Thursday, 2015-04-02

*** madhu_ak has quit IRC00:05
*** Aish has quit IRC00:11
*** barclaac has joined #openstack-lbaas00:27
*** jamiem has quit IRC00:27
*** SumitNaiksatam has joined #openstack-lbaas00:31
*** barclaac has quit IRC00:33
*** barclaac has joined #openstack-lbaas00:47
*** Aish has joined #openstack-lbaas00:52
*** Aish has quit IRC00:52
*** barclaac has quit IRC00:52
openstackgerritMichael Johnson proposed stackforge/octavia: Implements Octavia Controller Worker  https://review.openstack.org/15149600:57
*** haigang has joined #openstack-lbaas00:59
*** logan2 has quit IRC01:19
*** logan2 has joined #openstack-lbaas01:21
*** crc32 has quit IRC01:25
*** sballe has quit IRC01:28
*** xgerman has quit IRC01:29
*** xgerman has joined #openstack-lbaas01:29
*** xgerman has quit IRC01:50
*** crc32 has joined #openstack-lbaas02:56
*** sbalukoff has quit IRC03:02
crc32so I have to import octavia.common.config before I import octavia.api.v1.handlers.queue.producer but this breaks the pep8 alphabetical ordering non sense.  How am I supposed to resolve this?03:25
*** rm_work|away is now known as rm_work03:31
*** SumitNaiksatam has quit IRC03:38
*** SumitNaiksatam has joined #openstack-lbaas03:38
*** SumitNaiksatam has quit IRC03:40
*** SumitNaiksatam has joined #openstack-lbaas03:42
*** woodster_ has quit IRC04:00
*** SumitNaiksatam has quit IRC04:07
*** SumitNaiksatam has joined #openstack-lbaas04:07
*** sbalukoff has joined #openstack-lbaas04:11
*** xgerman has joined #openstack-lbaas04:11
*** fnaval has quit IRC04:21
*** SumitNaiksatam has quit IRC04:30
*** rm_work is now known as rm_work|away04:32
*** SumitNaiksatam has joined #openstack-lbaas04:34
*** woodster_ has joined #openstack-lbaas04:35
openstackgerritCarlos Garza proposed stackforge/octavia: Implementation for bp/api-to-oslo-messing-handler  https://review.openstack.org/14907904:39
*** fnaval has joined #openstack-lbaas04:48
*** kiran-r has joined #openstack-lbaas05:00
*** xgerman has quit IRC05:02
*** mugu has quit IRC05:05
*** mugu has joined #openstack-lbaas05:05
dougwigcrc32: why the ordering issue?05:50
crc32octavia.common.config must be imported after octavia.api.blah.blah.blah.producer05:50
crc32with out import octavia.common.config the oslo_messaging group is not added to the config05:51
*** haigang has quit IRC06:02
openstackgerritCarlos Garza proposed stackforge/octavia: Implementation for bp/api-to-oslo-messing-handler  https://review.openstack.org/14907906:06
*** bharath has joined #openstack-lbaas06:08
*** haigang has joined #openstack-lbaas06:12
*** bharath has quit IRC06:21
*** crc32 has quit IRC06:28
*** bharath has joined #openstack-lbaas06:30
openstackgerritShane McGough proposed openstack/neutron-lbaas: Implement KEMPtechnologies lbaas driver v2 shim  https://review.openstack.org/16794506:31
*** kobis has joined #openstack-lbaas06:34
*** woodster_ has quit IRC06:40
*** apuimedo has quit IRC06:41
*** rm_work|away is now known as rm_work06:51
rm_workblogan / ptoohill: you there?06:51
*** chlong has quit IRC06:55
rm_workdougwig: ah, i see what he did07:08
rm_workdougwig: i will explain tomorrow how to use import_group07:08
rm_work(to him)07:09
*** bharath has quit IRC07:13
*** sbalukoff has quit IRC07:20
ptoohillrm_work: hey07:21
rm_workah07:21
*** sbalukoff has joined #openstack-lbaas07:25
*** haigang has quit IRC07:30
*** haigang has joined #openstack-lbaas07:35
openstackgerritPhillip Toohill proposed openstack/neutron-lbaas: WIP: Neutron LBaaS: TLS Barbican Scenario Test  https://review.openstack.org/16482807:40
*** jschwarz has joined #openstack-lbaas07:56
*** haigang has quit IRC07:56
*** haigang has joined #openstack-lbaas08:02
*** BrianShang has quit IRC08:20
*** BrianShang has joined #openstack-lbaas08:22
*** haigang has quit IRC08:24
*** kiran-r has quit IRC08:27
*** haigang has joined #openstack-lbaas08:28
*** _kiran_ has joined #openstack-lbaas08:52
*** amotoki has joined #openstack-lbaas09:05
*** amotoki has quit IRC09:05
*** amotoki has joined #openstack-lbaas09:06
openstackgerritAdam Harwell proposed stackforge/octavia: Implementation for bp/api-to-oslo-messing-handler  https://review.openstack.org/14907909:11
rm_workerr09:16
rm_workgah09:17
rm_workdidn't mean to do that09:17
rm_work<_<09:17
rm_worki should go to sleep before I accidentally push any more reviews that aren't mine09:18
*** haigang has quit IRC09:24
*** amotoki has quit IRC09:31
*** haigang has joined #openstack-lbaas09:40
*** amotoki has joined #openstack-lbaas09:52
*** ptoohill has quit IRC10:03
*** ptoohill has joined #openstack-lbaas10:04
*** amotoki has quit IRC10:12
*** bharath has joined #openstack-lbaas10:14
*** bharath has quit IRC10:18
*** haigang has quit IRC10:55
*** _kiran_ has quit IRC11:01
*** _kiran_ has joined #openstack-lbaas11:09
*** haigang has joined #openstack-lbaas11:23
*** _kiran_ has quit IRC11:29
*** woodster_ has joined #openstack-lbaas12:27
*** BrianShang has quit IRC13:00
*** BrianShang has joined #openstack-lbaas13:01
*** chlong has joined #openstack-lbaas13:01
*** chlong has quit IRC13:47
*** jamiem has joined #openstack-lbaas14:03
*** haigang has quit IRC14:08
*** mlavalle has joined #openstack-lbaas14:09
*** xgerman has joined #openstack-lbaas14:53
*** sbalukoff has quit IRC15:03
*** jamiem has quit IRC15:07
*** ajmiller has joined #openstack-lbaas15:24
*** SumitNaiksatam has quit IRC15:37
*** kbyrne has quit IRC15:48
*** kobis has quit IRC15:50
*** jamiem has joined #openstack-lbaas15:50
*** madhu_ak has joined #openstack-lbaas15:56
*** enikanorov__ has joined #openstack-lbaas15:58
*** madhu_ak has quit IRC16:01
*** barclaac has joined #openstack-lbaas16:01
*** enikanorov has quit IRC16:01
*** SumitNaiksatam has joined #openstack-lbaas16:04
*** madhu_ak has joined #openstack-lbaas16:12
*** kbyrne has joined #openstack-lbaas16:15
*** jschwarz_ has joined #openstack-lbaas16:15
*** mwang2 has joined #openstack-lbaas16:17
*** jschwarz has quit IRC16:17
*** bharath has joined #openstack-lbaas16:24
*** localloop127 has joined #openstack-lbaas16:25
johnsomWe need to get rm_work a sticker for his laptop that says "don't git sleepy"16:27
*** SumitNaiksatam has left #openstack-lbaas16:28
*** SumitNaiksatam has joined #openstack-lbaas16:30
openstackgerritPhillip Toohill proposed openstack/neutron-lbaas: Fixes cert_parser intermediate validation  https://review.openstack.org/17020316:38
ptoohillhttps://review.openstack.org/#/c/169206/16:38
ptoohillhttps://review.openstack.org/#/c/169633/16:39
ptoohillThats greate! johnsom16:39
ptoohillgreat*16:39
*** Miouge has quit IRC16:53
*** fnaval has quit IRC16:59
*** jschwarz_ has quit IRC17:01
*** barclaac|2 has joined #openstack-lbaas17:02
*** barclaac has quit IRC17:05
*** SumitNaiksatam has quit IRC17:12
*** sbfox has joined #openstack-lbaas17:16
*** SumitNaiksatam has joined #openstack-lbaas17:17
*** fnaval has joined #openstack-lbaas17:18
*** sbfox has quit IRC17:29
*** sbfox has joined #openstack-lbaas17:33
*** SumitNaiksatam_ has joined #openstack-lbaas17:45
*** SumitNaiksatam has quit IRC17:47
*** SumitNaiksatam_ is now known as SumitNaiksatam17:47
bloganxgerman, johnsom: ping18:04
xgermanhi18:05
bloganyall wanna talk about that network driver?18:05
xgermancan we move that after lunch (+2h)18:05
bloganxgerman: sure18:06
bloganxgerman: just ping me18:06
xgermanwill do18:06
madhu_akdougwig: https://review.openstack.org/#/c/165627/ could you please have a look? Thanks18:33
*** barclaac|2 has quit IRC18:43
*** kobis has joined #openstack-lbaas19:01
*** kobis has quit IRC19:06
*** barclaac|2 has joined #openstack-lbaas19:06
*** sbfox has quit IRC19:24
*** enikanorov__ has quit IRC19:27
*** crc32 has joined #openstack-lbaas19:43
*** mwang2 has quit IRC19:51
*** madhu_ak has quit IRC19:56
TrevorVxgerman you around my man?20:00
johnsomNot yet, just grabbed lunch so will be a few minutes20:02
xgermanTrevorV hi?20:03
TrevorVI talked with phil about the log location for HAProxy20:03
xgermanok, cool20:03
TrevorVHe said its configured already in the jinja config.  I think that's sufficient for a log directory specification20:03
TrevorVThoughts?20:03
TrevorV(was about to make that comment in the CR, but figured we should talk it out here and I'll put the answer in the comments for tracking)20:04
xgermanwell, I like it to be configurable in the config file since some operators have specific log locations on their vms (I need to put things into /mnt)20:04
johnsom+1 for configurable via config file and passed in the config down to the amp20:06
TrevorVRight, which is why you'd have a different template file.  Those template files, as I understand, are another for of configuration files.20:06
TrevorVform of config****20:06
xgermannot sure if it makes sense to have the end user replace template files20:07
johnsomSorry, was thinking about the agent, should be in octavia config and put in the haproxy config that goes down20:07
xgermanthen when we need to update the template for haproxy 1.6 the operator can't reuse their config files?20:08
TrevorVJust to make things clear then, should there be a section in the actual octavia.conf that directly relates to haproxy templating, and have us just read that and generate the full template off of config values?  That sounds like work duplication.20:08
xgermanno, the template adapts whatever we do to the right haproxy version20:09
xgermanthe config entries allow aus to adapt octavia to an operator's environment20:09
xgermanso ideally if haproxy 1.6 comes around we chaneg the template but I can still use my config values'20:10
johnsomI'm thinking of all of the people that will show up here with questions if we have them muck with the templates to set log directories and such20:11
TrevorVI can understand that, I just don't think its that much of a problem.20:13
TrevorVProbably because we're talking about Operators doing said configuration, where they may have to make more specific changes anyway, not just a log directory20:14
*** madhu_ak has joined #openstack-lbaas20:15
xgermanI can't think of a ton of changes anyway -- but the log directory is one I need ;-)20:15
TrevorVRight, which is easily done via the template update.  That's kinda the point of a template right?20:16
*** mwang2 has joined #openstack-lbaas20:16
TrevorVI guess what I'm seing is you guys are thinking of templates a little more concrete where the changes to it are made via config files, whereas I'm thinking of a simple template update to suit your environment, similar to a config in the first place.20:17
TrevorVseeing***20:17
johnsomI think of the template as abstracting the HaProxy bit as opposed to a commonly updated config.20:17
xgermanI understand but configs live in like /etc and survive updates whereas templates are not  -- it's just an odd choice for configuration20:18
TrevorVI can't say its "our" choice, right?  HAProxy made that ha ha.  Idk, I'm kinda speaking out without a TON of knowledge on this, and I'd have ptoohill here, but he had to step away for a bit.20:19
TrevorVI don't necessarily mind the log directory being in the config file and loaded into the template, I just want a clear line drawn: is this an exception, or is everything supposed to follow suit.  I feel like it shouldn't be both.20:20
xgermanagreed. For instance I am perfectly fine with the stats directory being only in the template20:20
xgermanbut I know log directory, etc. should be configurable since different operators have different layouts20:21
xgermanso the things we want to configure should be in the config files; things we think the values we put in are ok for everybody can just be in the template20:22
TrevorVOnce again, different operators should just have different templates... just like they have different configs.20:22
xgermanthen the template needs to be loaded from /etc/octavia20:22
xgermanand you need to have commnadline options to specify where to load from20:22
TrevorVI'm not sure if that's a problem.20:22
TrevorVI'd have to talk to ptoohill20:22
xgermanwell, in my opinion it is since then you can't upgrade templates automatically'20:23
TrevorVActually, I'm going to push back on that20:23
TrevorVDifferent operators have different drivers, isn't that correct?20:23
TrevorVIf an operator can change their driver, then they should be able to update their templates on their own as well.20:24
johnsomThat is correct, the template is used by the driver20:24
xgermanwell, I don't think that's the premise -- we shouldn't force Redhat to write a third haproxy driver20:24
xgermanso they can set a log directory20:24
TrevorVNo, they wouldn't, they'd make a different template if they needed.20:25
TrevorVEven then you're talking about the field labeled "log" having a value of "/var/log/redhat_said_here" in the template that came with the code base... Its not a difficult change to make there.20:26
xgermanyeah, but  a) should the template be a public interface where people make changes?  and b) are people expecting to put things there which normally go into a config file? c) does everybody need to learn haproxy.config and jinja2 to change their log file location?20:27
xgermanblogan - ping...20:28
bloganxgerman: hi20:28
johnsomI like to keep common config settings in the config file to reduce questions and mistakes.  I.e. keep them as simple as possible20:28
johnsomHi blogan20:29
xgermanI think we exchanged arguments and we have a packed day so let's talk network driver20:29
TrevorVThat's fair johnsom, but once again, where do we draw the line?  Do we generate it fully via config or not?  I disagree with a half and half.20:29
xgermanI like to not have to touch the template20:30
TrevorVThen we should make it ALL configurable, not just the one field you want to have configurable.  Does that make sense?20:30
xgermanyep, makes sense...20:30
johnsomI hear you on where the line is.  I think we, the folks contributing to octavia, should discuss what we all need to change to deploy and try to make that easy.20:31
xgerman+120:31
TrevorVI can understand that.  If everyone else is okay with it being partially configurable then I guess I'll deal, but I really dislike it.20:32
blogani think the log being configurable in the octavia.conf makes sense20:32
johnsom+1 blogan20:33
bloganand we don't need to have everything in the template in octavia.conf, evolve over time for when operators want to easily change somethign in the template, we can just add a config option20:34
blogandoesn't have to be all or nothing20:34
bloganxgerman, johnsom: networkdriver20:36
bloganso my intention was to have a network driver that works with a normal devstack build (normal to us at least).  Rackspace does not have the allowed-address-pairs extension enabled so it wouldn't work for us, if HP's cloud does great, but the intention is to just have a workable active/passive network driver20:37
*** mwang2 has quit IRC20:37
xgermansounds good20:38
johnsomWe do, but it doesn't work20:38
bloganlol yeah bc i messed it up a bit, pushing a new ps up soon20:38
blogani got reversed on which ips should go into the allowed_address_pairs20:38
xgermanok, johnsom meant to say address pair doesn't work on our cloud20:38
bloganohhh20:38
bloganso you have it enabled but it doesn't work?20:39
xgermanexactly20:39
johnsomYes, saddly20:39
blogansecurity groups issue?20:39
johnsomNo, some genius layer two optimization code that doesn't learn the arps right20:39
johnsomBut, I think we are going to ignore that and go forward as if that is fixed as it is in our Helion code20:40
bloganah okay, so it woudl be safe to say that HP's cloud isn't pure openstack? it does have some HP special sauce in certain places?20:40
xgermanof course...20:40
johnsom"special" is the right word20:40
bloganokay, then thats why we have the network drivers20:41
xgermanyep20:41
xgermanbut having one which works in devstack is rande20:41
johnsomSo, if you don't have address pairs either, what is the plan?20:41
xgermangrande20:41
bloganbut the ponit of this one was a quick way to have it work in a devstack deploy20:41
xgerman+100020:41
johnsom+120:41
bloganwell we have a plan for our networking infrastructure, but it definitely won't work for yall20:41
bloganso we'll be writing our own special driver20:41
johnsomI was also confused on the amp ips all on one port, but your clarification made it click for me20:42
bloganyeah the vip ip should be on the amp's vip ports20:42
bloganif that makes sense20:42
bloganwhich id ont think it did20:42
johnsomSo the new patchset you are pushing for network driver.  It will work with stock devstack kilo?20:43
xgerman(ajmiller happens to have an octavia devstack laying around BTW)20:43
ajmillerI'm working on getting that wrapped up.  Right now having permission/sudo problems, but will get that worked out soon../20:44
xgermanand then we have the network driver to go with it ;-)20:44
TrevorVWell xgerman, it looks as though my distaste for the "all or nothing" approach is vetoed.20:45
bloganjohnsom: theoretically, i ran through it manually a month or two ago and it worked, i was just now testing it out20:45
openstackgerritCarlos Garza proposed stackforge/octavia: Implementation for bp/api-to-oslo-messing-handler  https://review.openstack.org/14907920:46
johnsomOk, cool.  So I think we are good.  I don't have any more questions and will review when you push your new version up.20:46
bloganjohnsom: okay, also try it out too, i do think i might be missing some security-group commands as well, in fact im almost certain i am now20:47
johnsomYeah, will do20:47
bloganoh and it won't work until amphora have some kind of vrrp software running on it and a driver/agent that configures it20:48
bloganjohnsom, xgerman: yall are planning on doing an active/passive setup in production correct?20:48
xgermanwell, right now we are thinking "preview" without anyhting ;-)20:49
johnsomYes, Helion.  Though I have not been putting a lot of thought into it as that was a v1 goal.  No that we wouldn't want it sooner.20:49
johnsomAre you thinking VRRP?20:49
bloganby "without anything" you mean no HA basically?20:49
johnsomWe are really pushing for an end-to-end demo for the summit.  HA isn't necessary for that.20:50
xgermanyeah, our mind is really on that20:50
*** sbfox has joined #openstack-lbaas20:50
bloganoh of course, just wondering what the end product for HP would look like20:51
johnsomMy thought was we were going to "release" v0.5 without HA, then in v1 workout the HA (including controllers/mysql/etc)20:51
johnsomIf you have syncs to accelerate HA I support it20:52
xgerman+!20:52
johnsomMan, lunch coma, cycles20:53
bloganwell this network driver should support HA and no HA setups in devstack, so if you only have one VM it should still work20:53
xgermanwe like HA but I think having something to show customers and get their feedback is our top priority ;-)20:53
blogani read the syncs and thought you had some new lingo20:53
johnsomThanks for the extra credit20:53
bloganwell hell no HA woudl be quite a simple network driver20:54
xgermanyep20:54
johnsomThat said, if you see things that aren't HA ready, call them out and/or fix them.  We are really focused on getting an end-to-end working20:55
xgermanyep, we need to get that show on the road...20:55
bloganoh of course, i made some changes to the network interface writing this bc it was with HA in mind20:56
johnsomExactly, the road to Vancouver20:56
johnsomThat is cool20:56
bloganand should also support Rackspace's active/active topology as well20:56
johnsomI still have the mental model, v0.5 is working, v1 is HA, v2 is horiz scale20:56
bloganand i think its just generic enough to support other implementations20:57
johnsomExcellent20:57
bloganfamous last words :(20:57
xgermanwwe gotta run to our next meeting... dense week :-(20:57
blogangood luck!20:58
*** mwang2 has joined #openstack-lbaas20:58
*** bharath has quit IRC20:59
*** bharath_ has joined #openstack-lbaas20:59
*** bharath_ has quit IRC21:00
*** bharath has joined #openstack-lbaas21:00
*** ajmiller_ has joined #openstack-lbaas21:10
*** fnaval has quit IRC21:13
*** ajmiller has quit IRC21:14
openstackgerritTrevor Vardeman proposed stackforge/octavia: Amphora SSH Driver  https://review.openstack.org/16096421:21
*** sbalukoff has joined #openstack-lbaas21:22
*** dougwig has quit IRC21:35
*** rm_work has quit IRC21:37
*** rm_work has joined #openstack-lbaas21:41
*** rm_work has quit IRC21:41
*** rm_work has joined #openstack-lbaas21:41
*** dougwig has joined #openstack-lbaas21:45
*** fnaval has joined #openstack-lbaas21:46
openstackgerritAl Miller proposed stackforge/octavia: Add devstack plugin for octavia  https://review.openstack.org/16779621:51
*** localloop127 has quit IRC21:56
openstackgerritAishwarya Thangappa proposed stackforge/octavia: Implements Octavia Controller Worker  https://review.openstack.org/15149622:02
*** madhu_ak has quit IRC22:06
*** madhu_ak has joined #openstack-lbaas22:06
*** SumitNaiksatam has quit IRC22:12
*** crc32 has quit IRC22:20
*** barclaac has joined #openstack-lbaas22:49
*** barclaac|2 has quit IRC22:53
*** barclaac|2 has joined #openstack-lbaas23:01
*** barclaac has quit IRC23:03
*** sbfox has quit IRC23:09
*** jamiem has quit IRC23:26
*** jamiem has joined #openstack-lbaas23:26
*** barclaac has joined #openstack-lbaas23:27
*** barclaac|2 has quit IRC23:30
johnsomdougwig Can we get this in today so our coverage run tomorrow has the updated numbers?  https://review.openstack.org/#/c/169050/23:31
johnsomIt has a couple of +1s and a +2 already23:32
dougwigI'm at the furious 7 premiere. I can after.  Unless mestery is online and can do it now for us. (Ping mestery)23:32
johnsomAwesome, enjoy.23:33
*** barclaac|2 has joined #openstack-lbaas23:34
*** barclaac has quit IRC23:34
rm_youoh nice23:41
rm_youI forgot that was tonight or i would be there too dougwig T_T23:42

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