Friday, 2015-05-01

openstackgerritOpenStack Proposal Bot proposed openstack/neutron-lbaas: Merge tag '2015.1.0'
*** SumitNaiksatam has joined #openstack-lbaas00:24
*** madhu_ak has quit IRC00:26
*** hitalia has quit IRC00:31
*** madhu_ak has joined #openstack-lbaas00:39
*** clev-away is now known as clev00:44
*** mwang2_ has quit IRC00:44
*** clev is now known as clev-away01:00
openstackgerritAl Miller proposed stackforge/octavia: Add devstack plugin for octavia
*** Aish has quit IRC01:25
*** ajmiller has quit IRC01:25
*** bharath has quit IRC01:28
*** madhu_ak has quit IRC01:28
*** fnaval has quit IRC01:51
*** fnaval has joined #openstack-lbaas02:07
*** clev-away is now known as clev02:27
*** sbalukoff has joined #openstack-lbaas02:52
*** hitalia has joined #openstack-lbaas03:04
bloganquery ptoohill03:07
rm_you|query blogan03:08
*** rm_you| is now known as rm_you03:08
bloganquery rm_you|03:08
bloganchanged your name03:08
*** clev is now known as clev-away03:09
ptoohillquery blogan03:10
*** clev-away is now known as clev03:18
*** SumitNaiksatam has quit IRC03:59
*** SumitNaiksatam has joined #openstack-lbaas04:00
*** bharath has joined #openstack-lbaas04:29
*** bharath has quit IRC04:34
*** SumitNaiksatam has quit IRC04:51
*** SumitNaiksatam has joined #openstack-lbaas04:53
*** clev is now known as clev-away04:57
*** clev-away is now known as clev05:11
*** clev is now known as clev-away05:28
*** rm_work|away is now known as rm_work06:05
rm_worksuch query06:06
*** chlong has quit IRC06:21
*** chlong has joined #openstack-lbaas06:35
*** chlong has quit IRC07:02
*** woodster_ has quit IRC07:10
*** chlong has joined #openstack-lbaas07:26
*** chlong has quit IRC07:37
*** chlong has joined #openstack-lbaas07:49
*** chlong has quit IRC07:53
*** chlong has joined #openstack-lbaas07:55
*** chlong has quit IRC08:01
*** chlong has joined #openstack-lbaas08:18
*** chlong has quit IRC08:40
*** chlong has joined #openstack-lbaas10:14
*** hitalia has quit IRC10:14
*** Miouge has joined #openstack-lbaas10:31
*** chlong has quit IRC11:01
*** chlong has joined #openstack-lbaas11:04
*** SumitNaiksatam has quit IRC11:11
*** chlong has quit IRC11:15
*** Miouge has quit IRC11:46
*** Miouge has joined #openstack-lbaas11:48
*** Miouge has quit IRC11:55
*** Miouge has joined #openstack-lbaas11:57
*** woodster_ has joined #openstack-lbaas12:01
*** Miouge has quit IRC12:05
*** clev-away is now known as clev12:25
*** TrevorV|Home has joined #openstack-lbaas12:48
*** clev is now known as clev-away13:00
*** clev-away is now known as clev13:00
*** ajmiller has joined #openstack-lbaas13:10
*** Miouge has joined #openstack-lbaas13:40
*** Miouge has quit IRC13:40
*** Miouge has joined #openstack-lbaas13:41
*** openstackgerrit has quit IRC13:51
*** openstackgerrit has joined #openstack-lbaas13:51
*** fnaval has quit IRC14:14
*** fnaval has joined #openstack-lbaas14:30
ajmillerptoohill blogan I'm about to push a patch to octavia devstack that addresses your comments about the executable paths.14:43
TrevorV|Homeajmiller, they'll be online within an hour or so, but its not typical they get online before that14:52
openstackgerritTrevor Vardeman proposed stackforge/octavia: Update vip and network plug methods
openstackgerritAl Miller proposed stackforge/octavia: Add devstack plugin for octavia
bloganajmiller: ping15:19
ajmillerNothing big, just wanted to touch base on fixing executable paths.  I submitted a patch to that, and am not trying to figure out what I did with phil's code, it looks like I blew away some of his changes with a patch I made last night.15:20
bloganajmiller: why did you remove the api execution in the devstack plugin?15:20
bloganah okay15:21
ajmilleryeah, I think it was a git fubar on my end15:21
ajmillerNothing I intended t odo.15:21
blogangit + gerrit = mass fubars15:21
ajmillerSorry about that.15:22
*** jorgem has joined #openstack-lbaas15:31
ajmillerblogan ptoohill I think the best thing to do is for me to try to put things back like they were before I made my checkin last night, then move forward from there.  Comments?15:35
ajmillerI've fetched that patchset into my repo,15:36
ajmillerand if I push that as a review, the end of the patchset chain should be back to how it was before I messed it up last night.15:37
rm_workyep sounds right15:38
bloganajmiller: thats usually what i do when i've done something like that15:38
blogango to the last known good state and replay the smaller changes15:39
ajmilleryup, ok, thanks.  The next review should put things back like they were at patchset 3215:39
openstackgerritAl Miller proposed stackforge/octavia: Add devstack plugin for octavia
ajmillerblogan can you have a look and double-check that I brought back the right stuff?15:46
bloganajmiller: sure15:46
ajmillerAfter that I'll take a breather and then redo my changes.15:47
bloganajmiller: if all you did was pull down the old patchset and repush it, that'll definitely have those changes15:49
ajmillerThat's what I did, pulled patchset 32, amended the commit so that it has a new hash (otherwise gerrit woudlnt' accept it), and sent it back up.15:49
ajmillerAnd I do see the API startup stuff in there that I didn't write.15:50
ajmillerSo we should be good to go.15:50
bloganyeah i think so15:50
*** mlavalle has joined #openstack-lbaas15:55
*** mlavalle has quit IRC15:57
*** mlavalle has joined #openstack-lbaas15:57
openstackgerritTrevor Vardeman proposed stackforge/octavia: Update vip and network plug methods
*** jorgem has quit IRC16:15
xgermandougwig somehow I can’t vote on
dougwigonly TC members can vote on governance patches.  I'd just leave your opinion in the comments16:24
dougwigthat's what i usually do.16:24
rm_workis it worth noting if you're a core reviewer when commenting?16:28
rm_workis there a "vote" that happens between cores + PTL of new project, or is it just the PTLs that decide?16:28
rm_workdougwig: ^^16:29
xgermanwe had a proposal on the ML16:29
rm_workah, so voting was ther?16:29
xgermansort of16:29
rm_workwell, responded anyway16:29
dougwigthe only "votes" that are counted in that patch is kyle + the TC.  though i expect that if any of us object, that'd be taken notice of.16:30
rm_workinteresting though :P16:33
rm_workcores could all be like "wtf?" and have their project taken over by the TC :P16:33
rm_workbut I guess that kinda makes sense in a weird way16:33
rm_workand yeah, in practice, even non-voting opinions would probably be taken into account16:34
dougwigoh yes, i'm sure anyone here could veto it.16:35
*** ajmiller has quit IRC16:40
*** TrevorV|Home has quit IRC16:52
*** HenryG is now known as floccinaucinihil16:54
*** floccinaucinihil is now known as HenryThe8th16:55
bloganjohnsom: i have some various fixes too, they definitely overla17:09
bloganp with that review17:09
johnsomUmm, ok17:10
bloganjust saying i'll pull yours down and compaire17:10
blogani don't think we need to do the update_amphora return though17:11
bloganif you just pass in amphora_id17:11
johnsomI wondered about that, but your comment said we needed to, so I trusted you...17:11
bloganno i just said we need to update the amphora17:12
bloganor whatever is being passed17:12
bloganreload from db17:12
johnsomI'm not following.  We needed to update the loadbalancer object after we assigned an amphora to it.  That is what this patch does by adding a task the creates a new "updated_loadbalancer" object and swaps it in for the rest of the networking flow17:13
johnsomIf we don't do that the network tasks won't see the amp and wont plug the network in.  If we pass around ids were are querying the DB in each of the network steps.17:14
bloganwhat i mean is taskflow won't allow you to require amphora nad also provide it from teh same task, but you can provide amphora and require amphora_id, which would just overwrite the old amphora, so amphora is now updated17:14
bloganso if all you ever need is the amphora id, then all you need to do is just require amphora_id, and only provide amphora whenever you absolutely need it reloaded, which in that task would17:15
johnsomThat doesn't work for loadbalancer however, since it was injected into the flow at flow load time.  The original loadbalancer object will get used by the subflow17:16
bloganyeah so i just passed in loadbalancer_id, and let a task retrieve it for me and provide, so now it can be overwritten17:16
bloganand you don't have to reload it after every task, just when you know you will need it updated17:16
*** madhu_ak has joined #openstack-lbaas17:17
johnsomI'm not reloading it after every task, only the one time it is updated17:17
ptoohillthank you ajmiller17:18
blogani know but i think it'd be better if you just had amphora and loadbalancer, and just kept overwriting those becuase i don't see a case where a task will need the original amphora and loadbalancer17:18
blogani dont know why taskflow doesn't allow a task to require and provide the same thing though, that would make it simpler, and overwriting store objects17:20
johnsomI agree, you should be able to overwrite them inside the task, probably by accessing the engine storage, but they don't seem to expose that (yet).17:21
bloganyeah but just providing lb_id in the store and providing loadbalancer from a task should get around that and its not terribly hacky17:22
johnsomblogan So, have you coded this up and tested it?  I think I see what you have done by changing some of the tasks.  Also, it would be cool if we keep in sync on the bugs in launchpad so we aren't duplicating effort.17:23
johnsomLet's post it and I'll give it a look and abandon my patch.17:24
bloganjohnsom: actually i was more jsut trying to get it all working, but yeah ive got it successfully completing17:24
bloganjohnsom: no dont abandon yet, it may not be right, but ill post it17:24
bloganhaven't modified the tests though17:25
bloganactually im giong to post the "test" i have for it that creates all the setup resources17:25
johnsomThat's ok17:25
bloganeventually that'll ajmiller's devstack plugin will do all that for the test17:27
johnsomYeah, I have sql scripts and a python test framework that walks through create/update/delete17:28
blogani just have create right now17:28
*** hitalia has joined #openstack-lbaas17:29
*** jorgem has joined #openstack-lbaas17:30
johnsomblogan did you also update the network sub-flow plug network/plug vip stuff?17:32
*** barclaac has joined #openstack-lbaas17:40
bloganjohnsom: yeah17:41
barclaacMorning all.17:41
johnsomOk, how about
openstackLaunchpad bug 1450524 in octavia "Controller worker needs to call allocate vip" [Critical,In progress] - Assigned to Michael Johnson (johnsom)17:41
johnsomblogan I'm trying to decide what to work on next17:41
*** Aish has joined #openstack-lbaas17:42
barclaacQOTD (at least for me anyway ;-) If I have setup with LBaaS v1 and I want to upgrade it to v2 what things in the v1 data model can't be represented in v2?17:42
bloganwhat about breaking it up by loadbalancers listeners pools members and healtmonitors17:43
bloganbarclaac: many to many health monitors to pools17:43
barclaacAnd v2 is a one to many only?17:43
bloganbarclaac: 1:1 right now17:43
barclaacso I could achieve an upgrade by duplicating any health monitors?17:44
bloganjohnsom: what i mean is pick the curd operations in the  for those entities and whomever wants to work on those can call it, we can put a LP bug in17:44
bloganbarclaac: if the pool has many healthmonitors then you'd basically have to choose one, if a health monitor exists on many pools then yeah duplication will work17:45
bloganjohnsom: or we can get even more granular and specify object and cud operation17:46
barclaacGot it. So an upgrade sounds quite feasible then. In the former case I've just got to ask "which one do you want to keep?"17:46
bloganbarclaac: yep, that should be the only change a user sees, as for doing the migration for the operator, thats going to be fun, but doable17:47
bloganbarclaac: actually there might be one more, dougwig inquired about it a while back but i can't remember17:48
barclaacblogan: I could also release note it and tell any v1 people "don't even think of using multiple monitors" to save pain later17:48
blogandougwig do you happen to remember any of the v1 to v2 changes that v1 coudl do but v2 can't?17:48
johnsomblogan I guess I'm not following (or IRC is dropping messages).  I'm looking at the bugs.  Are you talking about what areas we should be testing/fixing?17:48
bloganjohnsom: there aren't any bugs, but yeah put a bug in LP for create lb, and someone tests and fixes that, also have one for update lb, delete lb, create listener, etc17:49
blogani mean there aren't any bugs in LP yet17:49
johnsomYes there are, it's what I've been working against17:50
bloganno i mean the generic bugs im suggesting17:50
blogansorry im thinking too fast and my fingers can't keep up17:50
johnsomOther than the bugs listed, create and update through all of the components work.  I haven't had a chance to test delete yet.17:52
bloganyou've tested create and update for all components?17:52
johnsomWith work arounds for the bugs listed.  After we get those fixed new issues may pop up.17:53
xgermanjohnsom rocks - as you know :-=)17:53
johnsomController worker down to ssh driver to clarify17:54
bloganokay so in that case, thats where some overlap in thsi will happen bc i'd like to test them all out, and if i find issues and fix them they might overlap with what you've done, bc if i find a bug i'm usually just going to fix it right then and there (after talking to you and whomever first)17:54
johnsomYeah, or just pop a bug in LP and claim it17:55
bloganxgerman: yes i know, he's gotten a ton done17:56
*** crc32 has joined #openstack-lbaas18:02
*** bharath has joined #openstack-lbaas18:07
*** mlavalle has quit IRC18:17
*** mlavalle has joined #openstack-lbaas18:17
*** mlavalle has quit IRC18:22
*** mlavalle has joined #openstack-lbaas18:22
dougwigbarclaac: things you can't do: 1) TLS, 2) if you have multiple health monitors associated, you have to pick one.  the rest is straight-forward translation.18:35
barclaacThanks dougwig - TLS was the hook I was going to use to get people to upgrade :-D18:35
*** kobis has joined #openstack-lbaas18:39
*** TrevorV|Home has joined #openstack-lbaas18:40
openstackgerritBrandon Logan proposed stackforge/octavia: Fixed various issues with create lb and amphora flow
bloganjohnsom ^^18:44
johnsomCool, will have a look18:44
*** clev is now known as clev-away18:46
*** clev-away is now known as clev18:47
*** hitalia has quit IRC19:10
*** madhu_ak has quit IRC19:25
ptoohill, I suppose I'm the only one running into these errors, besides maybe Trevor. But would like this to be evaluated.19:40
bloganptoohill: i ran into them too19:43
ptoohillGood deal, does this change sound right?19:45
*** HenryThe8th has quit IRC19:50
*** HenryG has joined #openstack-lbaas19:54
*** HenryG has quit IRC19:55
*** HenryG has joined #openstack-lbaas19:56
*** bharath_ has joined #openstack-lbaas19:58
*** bharath has quit IRC19:58
*** madhu_ak has joined #openstack-lbaas20:07
*** hitalia has joined #openstack-lbaas20:08
*** hitalia has quit IRC20:11
*** TrevorV|Home has quit IRC20:13
*** mwang2 has joined #openstack-lbaas20:15
*** HenryG has quit IRC20:16
*** jorgem has quit IRC20:18
openstackgerritOpenStack Proposal Bot proposed openstack/neutron-lbaas: Updated from global requirements
*** mwang2 has quit IRC20:19
*** TrevorV|home has joined #openstack-lbaas20:20
*** HenryG has joined #openstack-lbaas20:21
openstackgerritMerged stackforge/octavia: Remove extra requirements
openstackgerritMerged stackforge/octavia: Update nova_driver to use amp_network
bloganjohnsom: do you prefer explicit requires in the flow or taskflow inferring from the arguments list? or specify requires in some particular cases?20:29
xgermanI think I went with explicit...20:30
johnsomI have tried to be explicit in an attempt to keep the namespace sane (i.e. not hunting for which task is injecting what)20:30
bloganwell i think we have to explicity with the provides, which is what injects into the namespace20:31
bloganbut yeah explicit is good if only being able to read it from looking at the flows20:32
*** mwang2 has joined #openstack-lbaas20:32
openstackgerritPhillip Toohill proposed stackforge/octavia: Add devstack plugin for octavia
*** hitalia has joined #openstack-lbaas20:41
johnsomblogan If you name the tasks different you can reuse them inside a flow20:49
*** mwang2 has quit IRC20:56
openstackgerritPhillip Toohill proposed stackforge/octavia: Add devstack plugin for octavia
bloganjohnsom: you mean rename class?21:09
*** kobis has quit IRC21:10
*** blogan_ has joined #openstack-lbaas21:18
blogan_one day ill switch off time warner21:18
xgermanthey are giving me pan, too21:33
blogan_pretty sure their slogan is "we give you pain so we don't have it"21:35
*** crc32 has quit IRC21:39
xgermanwell, my alternative is AT&T...21:39
xgermanso the lesser evil21:39
*** nmagnezi has joined #openstack-lbaas21:40
blogan_mine too :(21:53
blogan_though they can't be any worse honestly with stability21:53
*** blogan_ has quit IRC22:00
*** barclaac has quit IRC22:03
johnsomblogan Sorry, had to be away for a bit talking to a potential intern.  You can "name" tasks when building the flow22:05
*** barclaac has joined #openstack-lbaas22:06
*** nmagnezi has quit IRC22:06
*** barclaac has quit IRC22:16
*** barclaac has joined #openstack-lbaas22:17
*** TrevorV|home has quit IRC22:17
*** Miouge has quit IRC22:21
*** blogan_ has joined #openstack-lbaas22:27
*** mlavalle has quit IRC22:28
*** mlavalle has joined #openstack-lbaas22:33
*** barclaac has quit IRC22:42
*** barclaac has joined #openstack-lbaas22:47
*** barclaac has quit IRC23:04
*** mlavalle has quit IRC23:05
*** mlavalle has joined #openstack-lbaas23:10
*** barclaac has joined #openstack-lbaas23:14
*** mlavalle has quit IRC23:20
madhu_akblogan_, blogan: could you have a look into this patch and offer your suggestions, if any:
openstackgerritMerged openstack/neutron-lbaas: Updated from global requirements
*** bharath_ has quit IRC23:27
*** Aish has quit IRC23:29

Generated by 2.14.0 by Marius Gedminas - find it at!