Monday, 2016-02-29

*** Aish has left #openstack-lbaas00:11
sbalukoffThanks for the +2 ptoohill!00:31
sbalukoffTrevorV: Stacking to checkout the latest one-call-create patches (this time incorporating the controller patch as well. XD)00:35
sbalukoffDoes anyone know who the cores are for the python-neutronclient project?00:56
*** hayato_ has joined #openstack-lbaas01:00
*** paco20151113 has joined #openstack-lbaas01:38
*** yamamoto_ has joined #openstack-lbaas01:50
*** ducttape_ has joined #openstack-lbaas01:53
*** yamamoto_ has quit IRC01:55
*** ducttape_ has quit IRC02:05
*** yamamoto_ has joined #openstack-lbaas02:08
*** yamamoto_ has quit IRC02:09
*** manishg has quit IRC02:12
*** ajmiller has quit IRC02:25
*** yamamoto has joined #openstack-lbaas02:33
*** yamamoto has quit IRC02:40
*** manishg has joined #openstack-lbaas02:43
*** mdavidson has quit IRC02:46
*** paco20151113 has quit IRC02:47
*** manishg has quit IRC02:48
*** paco20151113 has joined #openstack-lbaas02:48
*** kfox1111 has quit IRC02:49
*** kfox1111 has joined #openstack-lbaas02:49
*** mdavidson has joined #openstack-lbaas02:50
*** manishg has joined #openstack-lbaas03:05
*** manishg has quit IRC03:09
*** neelashah has joined #openstack-lbaas03:22
*** yamamoto_ has joined #openstack-lbaas03:23
*** Purandar has joined #openstack-lbaas03:26
*** hayato__ has joined #openstack-lbaas03:39
*** hayato_ has quit IRC03:41
*** links has joined #openstack-lbaas03:44
*** neelashah has quit IRC03:51
*** links has quit IRC04:01
*** hayato_ has joined #openstack-lbaas04:05
*** manishg has joined #openstack-lbaas04:06
*** hayato__ has quit IRC04:08
*** hayato__ has joined #openstack-lbaas04:09
*** hayato___ has joined #openstack-lbaas04:10
*** hayato_ has quit IRC04:10
*** hayato_ has joined #openstack-lbaas04:11
*** manishg has quit IRC04:11
*** hayato__ has quit IRC04:11
*** hayato__ has joined #openstack-lbaas04:11
*** hayato___ has quit IRC04:15
*** hayato_ has quit IRC04:15
*** links has joined #openstack-lbaas04:18
openstackgerritMadhusudhan Kandadai proposed openstack/neutron-lbaas: [WIP] Neutron LBaaS: TLS Barbican Scenario Test
*** numans has joined #openstack-lbaas04:59
*** manishg has joined #openstack-lbaas05:07
*** manishg has quit IRC05:11
*** manishg has joined #openstack-lbaas05:12
rm_yousbalukoff: does trevor's fix look good? was just about to test it05:13
sbalukoffAm just getting back to it. Looks like it works to create a complicated LB tree...05:13
sbalukoffTrying to break it now. :)05:13
sbalukoffIt caught my l7rule error...05:13
sbalukoffSo that's good.05:13
sbalukoffGoing to try an l7policy error.05:14
rm_youi have been mostly testing happy-path, since if it passes that, i'll +2 for merge and then file bugs for any issues on the other paths05:14
sbalukoffyeah, so far happy path looks good.05:14
sbalukoffI'm suspecting that I can break it with an l7policy error... but I would blame that more on how I was doing validations...05:15
sbalukoff(There's a patch go fix that...)05:15
sbalukoff(But I don't want that patch getting in the way of this feature merging.05:16
sbalukoffsince that's a bugfix and the feature freeze is tomorrow.)05:16
sbalukoffOh, awesome! He catches it!05:19
sbalukoffSo yeah... it looks like this works.05:19
sbalukoffI want to go over the code some more, but from what I can see, this is looking like a +2 from me so far with caveats (again, I want a bug to back-fill the documentation on this, and a bug showing shared-pools support is missing from this.)05:20
sbalukoffBut first, I'mma eat this pizza. :)05:21
rm_youi want to +2 it as well, but first I'mma let this tequila wear off05:22
*** yuanying has quit IRC05:24
*** yuanying_ has joined #openstack-lbaas05:24
xgermanwell, sbalukoff those LBaaS reviewers are slackers — I didn’t get much bites for my cascading delete stuff...05:29
*** prabampm has joined #openstack-lbaas05:46
*** fawadkhaliq has joined #openstack-lbaas05:54
*** amotoki has joined #openstack-lbaas05:56
*** kevo has joined #openstack-lbaas06:01
*** manishg has quit IRC06:01
openstackgerritlei zhang proposed openstack/octavia: Replace deprecated LOG.warn with LOG.warning
*** kevo has quit IRC06:19
sbalukoffxgerman: I think it helps to poke them directly.06:26
rm_youyeah generally i get best results when i find people who should be reviewing and poke them in IRC06:29
rm_youas a reviewer, often i don't have time to just randomly browse (even though ideally i'd06:29
rm_you*i'd be doing just that)06:29
*** kevo has joined #openstack-lbaas06:30
*** manishg has joined #openstack-lbaas06:32
*** manishg has quit IRC06:36
*** Purandar has quit IRC06:38
*** kobis has joined #openstack-lbaas06:42
*** outofmemory is now known as reedip_06:44
*** rcernin has joined #openstack-lbaas06:44
*** rcernin has quit IRC06:49
sbalukoffxgerman: Are you still around?06:50
sbalukoffI wanted to talk with you about your comments on the one-call-create controller tasks (and for loops therein)06:50
*** kevo has quit IRC06:56
*** hayato__ has quit IRC07:04
*** bana_k has quit IRC07:05
*** fnaval has joined #openstack-lbaas07:12
*** fawadk has joined #openstack-lbaas07:15
*** fawadkhaliq has quit IRC07:15
*** fawadkhaliq has joined #openstack-lbaas07:15
*** amotoki has quit IRC07:19
*** fawadk has quit IRC07:20
*** rcernin has joined #openstack-lbaas07:24
*** amotoki has joined #openstack-lbaas07:31
*** manishg has joined #openstack-lbaas07:33
openstackgerritYi Zhao proposed openstack/neutron-lbaas: Fixed typo in
*** amotoki has quit IRC07:35
*** manishg has quit IRC07:38
*** pcaruana has joined #openstack-lbaas07:43
*** fawadkhaliq has quit IRC07:43
*** evgenyf has joined #openstack-lbaas07:46
*** amotoki has joined #openstack-lbaas07:49
*** amotoki has quit IRC07:49
*** fnaval has quit IRC07:50
*** fnaval has joined #openstack-lbaas07:59
*** fnaval has quit IRC08:04
*** ihrachys has joined #openstack-lbaas08:06
*** fnaval has joined #openstack-lbaas08:07
*** kevo has joined #openstack-lbaas08:08
*** fnaval has quit IRC08:11
*** kevo has quit IRC08:15
*** hayato_ has joined #openstack-lbaas08:15
*** hayato_ has quit IRC08:15
*** hayato_ has joined #openstack-lbaas08:16
*** fawadkhaliq has joined #openstack-lbaas08:27
*** rcernin has quit IRC08:29
*** manishg has joined #openstack-lbaas08:34
*** amotoki has joined #openstack-lbaas08:35
*** amotoki has quit IRC08:36
*** rcernin has joined #openstack-lbaas08:37
*** fawadkhaliq has quit IRC08:38
*** manishg has quit IRC08:38
*** amotoki has joined #openstack-lbaas08:45
*** jschwarz has joined #openstack-lbaas08:47
reedip_sbalukoff: ping08:49
*** eezhova has quit IRC08:51
reedip_sbalukoff: I have mentioned 2 syntaxes which are used to mention a list in NeutronClient . Considering that you are using Syntax 1, IMHO, it seems that changing the message for --sni-container-refs would be enough.08:54
reedip_sbalukoff: but I have given option#2 , considering the 3rd Syntax.08:54
reedip_sbalukoff: which is more useful08:55
*** nmagnezi has joined #openstack-lbaas08:56
*** amotoki has quit IRC09:10
*** amotoki has joined #openstack-lbaas09:13
*** nmagnezi has quit IRC09:15
*** amotoki has quit IRC09:24
*** nmagnezi has joined #openstack-lbaas09:28
*** amotoki has joined #openstack-lbaas09:34
*** manishg has joined #openstack-lbaas09:34
*** paco20151113 has quit IRC09:36
*** manishg has quit IRC09:39
*** evgenyf has quit IRC09:53
*** ihrachys has quit IRC09:54
*** ihrachys has joined #openstack-lbaas09:59
*** kobis has quit IRC10:20
*** kobis has joined #openstack-lbaas10:22
*** manishg has joined #openstack-lbaas10:35
*** hayato_ has quit IRC10:39
*** manishg has quit IRC10:40
*** yamamoto_ has quit IRC10:53
*** amotoki has quit IRC11:00
*** amotoki has joined #openstack-lbaas11:03
*** amotoki has quit IRC11:11
*** yamamoto has joined #openstack-lbaas11:31
*** yamamoto_ has joined #openstack-lbaas11:33
*** otsuka_ has joined #openstack-lbaas11:36
*** yamamoto has quit IRC11:36
*** manishg has joined #openstack-lbaas11:38
*** kobis has quit IRC11:38
*** otsuka has quit IRC11:39
*** kobis has joined #openstack-lbaas11:41
*** manishg has quit IRC11:42
*** kobis has quit IRC11:42
*** kobis has joined #openstack-lbaas11:43
*** kobis has quit IRC11:47
*** kobis has joined #openstack-lbaas11:48
*** kobis has quit IRC11:53
*** kobis has joined #openstack-lbaas11:56
*** otsuka has joined #openstack-lbaas12:13
*** otsuka_ has quit IRC12:15
*** otsuka_ has joined #openstack-lbaas12:23
*** otsuka has quit IRC12:26
*** amotoki has joined #openstack-lbaas12:30
*** H3y has joined #openstack-lbaas12:35
*** otsuka has joined #openstack-lbaas12:37
*** manishg has joined #openstack-lbaas12:38
*** otsuka_ has quit IRC12:41
*** rtheis has joined #openstack-lbaas12:42
*** manishg has quit IRC12:43
*** amotoki has quit IRC12:45
*** pcaruana has quit IRC12:53
*** prabampm has quit IRC12:54
*** amotoki has joined #openstack-lbaas13:02
*** Oku_OS has joined #openstack-lbaas13:05
*** pcaruana has joined #openstack-lbaas13:08
*** otsuka has quit IRC13:18
*** otsuka has joined #openstack-lbaas13:18
*** links has quit IRC13:18
*** amotoki has quit IRC13:39
*** manishg has joined #openstack-lbaas13:39
*** otsuka has quit IRC13:40
*** fawadkhaliq has joined #openstack-lbaas13:43
*** manishg has quit IRC13:44
*** amotoki has joined #openstack-lbaas13:53
*** neelashah has joined #openstack-lbaas13:58
*** links has joined #openstack-lbaas13:58
*** otsuka has joined #openstack-lbaas14:06
*** otsuka has quit IRC14:07
*** otsuka has joined #openstack-lbaas14:08
*** localloo1 has joined #openstack-lbaas14:12
nmagnezidougwig, ping re: haproxy namespace14:13
*** links has quit IRC14:16
*** ajmiller has joined #openstack-lbaas14:18
*** ajmiller has quit IRC14:25
*** localloo1 has quit IRC14:37
openstackgerritIhar Hrachyshka proposed openstack/octavia: glance: support relying on tags to extract image id
*** pcaruana has quit IRC14:38
*** H3y has quit IRC14:38
*** H3y has joined #openstack-lbaas14:39
*** manishg has joined #openstack-lbaas14:40
dougwignmagnezi: ack14:43
*** manishg has quit IRC14:45
*** woodster_ has joined #openstack-lbaas14:49
*** pcaruana has joined #openstack-lbaas14:52
*** ducttape_ has joined #openstack-lbaas14:53
*** doug-fish has joined #openstack-lbaas14:56
*** TrevorV has joined #openstack-lbaas15:00
*** otsuka has quit IRC15:03
*** H3y has quit IRC15:04
*** H3y has joined #openstack-lbaas15:05
*** ajmiller has joined #openstack-lbaas15:16
TrevorVsbalukoff I don't expect you to be around just yet, but when you jump online can you hit me up?  I want to chat with you a bit about the shared pool single-create situation15:17
*** fnaval has joined #openstack-lbaas15:18
*** mestery has quit IRC15:19
*** mestery has joined #openstack-lbaas15:19
*** manishg has joined #openstack-lbaas15:29
*** numans has quit IRC15:47
*** armax has joined #openstack-lbaas15:50
openstackgerritIhar Hrachyshka proposed openstack/octavia: glance: support relying on tags to extract image id
ihrachysbharathm: thanks for reviews on glance tag patch, and sorry for the churn: spending most of my time in neutron code, I already managed to forget how to manage config files manually ;)15:56
bharathmihrachys: Heh.. Happens sometimes. Thanks for fixing it. Looks good now16:00
*** nmagnezi has quit IRC16:05
openstackgerritEvgeny Fedoruk proposed openstack/neutron-lbaas: L7 support for Radware LBaaS v2 driver
*** Purandar has joined #openstack-lbaas16:13
*** rcernin has quit IRC16:16
*** fawadkhaliq has quit IRC16:17
*** pcaruana has quit IRC16:17
*** Aish has joined #openstack-lbaas16:19
*** neelashah has quit IRC16:20
*** jwarendt has joined #openstack-lbaas16:22
*** Purandar has quit IRC16:23
*** neelashah has joined #openstack-lbaas16:33
*** sbalukoff has quit IRC16:38
*** bana_k has joined #openstack-lbaas16:38
*** Purandar has joined #openstack-lbaas16:41
*** klindgren has quit IRC16:44
*** manishg has quit IRC16:52
*** manishg has joined #openstack-lbaas16:53
johnsomTrevorV ping16:53
TrevorVjohnsom pong16:54
johnsomWhy do we need the peer port task?16:54
*** yamamoto_ has quit IRC16:54
*** yamamoto has joined #openstack-lbaas16:55
TrevorVHonestly johnsom I'm not sure I'm the right person to answer that... That's something i pulled from the review earlier that blogan wrote, so I'd have to really look at what it was doing... sorry (my fault)16:56
johnsomOh, ok.  sbalukoff changed this recently, so maybe that can go away16:56
bloganTrevorV, johnsom: i don't remember honestly, so it may not be needed16:56
bloganif i did it, i'm sure i had a good reason, but i also could have just been crazy16:57
TrevorVWell we know you cray16:57
johnsomCool, well, that will remove some comments... grin16:57
TrevorVI'll delete that task and anywhere its referenced, and see if it 'splodes real quick16:58
johnsomAs for the other "discussion", I think the try/catch blocks would be good, just so we don't halt the revert.  The repo updates are an enhancement and can be done later if we open a bug (which I can do for you if you would like16:58
TrevorVrepo updates?16:59
johnsomYeah, have one repo call to update all of the listeners on an LB instead of looping and opening multiple db connections16:59
TrevorVIsn't it all done via one session though?16:59
TrevorVso X queries in one session?  Did I misread that bit?17:00
*** kobis has quit IRC17:00
*** sbalukoff has joined #openstack-lbaas17:00
johnsomYes, I meant DB calls17:00
johnsomHaven't had all of by coffee rations for the morning17:00
TrevorVjohnsom that's fine ha ha, but then doesn't that mean that its behaving the way we should want?  If one of the calls in the session were to fail, it would revert, which is "expected" behavior, right?17:01
TrevorVwouldn't it be best to not have one listener ACTIVE and the others "ERROR" just because we had a failure updating the "ACTIVE" one in the DB?17:01
johnsomAn exception inside the revert itself would halt the revert17:01
TrevorVBut the session won't update the rows until its exited, right?17:02
*** neelashah has quit IRC17:02
TrevorVSo an exception would trickle up, and close the session, updating the rows to all "ACTIVE" which *may* be better than spotty "ERROR" here and there.17:02
TrevorVEspecially when the one at fault WASN'T marked as ERROR, potentially17:02
TrevorVKeep in mind, I might be overthinking this, and trusting too much on SQL Alchemy, I'm just talking right now, you know?17:03
johnsomNot 100% sure I follow you, but let's take the mark LB active revert() method as an example.  If we get an exception, during that loop setting the listeners to ERROR, it would halt the revert.  So the prior tasks in that flow will not have their revert() called.  This could mean that the LB itself would not get marked in ERROR because one listener did not update right.17:05
*** neelashah has joined #openstack-lbaas17:05
johnsomAt least I think that is what happens when the revert itself gets an exception17:06
*** yamamoto has quit IRC17:10
*** amotoki has quit IRC17:13
*** kevo has joined #openstack-lbaas17:15
TrevorVjohnsom I see your point.  I'm adding a try-catch right now, and am about to go to lunch, so I'll toss up the change here in a second and then be AFK17:17
*** openstackgerrit has quit IRC17:18
*** openstackgerrit has joined #openstack-lbaas17:19
xgermanTrevorV thanks — I think each call to the repo opens its own DB session… so they will all be committed17:20
TrevorVjohnsom also removing the peer ports task, and running tox.17:20
sbalukoffMajor pet peeve of mine: Waiting until the day of a major deadline to point out what are essentially cosmetic errors and -1'ing an important patchset because of them.17:20
TrevorVxgerman yeah that's what johnsom was alluding to.17:20
TrevorVIts something I didn't follow until I looked and saw the "with" context in the DB method17:20
TrevorVIt made more sense there.17:20
sbalukoffCould someone please throw some +1's at these:
johnsomsbalukoff re-adding a task to allocate peer ports is not cosmetic17:21
*** evgenyf has joined #openstack-lbaas17:21
sbalukoffjohnsom: I wasn't talking about that. XD17:21
sbalukoffI was talking about the python-neutronclient comments that came in last night.17:21
johnsomAh, ok.  Just feeling the hate17:21
*** jschwarz has quit IRC17:21
xgermanpython-neutronclient hates everybody - ask jwarendt17:21
sbalukoffNaw-- you and xgerman make good points in your comments on that review.17:22
johnsomYeah, neutronclient can be slow going17:22
TrevorVYeah, they were valid concerns.  They could have waited for bug stuffs, but only because I'm afraid of the gate right now o_017:22
TrevorVha ha17:22
xgermanwell, we can always do that...17:23
*** kevo has quit IRC17:23
TrevorVIts fine, I'm about to push up the changes.17:23
TrevorVWe'll hoep for the best.17:23
*** bana_k has quit IRC17:24
xgermanthanks… my delete-lb stuff is moving slow as well… but I am secretly hoping for an extension ;-)17:24
openstackgerritTrevor Vardeman proposed openstack/octavia: Get Me A Load Balancer Controller
TrevorVjohnsom xgerman sbalukoff new patch up.  Take a gander if you don't mind.17:25
TrevorVHopefully I didn't skip one of the suggestions... o_017:25
TrevorVbrb, goin to lunch17:25
sbalukoffWell... when it comes to L7 code, the CLI is actually the least important thing to get in before the deadline; But this is still important to get merged today if at all possible:  Could one or more of y'all neutron-lbaas cores please have a look?17:26
sbalukoffTrevorV: Having a look now.17:26
openstackgerritGerman Eichberger proposed openstack/neutron-lbaas: Adds Cascade option for load balancer drivers
*** Aish has quit IRC17:28
xgermanTrevorV you forgot a try..catch :-(17:32
sbalukoffxgerman: +117:34
sbalukoffIf we're going to wrap one, we should wrap them all. :)17:34
*** fawadkhaliq has joined #openstack-lbaas17:36
*** Aish has joined #openstack-lbaas17:40
*** bana_k has joined #openstack-lbaas17:43
*** manishg has quit IRC17:44
*** ihrachys has quit IRC17:45
*** ajmiller has quit IRC17:50
sbalukoffWoo-hoo! Thank you johnsom for the +A on the neutron-lbaas L7 patch!17:51
sbalukoffxgerman: Checking out your latest patchset now.17:52
johnsomThe issue I had with it was fixed, so looked good to go17:52
*** ajmiller has joined #openstack-lbaas17:52
xgermansbalukoff that only removed WIP17:56
sbalukoffjohnsom: Yeah, evgenyf and I went through a bunch of back-and-forth revision and feedback cycles over the last few days ironing out all the bugs I was able to find.18:00
sbalukoffxgerman: Aah, good to know.18:00
*** madhu_ak|away has joined #openstack-lbaas18:04
*** madhu_ak|away is now known as madhu_ak18:06
*** kevo has joined #openstack-lbaas18:06
*** H3y has quit IRC18:06
openstackgerritMerged openstack/neutron-lbaas: L7 capability extension implementation for lbaas v2
* sbalukoff pops the cork out of the champagne bottle.18:15
*** yamamoto has joined #openstack-lbaas18:16
*** kobis has joined #openstack-lbaas18:18
*** yamamoto has quit IRC18:21
* johnsom worries that we drove sbalukoff to drink....18:29
sbalukoffHaha! Oh, I have contemplated it maaaany times this last month. XD18:29
*** Purandar has quit IRC18:30
*** doug-fish has quit IRC18:32
*** madhu_ak has quit IRC18:34
*** Purandar has joined #openstack-lbaas18:35
*** madhu_ak has joined #openstack-lbaas18:35
*** evgenyf has quit IRC18:42
*** doug-fish has joined #openstack-lbaas18:46
*** nmagnezi has joined #openstack-lbaas18:46
*** doug-fish has quit IRC18:46
dougwigjust landed.  :)  is the l7 etherpad up to date?18:46
dougwigwent to look at the extension, it's in.  sweet.18:46
*** doug-fish has joined #openstack-lbaas18:47
*** madhu_ak_ has joined #openstack-lbaas18:49
TrevorVxgerman I'm uploading that one here in a second...18:50
TrevorVMy bad18:50
xgermandougwig, now if you can have a look at the cascading delete stuff18:51
dougwigwas going to look at l7 client side first.18:51
openstackgerritTrevor Vardeman proposed openstack/octavia: Get Me A Load Balancer Controller
fnavalreview requested on this please:
TrevorVxgerman updated18:53
*** madhu_ak has quit IRC18:54
*** neelashah has quit IRC18:55
openstackgerritAndreas Jaeger proposed openstack/neutron-lbaas: Remove unused pngmath Sphinx extension
sbalukoffdougwig: Thanks for the +2 on the shared pools patch!18:57
sbalukoffOk, so talking with armax, it appears we need OpenStack doc updates to consider L7 "feature complete". Apparently we have a few more days to get that done.18:58
sbalukoffSo, I'mma finish a couple more reviews here and then dive into that.18:58
*** fnaval_ has joined #openstack-lbaas18:58
*** fnaval has quit IRC19:01
*** neelashah has joined #openstack-lbaas19:02
dougwigsbalukoff: yep.  ping Sam-I-Am about what they need.  might be they just need the raw content, or he might want you to take a stab at the real docs.19:03
*** fnaval_ has quit IRC19:03
*** doug-fis_ has joined #openstack-lbaas19:03
dougwigsbalukoff: he's over in the neutron channel.  docs guy.19:04
sbalukoffOh, sweet, he's an IBMer. I can contact him in ways he can't ignore!19:04
*** fnaval has joined #openstack-lbaas19:04
* sbalukoff cackles maniacly.19:04
dougwigit's ok, Matt's good at biting back.  :)19:04
sbalukoffIBM had no idea what it was doing by acquiring my employment along with Blue Box...19:05
*** doug-fi__ has joined #openstack-lbaas19:06
*** doug-fi__ has quit IRC19:06
*** doug-fi__ has joined #openstack-lbaas19:07
*** doug-fish has quit IRC19:07
*** doug-fis_ has quit IRC19:08
*** neelashah has quit IRC19:09
*** neelashah has joined #openstack-lbaas19:10
*** kobis has quit IRC19:13
nmagnezidougwig, hi :) so followed the instructions we discussed (gate hook for haproxy). despite everything i do i get: WARNING neutron_lbaas.drivers.haproxy.namespace_driver [-] Stats socket not found for loadbalancer19:15
*** doug-fi__ is now known as doug-fish19:16
*** ducttape_ has quit IRC19:16
sbalukoffI am being told by sam-i-am that neutron-lbaas needs release notes and that we should move all our wiki documentation into the official manual, apparently in the openstack-manuals repo.19:18
sbalukoffHave we ever done release notes in neutron-lbaas before?19:18
*** ihrachys has joined #openstack-lbaas19:20
sbalukoffBah. Whatever. I'll write a neutron-lbaas patch for the release notes later today.19:22
sbalukoffWe have a few days to get the manual updates in.19:22
johnsomRelease notes is on my list to go learn.  We have not been using them to my knowledge19:22
sbalukoffjohnsom: Just had to do them for the CLI updates... let me find the (haha!) wiki doc describing how to use them.19:23
johnsomI think we are missing gate jobs for the release notes stuff too.  Thanks sbalukoff19:23
*** ducttape_ has joined #openstack-lbaas19:23
sbalukoffMy bad...19:24
sbalukoffI guess that is in the official docs.19:24
sbalukoffIn any case... I see what I can do to make that work in neutron-lbaas today.19:25
sbalukoffEr.. I'll see...19:25
johnsomWhile you are at it, setup Octavia too19:25
sbalukoffIf I can.19:27
openstackgerritMerged openstack/neutron-lbaas: fix mismatch between v1 keyword and v2 driver in gate_hook
johnsomsbalukoff there is also this:
sbalukoffHow old is that etherpad?19:34
sbalukoffBecause the doc I linked describes using reno...19:34
johnsomOld, November-ish19:34
johnsomYeah, someone pointed me to the etherpad for background.  Haven't had time to look at it yet19:34
sbalukoffYeah, thanks for the link!19:35
sbalukoffI know people hate writing docs... but if you want people to actually use the code you spent months perfecting... they ain't unless it's well documented how to use it. :/19:35
*** bana_k has quit IRC19:42
*** bana_k has joined #openstack-lbaas19:43
sbalukoffI'm grilling sam-i-am on the all important graphic format question.19:43
*** ihrachys has quit IRC19:44
sbalukoffThe answer I'm getting is pretty much "everything sucks, so just commit binary .pngs, and if you used a proprietary source file to generate it, commit that too."19:44
sbalukoffWhich is... sad.19:44
madhu_ak_rm_work, ping19:44
*** evgenyf has joined #openstack-lbaas19:44
sbalukoffApparently most people use OmniGraffle or
*** madhu_ak_ is now known as madhu_ak19:44
sbalukoffAlso apparently .svg's have the same inconsistent rendering problem that GraphViz files do. :P19:45
sbalukoffWell... I guess that at least makes things easier. :P19:45
dougwigsbalukoff: we used to relnote in the wiki, like everyone else. mestery added reno support, and i've seen a few folks add reno notes with their features. it's likely not complete.19:56
mestery++ to reno19:56
*** armax has quit IRC19:59
*** Guest71383 is now known as redrobot20:01
*** neelashah has quit IRC20:02
*** neelashah has joined #openstack-lbaas20:07
*** hockeynut_afk is now known as hockeynut20:13
*** fawadkhaliq has quit IRC20:22
*** Purandar has quit IRC20:23
*** doug-fish has quit IRC20:24
*** nmagnezi has quit IRC20:26
*** doug-fish has joined #openstack-lbaas20:30
sbalukoffdougwig and mestery:  Could I get you to take another look at this?  (Rebase after merge conflict + update to release notes as requested by Akihiro)20:32
*** doug-fis_ has joined #openstack-lbaas20:32
xgermandoug-fish yt?20:33
*** mixos has joined #openstack-lbaas20:34
*** doug-fish has quit IRC20:34
*** doug-fis_ has quit IRC20:37
*** neelashah1 has joined #openstack-lbaas20:40
*** madhu_ak has quit IRC20:41
*** madhu_ak has joined #openstack-lbaas20:42
*** neelashah has quit IRC20:43
*** fnaval has quit IRC20:43
TrevorVxgerman alright, the last for-loop listener update I could find is updated...20:44
*** fnaval has joined #openstack-lbaas20:44
xgermansorry, our code is  a mess (hence didn’t give a -1 in case we need to rush it through)20:45
*** bana_k has quit IRC20:46
TrevorVxgerman you mean on my review just now?20:46
*** bana_k has joined #openstack-lbaas20:47
madhu_akrm_work, you there?20:48
*** ducttape_ has quit IRC20:50
openstackgerritTrevor Vardeman proposed openstack/octavia: Get Me A Load Balancer Controller
*** neelashah has joined #openstack-lbaas20:52
*** neelashah1 has quit IRC20:52
rm_workmadhu_ak: yes20:54
rm_workbut barely, whats up20:55
madhu_akso what was the status of this thread? ? I see you responded to that thread, bcoz I am reproducing the error20:58
madhu_akrm_work ^^20:58
*** chlong_ has joined #openstack-lbaas21:00
*** neelashah1 has joined #openstack-lbaas21:01
*** manishg has joined #openstack-lbaas21:02
*** ducttape_ has joined #openstack-lbaas21:02
*** neelashah has quit IRC21:03
*** Purandar has joined #openstack-lbaas21:03
mesterysbalukoff: Done!21:04
sbalukoffmestery: Sweet!  One more for you:
sbalukoffAnd thanks!21:05
sbalukoff(I hope that Akihiro is still online...)21:06
TrevorVxgerman sbalukoff blogan johnsom I have the review up to date with try/catches.  Let me know if I missed anything that wasn't "bug-report"-able for the near future.21:07
xgermanI thought you have now two +2s on it21:07
*** minwang2 has joined #openstack-lbaas21:10
*** crc32 has joined #openstack-lbaas21:10
*** neelashah has joined #openstack-lbaas21:10
*** jpomeroy has joined #openstack-lbaas21:10
*** Aish has quit IRC21:11
jpomeroyajmiller, for LBaaS v2 the listener and pool protocols need to match?21:11
TrevorVxgerman I saw that, but that doesn't mean +A, which means we have to make sure gate gets them o_021:11
TrevorVjohnsom are you swamped right now?21:11
*** Aish has joined #openstack-lbaas21:12
sbalukoffTrevorV: Will have another look right now.21:12
TrevorVsbalukoff thanks21:12
johnsomTrevorV no, just working through reviews21:12
TrevorVjohnsom I was just trying to make sure I had all the bugs listed that we identified.21:13
TrevorVSo we had 2 missing unit tests21:13
*** neelashah1 has quit IRC21:13
TrevorVThen we had "shared pools support in single-create"21:14
TrevorVWhich I want to discuss a little before that gets figured in21:14
TrevorVI also need to get the documentation updated21:14
TrevorVWas that everything?21:15
johnsomI think that is what I know of.21:15
johnsomMaybe an RFE for the repo update to do the listener updates in one DB call.21:15
TrevorVjohnsom I actually want to talk about the pros/cons of that behavior, if you guys don't mind entertaining that point as well.21:16
TrevorVAlright, so let me figure out how to report a bug, ha ha, and then we can get to talkins21:17
TrevorVYeah, I was just doing that, just was taking me a second to make sure of the steps.  sorry, new to this21:20
openstackLaunchpad bug 1551427 in octavia "Missing branch-coverage unit tests" [Undecided,New] - Assigned to Trevor Vardeman (trevor-vardeman)21:23
TrevorVdoes that look right?21:23
TrevorVAlright, so the discussion with the database objects, and anyone should interject if they like ( xgerman sbalukoff you guys specifically had comments about it in the review)21:26
johnsomrm_work this one too:  (I am assuming you would like a chance to look at these)21:26
TrevorVOne side of the argument, johnsom, is when you combine all the updates into one call, and then that call fails, would it solve anything that doing a singular update at a time won't?21:27
johnsomOther than save database round-trips, no.21:27
TrevorVI'm also concerned with diagnostics.21:27
xgermanand not switches things to active21:27
jpomeroyanyone know?  LBaaS v2 the listener and pool protocols need to match?21:28
jpomeroygetting errors if we don't set them to the same thing21:28
xgermanjpomeroy no21:28
TrevorVxgerman but a roll-back would also not update the status21:28
TrevorVI mean a DB roll-back21:28
*** manishg has quit IRC21:28
xgermanjpomeroy Terminated TLS/HTTP21:28
sbalukoffjpomeroy: Not exactly. HTTP pool can work with HTTP or TERMINATED_HTTP listener.21:28
xgermanshould ork21:28
sbalukoffjpomeroy: But HTTP listener doesn't work with TCP pool.21:29
xgermanTrevorV right21:29
*** manishg has joined #openstack-lbaas21:29
xgermanbut things would not switch to the next state21:29
TrevorVI'm confused, what stance are you taking?  It should all be one call or it shouldn't?21:30
sbalukoffHonestly, I'm pretty sure most of our revert code needs a good overhaul anyway:  I'm pretty sure we are missing opportunities for more graceful failure modes, and probably not doing the right thing in some cases...21:30
xgermanit should be one call. johnsom’s argument regarding performance is the main one21:30
TrevorVSure, but doesn't that potentially muddy up diagnostics?21:30
sbalukoffEspecially once we get job-board, maybe it would be better to queue up revert DB calls if they fail to be retried again at some point?21:30
johnsomsbalukoff this isn't for revert, but for the execute path21:31
TrevorVFor example, how would we distinguish the failed entry into the DB given the arguments?21:31
sbalukoffjohnsom: Aah.21:31
TrevorVIf that one call fails, we won't know a "culprit" so to speak21:31
TrevorVIf we do each update individually, it can make things more clear, can it not?21:31
xgermanwell, we implicitly decided that if I create/update two listeners one succeeds and one fails - both need to be rolled back21:31
*** neelashah has quit IRC21:32
TrevorVSure, "rolled back" being relative here where all we do is update a status.21:32
xgermanI would like that to be explicit: no partial success21:32
johnsomAt least not the code I am thinking of.  In the execute() we are looping and sending DB updates for each listener on an LB, where I argue we can save DB round trips with a smarter repo call that updates them all to Active21:32
xgermanyep, if we need to do single calls I like each update be it’s own task21:32
TrevorVjohnsom sure, I'm getting at diagnostics if that should fail21:32
xgermanwell, for diagnostics it would be best to have single tasks and not loop inside a task21:33
sbalukoffxgerman: +121:33
TrevorVIf you make a potentially large query to update a list of objects, can we easily decipher the issue if one specific thing fails?21:33
*** manishg has quit IRC21:33
sbalukoffBut if those tasks are executed in some sequence, that doesn't save us DB round-trip.21:34
sbalukoffSo, they'd need to be unordered, essentially.21:34
johnsomTrevorV I don't think it makes much difference.  It's a DB call failure either way.  The content may be different, but basically the same result.21:34
TrevorVsbalukoff unordered or otherwise its still the same number of executions to the DB based on the number of objects in the list.21:34
sbalukoffI agree21:34
johnsomYeah, unordered doesn't buy much here.21:35
sbalukoffAnd a failure would result in the same kind of revert, right?21:35
sbalukoff(A failure in any one of them, task or not.)21:35
TrevorVjohnsom I'm only trying to play devil's advocate here.  I'm okay with the iterations, and I'm okay with a large update.  I just want to make sure we vet each option to find the least of all issues.21:35
TrevorVI'm mostly concerned about error diagnostics providing false positives given each situation21:35
johnsomYeah, no problem21:35
*** Purandar has quit IRC21:36
TrevorVOn the hand where we do individual updates, and one of them fails, so the task reverts, AND one of the revert try/catches fail, then we have x listeners marked ERROR, and y listeners marked "ACTIVE", so we start figuring out why they're still active.21:36
TrevorVIf we do a batch-update to the DB, we have ALL listeners "ACTIVE" and have to look at every single one individually most likely to diagnose the issue21:36
xgermanI think if the database repeatedly balks we have bigger problems21:37
TrevorVxgerman i meant where we have half success and half failure21:37
xgermanI can see it being shaky while the DB fails over/etc.21:37
TrevorVLIke lets say we have 4 listeners21:37
*** neelashah has joined #openstack-lbaas21:37
TrevorVif 2 go active and 2 go ERROR, then we have a place to start21:37
xgermanTrevor V understood21:37
TrevorVOh okay, sorry21:37
xgermanbut if we have like more than 20 listeners in that state we have nigger problems21:38
johnsomActually with the single update, you still only have one thing to debug, the DB update call21:38
neelashahxgerman: are there any combination of different listerner and pool protocols allowed? or do they always have to be the same?21:38
bloganyikes xgerman!21:38
TrevorVAlright, I see your point johnsom .  So I guess its fine doing the batch update.21:39
TrevorVHow do you put in an enhancement request?21:39
sbalukoffneelashah: HTTP pool can work with both HTTP and TERMINATED_HTTPS listener.21:40
johnsomTrevorV You open a bug and put the tag RFE on it.21:40
neelashahneelashah: ah, ok got it21:41
neelashahsbalukoff: got it, thanks21:41
*** Purandar has joined #openstack-lbaas21:42
*** manishg has joined #openstack-lbaas21:46
TrevorVjohnsom assign it to myself probably?  or should I leave it for *anyone* to update?21:46
johnsomI would leave it open unless you are going to start working on it.  It's an RFE and not a bug against your code necessarily21:47
TrevorVRight, gotcha.  I'll leave it open for now.21:47
openstackLaunchpad bug 1551436 in octavia "Cascading LB delete is missing documentation updates" [Critical,New] - Assigned to German Eichberger (german-eichberger)21:47
sbalukoffblogan or rm_work or dougwig: Could one of you have a look at this and decide whether you're OK with +A on it?
TrevorVsbalukoff if that gets a +A, it might affect single-create... ha ha ha21:48
sbalukoffOh, no!21:49
TrevorVThen again, if single-create gets a +A, IT might affect THIS21:49
TrevorVThat seems... really bad21:49
sbalukoffWell, we're trying to get both single-create and single-destroy in today, right?21:49
sbalukoffSomeone's going to have to rebase. :/21:49
blogansbalukoff, xgerman: what's calling that controller worker method?21:49
sbalukoffblogan: Dependent patch.21:49
TrevorVI guess whoever gets the short straw gets the short straw, but with gate jobs taking 3-7 hours, one of them won't get in.21:49
*** fawadkhaliq has joined #openstack-lbaas21:50
ajmillerjpomeroy -- sorry, didn't see your ping.  It looks like you got your question answered?21:50
sbalukoffWell... I'm in the PST timezone...  it'll still be Feb. 29 here for a while. :)21:50
jpomeroyajmiller, yes thank you21:50
sbalukoffAlso, any of the other Octavia cores have a problem with merging whichever patch gets the short straw, if the only change is that it needs a rebase?21:51
sbalukoffTrevorV and xgerman: Other option is for one of you to rebase behind the other's patch.21:52
sbalukoffFrom my perspective, both of these new features are essentially ready.21:52
TrevorVI agree.21:52
TrevorVIf the cascade delete changes are already tested right now, then we should +A them asap so I can get that code now21:53
TrevorVIf we can do that.21:53
*** armax has joined #openstack-lbaas21:53
TrevorVThen we might be able to +A my stuff this evening.21:53
johnsomI'm reviewing the second delete now.  Though per-protocol I should give one of the other cores the +A21:54
sbalukoffOk, let me have another look at xgerman's second patch that is part of cascade-delete...21:54
sbalukoffOk, I already +2'ed it.21:57
sbalukoffblogan or rm_work or dougwig: Need one of y'alls stamp of approval on this:
sbalukoffYeah, both these patch sets are just waiting on a last +2/+A to merge.21:59
sbalukoffAnd yes, they conflict...  so TrevorV and xgerman: Y'all should work it out amongst yourselves which of you is going to rebase.22:00
TrevorVThe only question is who is waiting on zuul22:01
TrevorVxgerman are you waiting on zuul for the deletes?22:01
sbalukoffNote that xgerman's total line count is around 250, whereas TrevorV's is around 1200...22:02
sbalukoffOk, blogan just +A'ed xgerman's patch.22:02
johnsomWell, I'm equal opportunity, xgerman missed a unit test in
sbalukoffSo that one will get in first.22:02
sbalukoffTrevorV: Get to rebasing, man! :D22:03
TrevorVjohnsom are you not gonna just bug that and let it fly?22:03
TrevorVUgh... Alright :P22:03
bloganfull disclosure, i didn't test it out, i just looked over the code to make sure it made sense, but bugs can be fixed later22:03
johnsomTrevorV I would, but I am going to flag them as I see them and give the developer the option22:03
TrevorVjohnsom alright, well it looks like we're getting the +A on cascade delete, so bug-reporting is a necessity at this point.  Is xgerman in a meeting or something right now?22:04
johnsomblogan beat me on my -122:04
sbalukoffblogan: FWIW, I did test it out yesterday. Didn't see the code change significantly in the mean time.22:04
johnsomHe ran to lunch22:04
*** Purandar has quit IRC22:04
sbalukoffTrevorV: Ping me as soon as you've got a rebased patch set up there, and I'll run it through a quick test to make sure there aren't any unexpected complications from the rebase.22:05
TrevorVsbalukoff alright, will do.22:05
TrevorVI may be travelling home here in a second (to beat traffic) but it'll be up in the next couple hours ideally22:06
sbalukoffIn the mean time... lunch sounds like an excellent idea.22:06
*** doug-fish has joined #openstack-lbaas22:06
sbalukoffTrevorV: I'll be sticking around to see this one get through, so long as we have at least one more core willing to +2.22:06
TrevorVsbalukoff awesome, I'll be online to see that double +2 as well.  That's the best thing about working on a laptop while being able to go home to my tower :D22:07
*** Purandar has joined #openstack-lbaas22:08
*** TrevorV has quit IRC22:09
*** neelashah1 has joined #openstack-lbaas22:09
xgermansorry, was out for a bit. Thanks for merging my stuff...22:10
*** neelashah has quit IRC22:11
johnsomxgerman you picked up some bugs22:12
xgermanhow did I make so many mistakes in such a small code base?22:13
*** neelashah has joined #openstack-lbaas22:13
johnsomblogan sbalukoff When you get a chance:
*** neelashah1 has quit IRC22:16
sbalukoffWill check it out, eh...22:17
*** armax has quit IRC22:19
*** neelashah1 has joined #openstack-lbaas22:24
*** neelashah has quit IRC22:24
*** armax has joined #openstack-lbaas22:26
*** crc32 has quit IRC22:27
bloganjohnsom: that failed all scenario tests22:30
sbalukoffjohnsom, xgerman: Any merit to the nit I just added here, do you think?
rm_worksbalukoff: ok so we're going to try to merge single-create later tonight?22:30
rm_worksbalukoff: i'll be around if i need to be, around 11pm/midnight22:30
sbalukoffrm_work: I hope so, yes.22:30
rm_workor rather, i WILL be around 11pm/midnight :P22:30
rm_workregardless of whether i need to be... <_<22:30
bloganso for some reason the controller worker could not talk to that instance with the glance tags review22:31
bloganfor every test22:31
bloganthat ahs be concerned22:31
bloganhas me22:31
*** mixos has quit IRC22:31
sbalukoffblogan: +122:32
johnsomblogan I see some that connected22:32
bloganjohnsom: ah good point, but still all scenario tests failed22:33
johnsomHmm, but yeah, the main log looks bad22:33
johnsomI think computewait would have bombed if the instance didn't boot because of a messed up image ID22:35
*** neelashah1 has quit IRC22:35
bloganjohnsom: yeah i misread the logs, they all went active it looks like, except that last one22:35
*** doug-fis_ has joined #openstack-lbaas22:35
johnsomIt shouldn't get to the attempting to connect22:35
bloganwhich it just looks liek the logs stopped22:35
*** neelashah has joined #openstack-lbaas22:36
*** doug-fish has quit IRC22:37
*** manishg_ has joined #openstack-lbaas22:39
*** manishg has quit IRC22:39
*** neelashah has quit IRC22:43
openstackgerritMerged openstack/octavia: Improved LoadBalancer Delete
*** rtheis has quit IRC22:43
*** doug-fis_ has quit IRC22:48
openstackgerritMerged openstack/octavia: Cascading LB delete
*** doug-fish has joined #openstack-lbaas22:48
*** doug-fish has quit IRC22:53
*** doug-fish has joined #openstack-lbaas22:53
xgermanNow LBaaS V2… I think ajmiller was looking at that22:58
ajmillerxgerman Yes, I read through them this morning.23:00
ajmillerbeen mostly focussed on our internal stuff all that though.23:00
xgermanok, sounds good23:01
ajmillerblogan, how do you feel about xgerman's reply to your -1 on
*** jpomeroy has quit IRC23:04
*** TrevorV has joined #openstack-lbaas23:04
xgermanthanks for bringing that up23:05
TrevorVxgerman did cascade delete merge?23:06
TrevorVGonna rebase my changes on master and push change23:07
*** yamamoto_ has joined #openstack-lbaas23:11
*** yamamoto_ has quit IRC23:16
bloganajmiller: i feel xgerman should never disagree with me23:17
TrevorVsbalukoff xgerman johnsom blogan rm_work ha ha all the tags... Sooo it was a super simple rebase, that's the good news.  Its about to go up in a review.23:17
johnsomI disagree with that23:17
blogani also feel johnsom should never disagree with me either23:17
openstackgerritHenry Gessau proposed openstack/neutron-lbaas: Track alembic heads
openstackgerritTrevor Vardeman proposed openstack/octavia: Get Me A Load Balancer Controller
openstackgerritTrevor Vardeman proposed openstack/octavia: Get me a Load Balancer API
TrevorVjohnsom you disagree with it being a simple rebase?23:19
xgermanhe agrees to disagree23:20
sbalukoffTrevorV: Restacking to try out your new, rebased patches now!23:23
* TrevorV is scared23:23
sbalukoffWell, now that xgerman's cascade delete patch is merged... at this point it'll make for easy clean-up. ;)23:25
bloganxgerman: could you make an api tempest test for the cascade delete?23:26
bloganxgerman: in a follow up patch23:26
xgermanOctavia or LBaaS V223:26
bloganlbaas v223:27
xgermansure I can try - never made a tempest test… but it’s always a first23:27
xgermanalso cascading delete is the same as delete so should be copy&paste23:28
bloganxgerman: well your assertions would be different23:28
bloganalso, just so i'm clear, this will do a cascading delete if the dirver supports it, always?23:29
xgermanyes, that’s what we agreed at the midcycle23:29
bloganlike a DELETE /loadbalancers/{lb_id} will cascade delete always?23:29
xgermanyep, if the driver supports it23:29
bloganyeah i kind of remember that, but still makes me feel uneasy, but meh, it should be how the api was all along23:29
*** manishg_ has quit IRC23:30
bloganmeaning i wish we just had /loadbalancers endpoint, thats it and just ahd get me a lb and cascade delete23:30
*** manishg has joined #openstack-lbaas23:30
bloganbut with shared pools, we'ed need a /loadbalancers/{lb_id}/pools endpoint or a /pools endpoint23:30
xgermanlabs v3?23:30
bloganxgerman: i don't think we have api tests that try to delete a load balancer with children, if we did, those should ahve failed23:32
sbalukoffOctavia does have a loadbalancers/{lb_id}/pools endpoint.23:32
xgermanblogan, ok. I can add that as a test23:33
* xgerman is picking up lots of work today23:33
sbalukoffHeh! I know the feeling.23:33
sbalukoffOh... sorry, missed the context on that...23:34
openstackgerritEvgeny Fedoruk proposed openstack/neutron-lbaas: L7 support for Radware LBaaS v2 driver
blogansbalukoff: yea i was talking about neutron-lbaas23:36
bloganxgerman: mind if i rebase that patch?23:37
bloganor if you want to23:38
sbalukoffblogan: How would updates to objects subordinate to a loadbalancer work in that case?23:38
xgermanblogan, go ahead23:39
blogansbalukoff: probably just update by doing /loadbalancers/{lb_id}/listeners23:40
blogansimilar to octavia23:40
sbalukoffMakes sense.23:40
bloganwhich is really where we could take octavia's api23:41
sbalukoffIt's very similar in most other ways already.23:41
*** yamamoto_ has joined #openstack-lbaas23:42
openstackgerritBrandon Logan proposed openstack/neutron-lbaas: Adds Cascade Delete for LoadBalancers to Octavia Driver
TrevorVsbalukoff at the progress?!23:44
sbalukoffStill stacking... this machine is pretty slow.23:45
TrevorVAll good, just curious :D23:45
*** manishg has quit IRC23:49
*** fnaval has quit IRC23:52

Generated by 2.14.0 by Marius Gedminas - find it at!