one frontend, right?17:44
rm_workbecause it's literally /var/lib/octavia/<listener_id>/haproxy.conf17:45
johnsomWell, yes, I hate that too17:45
rm_workyeah but that means it's actually explicit which one it is choosing, lol17:45
johnsomBut I am thinking of the L7 case, where pool is tied to load balancer and l7 points to it17:45
rm_workso when it generates the config for THAT listener, the pool will get those headers17:45
rm_workbut when it generates the config for the OTHER listener, it'll get a different one17:46
rm_workallows a pool to serve for HTTPS and TLS_TERM17:46
rm_workand use headers17:46
rm_workbecause the TLS_TERM listener has the insert_headers17:46
rm_worksure, if they were on the pool we could just ignore them I guess17:46
rm_workbut we'd still have to check the listener type, right?17:46
*** fnaval has joined #openstack-lbaas17:46
johnsomRight, but say you have an LB, which has three pools.  One pool is tied to the listener, the others are tied to LB.  The listener has L7 policies that point to the other two pools.17:47
johnsomNo, we should check the pool protocol as that is what matters IF we can insert them17:47
rm_workright17:47
rm_workerrr17:48
rm_workhmm17:48
rm_workif the Listener protocol is HTTPS, we can't insert headers17:48
rm_workright?17:48
johnsomYes, you can, for the PROXY protocol17:48
rm_workhmm17:48
johnsomThis patch: https://review.openstack.org/#/c/45875517:48
*** cpuga has quit IRC17:48
*** cpuga has joined #openstack-lbaas17:49
johnsomWhich has bugs BTW17:49
rm_workthat isn't on the POOL though17:50
rm_workor at least not with that patch it isn't17:50
rm_workbecause he didn't change the SUPPORTED_PROTOCOLS17:50
rm_workso you can't set a pool's protocol to PROXY...17:50
johnsomYeah, I think there might be an issue there with the patch.  Plus it's under a check for HTTP protocol, where PROXY only really makes sense for TCP connections17:51
xgerman+!17:51
xgerman+117:51
rm_workxgerman: should I send another email to the OVH folks and be like "confirmed, we can't rely on images"?17:54
eanderssonI'll vote for you next cycle xgerman!17:55
xgermanno worries — I had a pretty good grip on my changes ;-)17:55
xgermanrm_work yep, let’s do that :-)17:55
xgermanalso johnsom not sure if you saw https://image.slidesharecdn.com/dockerconcilium-170418181631/95/dockercon-2017-cilium-network-and-application-security-with-bpf-and-xdp-59-638.jpg?cb=149253947617:58
xgermanXDP claims similat performance than DPDK17:58
johnsomHmm, I think it will depend on the use cases.  DPDK isn't just about bypassing network stacks18:00
xgermanFB is sing this for LB18:03
xgermanusing18:03
xgermanso making this a distributor ;-)18:03
johnsomI think I actually went to their talk in Barcelona: https://www.openstack.org/videos/barcelona-2016/leveraging-express-data-path-xdp-for-programmable-high-performance-data-path-in-openstack18:04
johnsomIt was very plumgrid focused18:04
xgermanprobably18:04
xgermante way FB is running it is to have this on eavh vm and bounce out traffic not meant for that vm18:05
xgermanhttp://prototype-kernel.readthedocs.io/en/latest/networking/XDP/use-cases/xdp_use_case_load_balancer.html18:10
*** diltram has quit IRC18:17
*** diltram has joined #openstack-lbaas18:27
*** amotoki has joined #openstack-lbaas18:39
*** jerrygb has quit IRC18:40
*** JudeC has joined #openstack-lbaas18:43
*** amotoki has quit IRC18:44
rm_workjohnsom: I feel like Pool and Listener shouldn't have the same constants.SUPPORTED_PROTOCOLS list for protocol18:49
*** jerrygb has joined #openstack-lbaas18:49
rm_worksince ... I think pools can't use PROTOCOL_TERMINATED_HTTPS :P18:49
*** med_ has quit IRC18:54
xgermanyep18:55
*** Guest76746 has joined #openstack-lbaas18:55
*** voelzmo has joined #openstack-lbaas19:01
rm_workjohnsom: do you have the etherpad somewhere that had our discussion about how single-create would work19:02
rm_workwith regard to the post body and referencing pools and such19:02
*** pcaruana has quit IRC19:02
johnsomHmm, I will look when I get done with lunch.  There is also a desription and example in the lb api-ref19:04
*** kbyrne has quit IRC19:04
rm_workkk19:05
johnsomrm_work: on the pool protocol I thought so too, be we are mapping it in the code so I documented it that way19:06
rm_workyeah so there is, i see19:06
rm_workjohnsom: yeah i meant, i think in the code this is a bug19:06
johnsomI guess we have to look at what nlbaas did19:08
johnsomI would support dropping terminated on pools as it makes no sense if we can19:09
rm_workright19:09
rm_workit should be a different list19:10
rm_workuhh so on the api-ref for LB single-create19:10
rm_workhow does the l7policy specify a pool19:10
rm_workI assume by name19:11
rm_workbut it isn't there19:11
rm_workI'm going to propose a different thing, along with docs changes19:15
rm_workwe'll see what you think19:15
johnsomSounds good, code speaks!  Grin19:17
*** cody-somerville has joined #openstack-lbaas19:28
*** csomerville has quit IRC19:30
johnsomrm_work I'm not finding an etherpad for single-call-create (get me a load balancer).19:33
johnsomI am not sure I ever saw one19:34
johnsomIt was brandon and trevor working on it, so maybe there was an internal one I never saw?19:34
rm_workhmm i thought it was a convo between US19:34
rm_workbut maybe i was thinking of that api-ref19:34
rm_workanywho19:34
rm_workdeep into making this WORK19:34
johnsomSweet19:35
rm_workkinda almost hoping I can show basics EOD19:35
johnsomI am working on reviews, slides, and pool api-ref.  I kind of want to make some formatting changes to listeners, should I wait for that to merge or do it now?19:35
johnsomI know you folks has started reviewing...19:36
*** gcheresh_ has joined #openstack-lbaas19:36
rm_worki started but hadn't made comments yet19:36
johnsomOk, maybe I should just do it then.  I decided on a new way to format the sub-objects while working on session persistence.  It's cleaner.19:37
*** voelzmo has quit IRC19:38
rm_workkk19:38
rm_workyeah that happens :P19:38
rm_worki always had to go back and redo stuff during the api patches19:38
rm_workas I refined the way i did the testing19:38
-openstackstatus- NOTICE: Gerrit will be offline briefly starting at 20:00 for scheduled maintenance http://lists.openstack.org/pipermail/openstack-dev/2017-April/115702.html19:42
*** blogan has quit IRC19:44
*** blogan has joined #openstack-lbaas19:45
xgermanmmh, my TLS example can’t read the cert while openssl can19:45
xgermanOctavia has issyes19:46
xgermanissues19:46
xgermanhttps://www.irccloud.com/pastebin/HNknZmn9/19:47
xgermanlooks like a bug to me19:47
rm_workyeah there's a bug filed already19:49
rm_workfor that19:49
rm_worki haven't had time19:49
xgermanthere is?19:50
rm_workyes19:50
xgermanbecause I filed a new one19:50
rm_work:/19:50
xgermanyou have the URL?19:50
rm_workwhich is yours19:50
xgermanhttps://bugs.launchpad.net/octavia/+bug/168534919:50
openstackLaunchpad bug 1685349 in octavia "Can't parse certificate - TypeError: initializer for ctype 'char[]' must be a str or list or tuple, not unicode" [Undecided,New]19:50
rm_workhttps://bugs.launchpad.net/octavia/+bug/1681595 and https://bugs.launchpad.net/octavia/+bug/162451219:51
openstackLaunchpad bug 1681595 in octavia "Octavia is not working with barbican" [Critical,Triaged]19:51
openstackLaunchpad bug 1624512 in octavia "AttributeError: 'unicode' object has no attribute 'tls_container_id' during TLS barbican scenario testing" [Critical,New]19:51
rm_workI believe both of those are already dupes19:51
rm_workor rather, one is19:51
xgerman:-(19:51
rm_workmaking yours the third :P19:51
xgermanok, I want to demo TLS in BOS so will see if I can fix19:51
rm_workhmm maybe just https://bugs.launchpad.net/octavia/+bug/168159519:52
openstackLaunchpad bug 1681595 in octavia "Octavia is not working with barbican" [Critical,Triaged]19:52
xgermanlet me look why it doesn’t work19:52
rm_workbut it's basically the same thing19:52
rm_workif you can't do it, i can look next week19:52
xgermank, I will leave it to johnsom to mark them all as dupes19:52
johnsomOk19:52
johnsomYeah, german's is a dup of mine, but I think the other two may be the same thing, but different parts of the code.19:55
johnsomSo I left those two19:55
*** ig0r_ has quit IRC19:55
xgermank19:55
-openstackstatus- NOTICE: Gerrit is offline briefly for scheduled maintenance http://lists.openstack.org/pipermail/openstack-dev/2017-April/115702.html20:03
*** ChanServ changes topic to "Gerrit is offline briefly for scheduled maintenance http://lists.openstack.org/pipermail/openstack-dev/2017-April/115702.html"20:03
*** jerrygb has quit IRC20:10
*** jerrygb has joined #openstack-lbaas20:25
*** aojea has joined #openstack-lbaas20:27
*** aojea_ has joined #openstack-lbaas20:35
*** ChanServ changes topic to "Welcome to LBaaS / Octavia"20:35
-openstackstatus- NOTICE: Gerrit is back in service and generally usable, though remote Git replicas (git.openstack.org and github.com) will be stale for the next few hours until online reindexing completes20:35
*** aojea has quit IRC20:38
*** amotoki has joined #openstack-lbaas20:40
*** aojea has joined #openstack-lbaas20:40
*** aojea_ has quit IRC20:44
*** amotoki has quit IRC20:46
*** openstackgerrit has quit IRC20:48
*** yamamoto has joined #openstack-lbaas20:54
*** gcheresh_ has quit IRC20:55
*** openstackgerrit has joined #openstack-lbaas20:57
openstackgerritGerman Eichberger proposed openstack/octavia master: Fixes Octavia not working with Barbican  https://review.openstack.org/45896820:57
*** yamamoto has quit IRC20:58
*** aojea_ has joined #openstack-lbaas21:02
*** aojea has quit IRC21:04
*** bread has joined #openstack-lbaas21:06
*** bread is now known as Guest4665821:07
*** cpuga has quit IRC21:11
openstackgerritGerman Eichberger proposed openstack/neutron-lbaas master: Octavia Proxy Plugin  https://review.openstack.org/41853021:15
*** aojea_ has quit IRC21:17
*** aojea has joined #openstack-lbaas21:18
*** aojea_ has joined #openstack-lbaas21:19
*** aojea__ has joined #openstack-lbaas21:22
*** aojea has quit IRC21:23
openstackgerritMichael Johnson proposed openstack/octavia master: Add v2 listener API section  https://review.openstack.org/45575521:25
*** aojea_ has quit IRC21:25
*** kbyrne has joined #openstack-lbaas21:39
*** amotoki has joined #openstack-lbaas21:42
johnsomJudeC Take a look at this bug and let me know if that is what you are looking for.21:46
johnsomhttps://bugs.launchpad.net/octavia/+bug/168536921:46
openstackLaunchpad bug 1685369 in octavia "Create OSC loadbalancer list command" [High,In progress]21:46
johnsomThe current loadbalancer list command might be missing some of these options.  I am trying to mirror the functionality of neutron to some degree.21:46
JudeCThis is perfect21:47
JudeCand yeah it is missing all of this functionality21:47
*** amotoki has quit IRC21:47
johnsomI missed the default output columns, adding to the description21:47
JudeCRight now 'loadbalancer list' just lists all loadbalancers in octavia.21:47
*** cpuga has joined #openstack-lbaas21:48
*** aojea__ has quit IRC21:48
johnsomYeah, I don't think filters have been implemented in the octavia API yet, so it might be something that needs to be finished later.21:48
rm_workJudeC / johnsom correct21:50
rm_workno filters yet21:50
rm_workand sorting/paging is still in review even21:50
JudeChttps://developer.openstack.org/api-ref/load-balancer/v2/index.html?expanded=update-a-load-balancer-detail21:50
JudeCis this wrong?21:50
JudeCFiltering and column selection¶21:50
JudeCThe Octavia API v2.0 supports filtering based on all top level attributes of a resource. Filters are applicable to all list requests.21:50
JudeCAccording to that doc21:50
johnsomJudeC It is correct and what we should have, but it just doesn't work  yet.21:51
JudeCahhh21:51
JudeCok21:51
JudeCthank you guys.21:52
johnsomOk, I will put in bugs for the load balancer commands.  You can list them with: https://bugs.launchpad.net/octavia/+bugs?field.tag=octavia-osc21:54
JudeCThis is fantastic thank you so much haha.22:02
johnsomSure, ping me if anything seems odd.  I might make a mistake.22:02
johnsomI'm not sure how you want to handle things that should be in the client but aren't implemented in the API yet.22:05
johnsomExample would be provider.  Should we just stub it out and return a blank?22:05
JudeChmm.22:05
johnsomWell, I am going to put a "note" in the bugs for things that are still coming.  I'm not sure which way is easier/better from the client code perspective.  We know they are coming, just not there yet.22:07
JudeCI mean maybe I will prioritize the stuff that is already implemented in the API22:08
JudeCand then some of the stuff that hasn't been completed yet get to at the end?22:08
JudeCand just stub it out22:08
johnsomYeah, if it's not a thing where it just makes sense to stub them, skip everything that has  a note.  We can just come back and do them as they are implemented22:08
JudeCPerfect22:09
*** fnaval has quit IRC22:13
*** csomerville has joined #openstack-lbaas22:20
*** cody-somerville has quit IRC22:20
*** csomerville has quit IRC22:26
*** blogan has quit IRC22:30
*** catintheroof has quit IRC22:46
*** cpuga has quit IRC22:50
rm_work@(*$YC*@)22:55
rm_workwsme22:55
rm_workaugh22:55
rm_workfighting this so hard22:55
JudeC??22:55
johnsomHa, it does take some mind bending.....22:55
johnsomOh, I know exactly what he is experiencing22:55
johnsom...22:55
johnsomrm_work let me know if you want a second set of eyes.  I also expect we may need more types for the single-call-create stuff22:56
rm_workyes22:57
rm_workyes we do22:57
rm_workthe problem is dynamically deciding which to return22:57
rm_worki have everything done but that22:57
rm_workpost() has a specific type set in the wsexpose()22:57
rm_workI can't figure out how to dynamically override that22:57
rm_workwithout doing it globally22:57
rm_workI can do:22:57
johnsomOh, yeah, umm, that is interesting...22:58
rm_workif isinstance(loadbalancer, LoadBalancerFullResponse):22:58
rm_work    self._wsme_attributes[0].datatype = LoadBalancerFullResponse22:58
johnsomMy gut is telling me I saw brandon use a trick for that at one point.22:58
rm_workbut that still sets it for EVERYTHING22:58
rm_workso the next request would get a Full response22:58
rm_workregardless of what type it is22:58
johnsomOh, hmm22:59
rm_workI *assume* it's threadsafe?!22:59
rm_workso22:59
rm_workmaybe I can just get away with specifically setting one or the other22:59
rm_workevery time22:59
rm_workyeah fuck it for now that'll work23:00
rm_workcan fix it in review23:00
johnsomThe other thought I have is include them both in the response type and just not fill it in for the either / or23:00
johnsomIt can hide things23:00
rm_workerr23:02
rm_workhow can you do that23:02
rm_workcan I specify multiple response types?!23:02
rm_workthat'd solve my problem cleaner than this...23:02
johnsomI don't think you can specify multiple return types23:05
rm_work...23:07
rm_workthen what do you mean exactly23:07
johnsomI'm trying to dig a little on that, but the summary is if the result for a field is null or "unset" the field isn't returned in the response at all.  The Intel folks had this problem a few times, I would tag them for missing fields in the response so they had to put in things like ""23:08
johnsomI'm just trying to figure out how "null" works vs. not there at all23:09
johnsomSo, you could put both fields in the response type, then use the "from_data_model" to swap them out.  At least that is my thought.  Which might be complete BS...23:09
johnsomrm_work Like brandon did in the v1 api: https://github.com/openstack/octavia/blob/master/octavia/api/v1/types/load_balancer.py#L5023:13
johnsomIt's that whole "children" thing23:14
rm_workerr23:14
rm_workso23:14
rm_workI can make them appear/disappear23:14
rm_workbut I can't assign two different types to the same field name23:15
rm_workbecause of the wsme magic23:15
rm_workso i could have the root type with two attributes23:15
rm_workloadbalancer, loadbalancer_full23:15
rm_workand I could have only one show up23:15
rm_workbut that's ...23:15
xgermanmmh, not sure if this is me: http://logs.openstack.org/68/458968/1/check/gate-octavia-v1-dsvm-scenario-kvm-centos.7-ubuntu-xenial-nv/ee26ff8/console.html23:15
johnsomRight.  I think we did listener_ids and listeners for example23:16
rm_workright but23:16
rm_workthis is the root type23:16
rm_work"loadbalancer" is the *not full* one23:16
rm_workso23:16
rm_workhaving a create return a different root type23:17
rm_workdepending on what data you send in23:17
rm_workis really dumb23:17
rm_workhere, one sec23:17
rm_workI basically just need to copy/paste some stuff and I'll push up something that should work23:17
rm_work"work"23:17
*** JudeC has quit IRC23:18
johnsomxgerman This points to a problem with the patch: 2017-04-21 21:26:23.776 11171 DEBUG octavia.amphorae.drivers.haproxy.rest_api_driver [-] Connected to amphora. Response: <Response [400]> request23:19
xgermanoh, well, will have a look after dinn3423:20
johnsomThe amp isn't happy with something being sent to it23:20
johnsomJude must work somewhere east of here and is done for the day....23:30
johnsomOr subscribes to the beer-thirty concept23:30
rm_workhe's on a little bit sporadically, like me, I think23:38
rm_workbut like all the time23:38
johnsomHa, ok.  I just wanted to mention that I'm using the term "output fields" for tables that have the values in rows and "output columns" for horizontal tables.23:39
johnsomI think I have a pretty good pass on the load balancer bugs for him23:40
*** yamamoto has joined #openstack-lbaas23:56
johnsomI guess I will work on the project update form next week...23:58
« Thursday, 2017-04-20 Index Saturday, 2017-04-22 »

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!