Friday, 2015-01-16

sbalukoffdougwig: Congrats on becoming a Neutron core reviewer, eh.00:04
dougwigty, but i have to get voted in first. :)00:05
sbalukoffEh, it'll happen.00:05
sbalukoffAnd you totally deserve it, too. ;)00:05
sbalukoff(In every sense of that word.)00:05
*** muhalam has joined #openstack-lbaas00:07
*** muhalam has quit IRC00:08
*** dougw2 has quit IRC00:10
*** ajmiller has joined #openstack-lbaas00:28
*** sbfox has quit IRC00:28
*** markmcclain has quit IRC00:28
*** sbfox has joined #openstack-lbaas00:38
*** ajmiller has quit IRC00:42
*** sbfox has quit IRC00:44
*** omnibus7 is now known as davidlenwell00:49
*** davidlenwell is now known as davidlenwell__00:50
*** davidlenwell__ is now known as davidlenwell00:50
*** barclaac|2 has quit IRC00:55
*** markmcclain has joined #openstack-lbaas00:55
*** xgerman has quit IRC01:01
*** barclaac has joined #openstack-lbaas01:06
*** barclaac|2 has joined #openstack-lbaas01:07
*** barclaac has quit IRC01:11
*** markmcclain has quit IRC01:23
openstackgerritPhillip Toohill proposed stackforge/octavia: taskflows to create, delete amphora
*** evgenyf has quit IRC01:31
*** openstack has joined #openstack-lbaas01:34
openstackgerritPhillip Toohill proposed stackforge/octavia: taskflows to create, delete amphora
openstackgerritPhillip Toohill proposed stackforge/octavia: taskflows to create, delete amphora
*** SumitNaiksatam has quit IRC04:09
*** mlavalle_ has quit IRC04:20
*** amotoki has quit IRC04:23
*** amotoki has joined #openstack-lbaas04:42
*** sbfox has joined #openstack-lbaas04:57
*** kobis has joined #openstack-lbaas05:00
*** kobis has quit IRC05:07
*** kobis has joined #openstack-lbaas05:17
*** kobis has quit IRC05:22
*** kobis has joined #openstack-lbaas05:23
*** kobis has quit IRC05:42
*** woodster_ has quit IRC06:10
*** sbalukoff has quit IRC06:39
*** sbalukoff has joined #openstack-lbaas06:52
*** sbfox has quit IRC06:56
*** sbfox has joined #openstack-lbaas07:05
*** kobis has joined #openstack-lbaas07:11
*** kobis has quit IRC07:19
*** chlong has quit IRC07:20
openstackgerritPhillip Toohill proposed stackforge/octavia: taskflows to create, delete amphora
*** Miouge has joined #openstack-lbaas07:52
*** apuimedo has joined #openstack-lbaas07:53
*** kobis has joined #openstack-lbaas09:39
*** kobis has quit IRC09:46
*** sbfox has quit IRC09:51
*** apuimedo has quit IRC11:22
*** kobis has joined #openstack-lbaas11:47
*** kbyrne has quit IRC12:06
*** kbyrne has joined #openstack-lbaas12:08
*** kobis has quit IRC12:14
*** apuimedo has joined #openstack-lbaas12:59
*** woodster_ has joined #openstack-lbaas13:18
*** Miouge_ has joined #openstack-lbaas13:56
*** Miouge has quit IRC13:59
*** Miouge_ is now known as Miouge13:59
*** ajmiller has joined #openstack-lbaas14:39
*** kobis has joined #openstack-lbaas14:45
*** kobis has quit IRC14:55
*** kobis has joined #openstack-lbaas14:55
*** amotoki has quit IRC14:57
*** fnaval has quit IRC15:30
*** fnaval has joined #openstack-lbaas15:31
*** fnaval has quit IRC15:35
*** fnaval has joined #openstack-lbaas15:47
*** kobis has quit IRC16:01
*** xgerman has joined #openstack-lbaas16:06
*** kobis has joined #openstack-lbaas16:07
xgermanblogan, dougiwg (sbalukoff) can we get some eyes on ? Otherwise I need to learn how to chain reviews... or you can just explain how that works...16:11
*** Miouge has quit IRC16:19
*** Miouge has joined #openstack-lbaas16:21
*** kobis has quit IRC16:34
*** jorgem has joined #openstack-lbaas16:40
*** Guest29310 is now known as redrobot16:41
ptoohillxgerman, to chain them, checkout whichever is the 'first' or the dependency, then cherry pick the other commit on top. submit that and you have a chained review16:57
*** amotoki has joined #openstack-lbaas16:58
sbalukoffptoohill: Did you see the comments I made last night on the amphora api spec?17:00
xgermanptoohill thanks -- hopefully I will play around with that ;-)17:00
ptoohillindeed i did sbalukoff, thank you :)17:00
sbalukoffIs that enough to unblock you?17:00
sbalukoffOr did you have more questions there?17:00
*** apuimedo has quit IRC17:01
ptoohillI believe so, i need to go back and reread the cert answer(read on my phone this am) and i may have a question about that one17:02
sbalukoffptoohill: The more I think about that one, the more I think relative paths probably won't work in the config file. So we should pick a path template and stick with that.17:03
sbalukoffAgain, probably under something like:  /var/lib/octavia/<listener_uuid>/<certificate_name.pem>17:03
ptoohillin that case, that resolves the question i would have had17:03
ptoohillIn the driver im only concerned with this because thats where i build the config. otherwise i wouldnt care where they were put17:04
sbalukoffYep, I understand!17:04
xgermansbalukoff, ptoohill we put all certs in /etc/ssl17:04
sbalukoffxgerman: Certs need to go on a memory filesystem.17:05
ptoohillthat could work too, and what i was testing with17:05
xgermannot sure if there is a standard where to put them or not17:05
xgermanyou can mount that fs there ;-)17:05
sbalukoffxgerman: listener_uuid needs to be part of the path.17:05
ptoohillas long as i know the base path i can generate the path without any problems17:05
sbalukoffor filename.17:05
xgermanfile location + if not memory fs are two different things17:05
sbalukoffxgerman: I agree.17:06
sbalukoffHowever, I also tend to try to avoid having automated tools put stuff under /etc17:06
sbalukoffThere's really no reason not to put this under /var/lib/octavia or something.17:06
ptoohillYea, the files can be stored anywhere, we just need to know where that is17:07
sbalukoffAnd in any case the exact root under which we put something is supremely unimportant.17:07
ptoohillthough, i dont have much experience with memoryFS so i am unsure of specific restrictions17:08
xgermanyes, i was just saying where we put things and the system tools seems to find certs there (there is always the issue with all the ca certs, etc.)17:08
*** mlavalle has joined #openstack-lbaas17:08
sbalukoffSystem tools aren't going to be useful for us in this case.17:08
dougwigmorning all17:08
xgermansbalukoff we usually use them to debug but I get your point17:09
dougwig'git review -d' makes it simple. you just have to be careful of rebases.17:09
xgermannow i actually need to write the code I intended to write ;-)17:10
*** amotoki has quit IRC17:11
*** amotoki has joined #openstack-lbaas17:15
*** sbfox has joined #openstack-lbaas17:18
*** kobis has joined #openstack-lbaas17:20
TrevorVgoooooooooood merning!17:22
*** amotoki has quit IRC17:27
*** apuimedo has joined #openstack-lbaas17:28
*** SantoshNetScaler has joined #openstack-lbaas17:29
bloganfinally back up and running on my real machine!17:34
blogani was forced into os x land for a week and i did not like it17:35
*** amotoki has joined #openstack-lbaas17:36
ptoohillLike old man stuck in his ways17:36
ptoohillTOUCHTONE PHONES, we dont need nun of them smart dillywhacker phones17:37
*** kobis has quit IRC17:45
sbalukoffblogan: I hear you. I gave os X land a try for a couple years and am glad to be back on linux.17:57
*** kobis has joined #openstack-lbaas17:58
*** Miouge has quit IRC18:00
dougwiggive me a professionally maintained window system with a unix kernel any day.18:09
*** sbfox has quit IRC18:10
sbalukoffdougwig: Haha! Just make sure you like doing things the Steve Jobs way.18:11
sbalukoffBecause if you don't, then too bad.18:12
dougwigit's true. though increasingly it's the "not steve jobs" way. the amount of spurious popup crap is approaching microsoft proportions these days.18:12
*** amotoki has quit IRC18:15
*** Miouge has joined #openstack-lbaas18:17
*** sbfox has joined #openstack-lbaas18:20
*** amotoki has joined #openstack-lbaas18:22
*** sbfox1 has joined #openstack-lbaas18:29
*** sbfox1 has quit IRC18:29
*** sbfox has quit IRC18:29
*** sbfox1 has joined #openstack-lbaas18:29
*** amotoki has quit IRC18:58
openstackgerritGerman Eichberger proposed stackforge/octavia: taskflows to create, delete amphora
xgermanI think what dougwig recommended went terribly wrong19:00
xgermanit bundled up the parent patch, too19:00
xgermangoing to a metting19:01
*** amotoki has joined #openstack-lbaas19:01
TrevorVsbalukoff you alive my man?19:23
ptoohillYea, seen that xgerman, the way i do it is to checkout the dep then cherry pick the needed-by on top(in this case patchset 14, so you can undo the merge)19:23
dougwigxgerman: can you post for command history, so we can see where it went wrong?19:26
bloganxgerman: all you need to do is pull that the review you depend on, and then add another commit on top of it19:27
*** amotoki has quit IRC19:41
*** sbfox1 has quit IRC19:58
*** sbfox has joined #openstack-lbaas19:59
*** sbfox has quit IRC19:59
*** sbfox has joined #openstack-lbaas20:00
*** apuimedo has quit IRC20:03
sbalukoffTrevorV: Partially. What's up?20:07
rm_workI think they MAY have just gone to do foosball :P20:11
rm_workxgerman: +1 for blogan and ptoohill's preferred method, 100% what I always recommend.20:12
rm_workcheckout latest master (or whatever dependency) and use the cherry-pick command to grab the change you want :)20:13
rm_workso what reviews do people need me to look at?! I'm in full-on review mode today20:17
rm_workI feel like we used to have a page somewhere with "latest reviews" but i don't know if it's up to date or if it included both neutron-lbaas and octavia or just one of the two20:18
*** SantoshNetScaler has quit IRC20:36
xgermanI will give that a spin20:38
bloganjust look at non WIP octavia reviews in gerrit20:38
rm_workblogan: yeah looking at a review now20:39
bloganxgerman: is this a WIP?
rm_workactually I am looking at a WIP right now, but by request <_<20:41
xgermanI wanted to base it on the in review  nova driver20:41
bloganoh okay20:41
bloganso pull the nova driver review down, with git review -d20:41
bloganthen do this20:42
blogangit fetch refs/changes/69/142569/15 && git cherry-pick FETCH_HEAD20:42
bloganthen git review and that'll be a dependency20:42
rm_work(assuming the cherry-pick finishes and doesn't have conflicts)20:43
xgermanok, let me try that20:44
xgermangit review20:47
xgermanYou are about to submit multiple commits. This is expected if you are20:47
xgermansubmitting a commit that is dependent on one or more in-review20:47
xgermancommits. Otherwise you should consider squashing your changes into one20:47
xgermancommit before submitting.20:47
xgermanThe outstanding commits are:20:47
xgerman7e964e0 (HEAD, review/german_eichberger/nova-driver-fix) taskflows to create, delete amphora20:47
xgermanb8089b9 (nova-driver-fix) Patches nova client to work with HP Cloud20:47
xgermanDo you really want to submit the above commits?20:47
xgermanType 'yes' to confirm, other to cancel:20:48
openstackgerritGerman Eichberger proposed stackforge/octavia: taskflows to create, delete amphora
blogantype yes20:48
rm_workusually good to check the commit hashes of the ones that aren't yours tho20:48
rm_workand verify they match the latest for the deps20:48
xgermanI did no before and squashed20:48
xgermanThanks for the help!!20:48
xgermanrm_work one more question -- It seems I need to wrap the nova driver in a singleton to avoid the keystone auth calls20:56
xgermanyeah, it feels hmm as well for me20:56
rm_workthe nova/keystone clients should already be a singleton20:56
bloganshould have done module methods20:56
rm_workso it shouldn't matter how many times the driver asks for it20:56
xgermanwell, when I was running my flow
xgermanit looked like I did a keystone auth for both tasks I was using the nova_client20:57
bloganif those keystone methods were module methods, you could do a singleton fairly easily20:57
sballe__@blogan Ho20:57
rm_workthe keystone methods are already set up to work that way tho20:57
sballe__blogan: What version of HA proxy is being pulled in with Neutron LBaaS v2 ?20:58
dougwigif it prompts you with multiple commits ,you *should* go compare the hashes in gerrit to what's in your terminal, and make sure they match.20:58
xgermanlet me double check the log20:58
blogansballe__: neutron lbaas v2 does not install a version, but for tls it will need 1.520:59
ptoohilldoesnt it pull 1.4? blogan20:59
xgermanrm_work -- double checked it does it once20:59
ptoohillwere still waiting for legit packages for 1.5 otherwise we need to set up ppa20:59
bloganptoohill: devstack would yes, but i neutron lbaas v2 code does not pull down haproxy21:00
rm_workxgerman: ok, so it is working as intended?21:00
xgermanyes, it is -- Thank you... soryy for gettign confused21:00
rm_workI was really hoping taskflow wasn't somehow subverting the static class stuff :P21:00
bloganif i felt like being overly nitpicky i would have said put those methods at the module scope21:00
rm_workblogan: uhh then how would they set class variables to store the objects?21:01
bloganmodule variables21:01
bloganNOVA_CLIENT = None21:01
rm_workhow is that not way uglier to you21:01
rm_workI would -1 that all day long, and say "go put that in a class"21:02
bloganbc its simpler and more pythonic in my opinion21:02
rm_workit is rife for abuse21:02
bloganyou mean putting code just in module scope is rife for abuse?21:02
rm_workfor instance when people realize "oh i just grab NOVA_CLIENT" right?21:02
bloganput _ infront of it whatever21:02
bloganwhats stopping someone from grabbving NovaKeystoneAuth._nova_client?21:03
rm_workit just says to me that you're avoiding a class on purpose21:03
rm_workthat's what I see21:03
rm_workbecause convention clearly dictates that's wrong21:03
bloganputting that in a class and just having nothing but class methods tells me you dont need a class21:03
rm_workand they can inspect the class to see the correct answer :P21:03
bloganand they can't inspect the module?21:04
*** kobis has quit IRC21:04
rm_workthey could, but it feels … icky21:05
rm_worki don't think either of us is going to move anywhere on this one21:05
rm_workfor me it's an instinctual code style issue21:05
bloganill agree its a matter of taste, but wrapping things in classes just for the sake of namespacing is overkill since the module is still a namespace21:06
bloganespecially since you can't instantiate a module, that class can be instantiated, but that would give you nothing21:07
bloganexcept confusion21:07
rm_workI guess so <_<21:15
rm_workbut it is very rare that I see any attributes/methods at module level that aren't static/config related21:16
bloganits acutally how a lot of code does singletons21:17
bloganhere's an example from barbican21:18
*** mlavalle has quit IRC21:19
rm_workugh that's weird21:23
bloganmaybe for some, but im not a big fan of putting code in __init__ files but I can see why people do it21:25
dougwigif it really has to happen at import, i'll do it, but as a one-liner to another module to keep it short.21:26
bloganyeah my thoughts on that too21:26
blogani also liked the oslo.namespace libraries over the new oslo_namespace, but i guess im just weird21:27
dougwigno, you're sane.21:28
bloganhmm a dougwig compliment21:30
mesteryblogan dougwig: You guys both got into trouble in school a lot didn't you?21:35
dougwigof course.  school was boring.21:36
blogani was a model student21:37
sballe__rm_work: I am assuming we'll need ha-proxy 1.5 for ipv6 too21:43
openstackgerritTrevor Vardeman proposed stackforge/octavia: haproxy reference amphora API client
*** barclaac has joined #openstack-lbaas22:25
rm_worklooks like that's possible22:27
rm_workwas ipv6 in our 0.5/1.0 plan?22:27
rm_workI hope so actually but I have often been accused of over-committing for our early target milestones22:27
*** barclaac|2 has quit IRC22:28
*** jorgem has quit IRC22:28
*** mlavalle has joined #openstack-lbaas22:41
sbalukoffrm_work: It's not, but I would like to see it soon, thereafter. :)23:01
sbalukoffIn any case, none of the code we're committing should be IPv6 unfriendly.23:02
*** mlavalle has quit IRC23:13
rm_workneed to really check xgerman's nova driver code23:18
rm_workto make sure it handles the return including ipv4 AND ipv6 addresses23:19
xgermanyeah, go for it23:19
*** openstackgerrit has quit IRC23:19
rm_workor maybe it's TrevorV's code, not sure which part of it is handling that23:19
rm_workxgerman: what does your nova return for that variable? is it a list like mine does?23:19
*** openstackgerrit has joined #openstack-lbaas23:19
xgermanbut if you have an ipv6 cloud handy just run it23:20
rm_workyeah, I don't actually know how I'd get any of the Octavia code "running" yet23:21
xgermanTrevor's code needed an extension we don't have in HP's cloud23:21
rm_workare you just importing pieces?23:21
rm_workat a python CLI?23:21
xgermanif you look at my flow commit the api manager in there will run23:21
rm_workhmm k23:21
rm_work ?23:21
xgermanit has a  main23:22
rm_workso i have pulled that down23:22
rm_workwhat do I "run"?23:22
xgermanthat;s the thing with the main and if your octavia configs are good you should be able to spin up vms23:22
rm_workI'll give that a shot23:23
rm_work… my configs are probably not good23:24
xgermanmy goal is to hammer the flows out + integrate/test any drivers as they become available23:24
rm_workis it expecting them in /etc/octavia/ ?23:24
xgermanyep, I linked there23:24
xgerman(based on your advice at the hackathon)23:24
*** mlavalle has joined #openstack-lbaas23:25
rm_workwhich parts of the config need to be right?23:25
rm_workjust keystone_authtoken and ...23:25
rm_workoh, networking probably23:25
xgermanand DB23:25
xgermanand networking23:26
rm_workDB has to be set up? ugh23:26
rm_workyeah I don't have this machine running mysql currently23:26
xgermanyep, it's my end-to-end test :-)23:26
rm_workalright I will probably try this out next week23:26
rm_workwhen I am not WFH23:26
rm_workI was planning on just doing reviews and such from home today which are easy without anything installed locally ;P23:26
xgermanyou can just comment out the peices which write to the DB BTW23:26
*** mlavalle has quit IRC23:27
xgermananyhow, that's how I test/develop -- and I have the big reckoning of writing days of unit tests ahead of me :-(23:27
*** mlavalle has joined #openstack-lbaas23:28
*** mlavalle has quit IRC23:28
*** mlavalle has joined #openstack-lbaas23:29
*** barclaac has quit IRC23:30
*** mlavalle has quit IRC23:30
*** mlavalle has joined #openstack-lbaas23:30
*** mlavalle has quit IRC23:58

Generated by 2.14.0 by Marius Gedminas - find it at!