Thursday, 2015-07-30

*** minwang2 has quit IRC00:15
*** abdelwas has quit IRC00:21
madhu_akfinally, jenkins is happy with the scenario tests for lbaasv2!
bana_kwhen is the next octavia hackthon?00:46
*** amotoki has joined #openstack-lbaas00:58
*** h00327910__ has joined #openstack-lbaas00:58
*** johnsom__ has quit IRC00:59
*** vivek-ebay has quit IRC01:12
*** vivek-ebay has joined #openstack-lbaas01:12
*** madhu_ak has quit IRC01:17
sbalukoffbana_k: There probably won't be one until the next mid-cycle. So probably December or January.01:27
sbalukoffAt least, not one where we meet in person and work on stuff / discuss stuff in the same room.01:27
sbalukoffIf we wanted to do a 'virtual hack-a-thon' where we try to do the same purely over tele-presence, this could probably happen (as it costs nothing, and a limiting factor here is travel budgets)...01:28
sbalukoffDid you have something in mind you wanted to work on, or was your question more about just meeting in person to have an easier time pairing up on work?01:29
*** chlong has joined #openstack-lbaas01:29
sbalukoffbana_k: As far as things you can help with right now:  I would love it if you could take a look at the sqlalchemy stuff in the Neutron LBaaS patch I have up right now and let me know if you think I'm doing it wrong. I suspect I am.01:33
sbalukoffThis is what I'm talking about, the all_pools methods / property I'm adding to the model:
sbalukoffThere are many ways I could be doing this wrong, and I suspect I'm violating at least a couple rules with this.01:34
sbalukoff(This all needs to be unit-tested out, as well.)01:35
bana_ksbalukoff: Oh I see. Nothing in particular. Was just thinking if its there, can attend and speedup the learning process :).01:40
bana_kI can definitely work on it01:41
*** Tiancheng has joined #openstack-lbaas01:52
*** mixos has joined #openstack-lbaas02:00
*** mixos has left #openstack-lbaas02:03
*** bana_k has quit IRC02:06
*** vivek-ebay has quit IRC02:26
*** mixos has joined #openstack-lbaas02:37
*** ajmiller has joined #openstack-lbaas02:41
*** KunalGandhi has quit IRC02:55
*** ajmiller has quit IRC02:58
*** ajmiller has joined #openstack-lbaas02:58
*** woodster_ has quit IRC03:14
*** fnaval has quit IRC03:34
*** fnaval has joined #openstack-lbaas03:43
*** fnaval has quit IRC03:47
*** fnaval has joined #openstack-lbaas03:47
*** vivek-ebay has joined #openstack-lbaas03:49
*** minwang2 has joined #openstack-lbaas04:05
*** bana_k has joined #openstack-lbaas04:31
*** Tiancheng has quit IRC04:34
*** madhu_ak has joined #openstack-lbaas04:43
*** vivek-eb_ has joined #openstack-lbaas05:06
*** vivek-ebay has quit IRC05:08
*** Tiancheng has joined #openstack-lbaas05:22
*** minwang2 has quit IRC05:26
*** minwang2 has joined #openstack-lbaas05:28
*** KunalGandhi has joined #openstack-lbaas05:31
*** vivek-eb_ has quit IRC05:37
*** KunalGandhi has quit IRC05:43
*** KunalGandhi has joined #openstack-lbaas05:43
*** jschwarz has joined #openstack-lbaas05:43
*** madhu_ak has quit IRC05:57
*** numan has joined #openstack-lbaas06:08
*** Tiancheng has quit IRC06:09
*** Tiancheng has joined #openstack-lbaas06:10
*** jschwarz has quit IRC06:39
*** minwang2 has quit IRC07:05
*** nmagnezi has joined #openstack-lbaas07:07
*** bana_k has quit IRC07:11
*** KunalGan_ has joined #openstack-lbaas07:16
*** KunalGandhi has quit IRC07:19
*** KunalGan_ has quit IRC07:21
*** chlong has quit IRC07:25
*** kiran-r has joined #openstack-lbaas08:25
*** enikanorov has joined #openstack-lbaas08:33
openstackgerritPhillip Toohill proposed openstack/octavia: Pluggable controller worker
openstackgerritPhillip Toohill proposed openstack/neutron-lbaas: Fixing private key loading
*** chlong has joined #openstack-lbaas09:07
openstackgerritPhillip Toohill proposed openstack/octavia: Adding new network driver
*** ig0r_ has joined #openstack-lbaas09:31
*** Tiancheng has quit IRC10:10
*** kbyrne has quit IRC10:21
*** kbyrne has joined #openstack-lbaas10:23
*** jschwarz has joined #openstack-lbaas10:52
*** Tiancheng has joined #openstack-lbaas12:05
*** _kiran_ has joined #openstack-lbaas12:06
*** kiran-r has quit IRC12:09
*** _kiran_ has quit IRC12:20
*** mixos has quit IRC12:40
*** nmagnezi has quit IRC12:41
*** nmagnezi has joined #openstack-lbaas12:56
*** ig0r_ has quit IRC13:10
*** ig0r_ has joined #openstack-lbaas13:16
*** rbrooker has joined #openstack-lbaas13:23
*** woodster_ has joined #openstack-lbaas13:24
*** kbyrne has quit IRC13:26
*** kbyrne has joined #openstack-lbaas13:27
*** ig0r_ has quit IRC13:31
*** Tiancheng has quit IRC14:08
*** mlavalle has joined #openstack-lbaas14:11
*** KunalGandhi has joined #openstack-lbaas14:20
*** ajmiller has quit IRC14:21
*** KunalGandhi has quit IRC14:24
*** ajmiller has joined #openstack-lbaas14:30
*** ig0r_ has joined #openstack-lbaas14:32
*** rbrooker has quit IRC14:38
*** madhu_ak has joined #openstack-lbaas14:47
*** fnaval has quit IRC14:57
*** nmagnezi has quit IRC15:01
*** jschwarz has quit IRC15:01
*** rbrooker has joined #openstack-lbaas15:02
*** ig0r_ has quit IRC15:03
*** ig0r__ has joined #openstack-lbaas15:03
*** fnaval has joined #openstack-lbaas15:07
*** logan2 has quit IRC15:08
*** ig0r_ has joined #openstack-lbaas15:09
*** ig0r__ has quit IRC15:09
*** ig0r__ has joined #openstack-lbaas15:13
*** ig0r__ has quit IRC15:15
*** ig0r_ has quit IRC15:15
*** vivek-ebay has joined #openstack-lbaas15:16
*** numan has quit IRC15:20
*** KunalGandhi has joined #openstack-lbaas15:21
*** madhu_ak has quit IRC15:26
*** KunalGandhi has quit IRC15:27
*** vivek-ebay has quit IRC15:27
*** xgerman has quit IRC15:29
*** minwang2 has joined #openstack-lbaas15:31
*** rbrooker has quit IRC15:45
*** ig0r_ has joined #openstack-lbaas15:52
*** woodster_ has quit IRC15:55
*** h00327910__ has quit IRC15:55
*** rbrooker has joined #openstack-lbaas15:59
*** rbrooker has quit IRC16:06
*** rbrooker has joined #openstack-lbaas16:08
johnsomblogan rm_work Anyone around to give this the +A?  Data model updates part 1 for VRRP16:14
*** rbrooker has quit IRC16:14
*** logan2 has joined #openstack-lbaas16:19
*** minwang2 has quit IRC16:28
*** woodster_ has joined #openstack-lbaas16:33
openstackgerritMerged openstack/octavia: Adding model changes to support active/standby
*** vivek-ebay has joined #openstack-lbaas16:36
*** madhu_ak has joined #openstack-lbaas16:37
*** xgerman has joined #openstack-lbaas16:37
*** KunalGandhi has joined #openstack-lbaas16:44
*** TrevorV has joined #openstack-lbaas16:57
*** bana_k has joined #openstack-lbaas17:01
*** h00327910__ has joined #openstack-lbaas17:08
*** minwang2 has joined #openstack-lbaas17:09
johnsommestery Any thoughts on the pyroute2 package?  We are considering it for setting up some routes we need on the amphora.17:42
johnsomI see that it has been proposed off and on for various things, being the last I see.17:43
johnsomHowever it isn't currently in global-requirements17:43
*** bharath has joined #openstack-lbaas17:43
mesteryjohnsom: I'd be in favor of it, yes. It's now dual-licensed (GPL and AV2), so that was the biggest hurdle.17:49
*** SumitNaiksatam has joined #openstack-lbaas17:51
johnsomOk, cool.  I will probably start making motions to get it added.17:51
xgermanjohnsom I can help you with adding17:53
johnsomOk.  I'm reading up on the process now17:53
*** KunalGandhi has quit IRC18:24
*** ig0r__ has joined #openstack-lbaas18:24
*** bana_k has quit IRC18:31
*** bana_k has joined #openstack-lbaas18:31
*** KunalGandhi has joined #openstack-lbaas18:31
*** kbyrne has quit IRC18:34
*** kbyrne has joined #openstack-lbaas18:35
*** KunalGandhi has quit IRC18:48
*** KunalGan_ has joined #openstack-lbaas18:48
sballemestery: Just saw you comment around the license. What GPL are we talking about?18:48
sballejohnsom: maybe you know what version of GPL we are talking about?18:48
mesterysballe: pyroute2 is dual licensed, so we can likely use it18:49
*** ig0r_ has quit IRC18:50
*** minwang2 has quit IRC18:51
sballeOk let's check with our our IP attorney.18:51
*** ig0r_ has joined #openstack-lbaas19:04
*** openstackgerrit has quit IRC19:46
*** chadix has joined #openstack-lbaas19:46
*** openstackgerrit has joined #openstack-lbaas19:47
*** ig0r_ has quit IRC19:56
openstackgerritMerged openstack/octavia: Pluggable controller worker
xgermanblogan yt?19:57
xgermanyou like to talk UDP design19:57
bloganxgerman: im here and there some :)19:58
xgermanok, so what do we want to send in those UDP packages/19:58
*** minwang2 has joined #openstack-lbaas19:58
xgermanMember status? Listener status? Stats?19:59
xgermanI like to send everything but I am also greedy :-)19:59
bloganhow much room do we have?19:59
xgerman64 KB I think19:59
bloganwell that should be plenty19:59
xgermanyeah and we gzip19:59
rm_workyou have to be careful with sending anything that is N-count20:00
rm_worklike members20:00
rm_workotherwise there is a bit of room...20:00
bloganwell hopefully we have limits on the number of members a lb can have20:00
bloganwe dont now of course20:00
rm_workah true20:00
bloganxgerman: would listener status be a check that haproxy is running?20:01
xgermanyes, you get all of that from the stats socket20:01
bloganxgerman: oh yeah i forgot about that, okay yeah if we could send all that info obviously i'd like that20:02
xgermansame here20:02
xgermanyou will also get stats from that socket aka bytes in/out20:02
bloganbc that adds better health monitoring knowing haproxy is running, however we may still need an extra bit of info to make sure it is receiving packets on the vip20:03
bloganbut i tink that can be something we as a follow up20:03
bloganbc not exactly sure what to do otehr than fire off a packet to the vip20:03
xgermanI think harpy has some monitoring page you can activate20:04
bloganlol harpy20:04
xgermanso you could query that over the VIP20:04
xgermanjust make sure to narrow down IPs for that page20:04
bloganhmmm would that make it accessible to users?20:04
xgermanwell they enter over the VIP so...20:05
blogani know20:05
blogannot sure i like them being able to access taht20:05
blogantehn again it might be a requested feature one day20:06
xgermanyep, now if you are HP you would just say the user has to monitor that20:06
xgermanand he has to set up some status page in his web app20:06
*** abdelwas has joined #openstack-lbaas20:08
*** openstackgerrit has quit IRC20:16
*** openstackgerrit has joined #openstack-lbaas20:17
abdelwasHi, this is Sherif20:24
abdelwasI am getting the following an error when I try to create a load balancer in octavia.20:24
abdelwasHere is the error details:
abdelwasAny suggestions ?20:25
xgermantghere has been some subnet changes… but not sure20:25
abdelwasThe subnet id in the VIP table is always null20:25
bloganabdelwas: did you provie a subnet_id in the vip field through the API?20:26
bloganare you using the client?20:26
bloganthe neutron client20:26
abdelwasI am using the neutron client20:26
abdelwasand the rest api driver20:27
bloganwhat task is it failing in?20:27
bloganon i see the line20:27
abdelwasPlugVIP, during the allocate_vip method20:28
abdelwasI get this debug message: DEBUG octavia.controller.worker.tasks.network_tasks [-] Allocate_vip port_id 443d31a2-e824-4539-a259-266389dc2c5e, subnet_id None,ip_address20:29
abdelwasthen everything fails apart afterwards.20:29
openstackgerritEvgeny Fedoruk proposed openstack/neutron-lbaas: Fixing Radware LBaaS v2 driver bugs
bloganthat exception is coming back from the plug_vip method20:31
abdelwasHow the subnet_id is passed at the Plug_VIP task? could it be reltated to changes from network_id to subnet_id20:32
blogansubnet_id will be part of the loadbalancer object and vip object, so it sholud be populated20:33
bloganallocate_vip should have populated subnet_id of the vip20:33
bloganwhich in this case looks like subnet_id is not populated20:33
abdelwasYeah but why it could be NULL, if the neutron command passed "private_net" lets say20:34
blogannot sure yet, would have to breadcrumb it20:34
bloganfollow the breadcrumbs20:34
bloganor just run debugger20:35
TrevorVI'm fairly certain this is a known issue when using octavia through neutron lbaas.  pothole I think you were the one that knows about this one20:35
johnsomYes, I think this is an known issue when going through client/lbaas api.20:39
johnsomIt was introduced with a patchset that switched network_id and subnet_id or something20:39
blogani thought that got fixed20:39
bloganyeah the octavia driver passes subnet_id now20:41
johnsomHmm, ok.  Maybe I will do a debug session with Sherif and help him figure it out20:41
bloganim starting one too20:42
bloganabdelwas: what patchset are you using of the octavia driver reveiw?20:42
minwang2i can try to us postman to create a lb see it is working  that way20:42
abdelwasJust an updated master20:42
bloganabdelwas: master won't matter, your localrc is set up to pull down a specific review in neutron-lbaas so your devstack pulls down the octavia driver since it has not been merged into master20:43
bloganabdelwas: so if you look in yoru localrc where it does the enable_plugin of neutron_lbaas, it should have a ref/changes/*/* at the end of the line20:45
*** openstackgerrit has quit IRC20:46
abdelwasHmm, let me check this20:47
*** openstackgerrit has joined #openstack-lbaas20:47
abdelwasI actually change what localrc do20:49
abdelwasWhen I checkout different branch.20:49
blogando you checkout the octavia driver reveiw?20:49
abdelwasNo! I just learned about that from you now :)20:50
abdelwasWhich patch set I should consider20:50
abdelwasOK let me test this20:50
bloganwell how was neutron lbaas-loadbalancer-create going to octavia before?20:50
abdelwasgive me few minutes. I will figure it out20:51
abdelwasforget about what I said20:53
abdelwasI didn't touch this part localrc is corret20:53
abdelwasand the driver is patched with this
minwang2i tried to use postman to create lb against octavia, it always gave me pending_create status20:54
abdelwascheck Phillip comment here:
abdelwasGot it!20:58
bloganabdelwas: should be patched with the 16th patchset20:58
blogannot 15th20:58
bloganminwang2: is it slow fro the amphora to become active?20:59
abdelwasThank ouy20:59
minwang2i am testing it now20:59
bloganup the timeout limit21:00
*** woodster_ has quit IRC21:00
*** openstack has joined #openstack-lbaas21:08
bloganbut basically you rebasing master locally could cause a lost patchset on a parent, if i had pushed up a patch set after you did a git review -d on yours, then that patchset would ahve been overwritten21:09
bloganits fine right now so no harm done, just be careful in the future and make sure you have the up-to-date parent patches before you do a rebase and push21:10
minwang2how can i make sure that i have the up-to-date parent patches21:11
minwang2just click the button’ rebase change’ ?21:12
bloganyeah, but there's no way to be 100% certain as there will always be a time gap between the time you check and the time you push21:13
bloganso just check before you push21:13
minwang2ok, good to know21:14
openstackgerritBanashankar k proposed openstack/octavia: Excluding virtual interfaces like eth0:1 while reading the stats file /sys/class/net/{interface}/statistics/{type}_bytes
openstackgerritTrevor Vardeman proposed openstack/octavia: Decouple amp driver from network using new models
openstackgerritTrevor Vardeman proposed openstack/octavia: Adding amphora failover flows
*** TrevorV has quit IRC21:20
pothole Can we get some eyes one these, please?21:22
minwang2i tried to create lb by postman, it is always in pending_create status21:24
potholeIf you query the lb after you create it is the status the same?21:25
bloganpothole: was just starting up on that, but i gotta get barbican running again21:26
bloganminwang2: is the amphora not going active in nova?21:27
potholeK, if you need a hand let me know blogan21:27
potholeI have an array of manequin hands in the closet I will ship via fedex if you so need.21:27
minwang2i put it here21:30
minwang2the lb’s status in db are pending_create, there are 2 nova instance, they are active21:30
bloganwell you'd have to look at the logs of the controller worker then21:31
*** jorgem has joined #openstack-lbaas21:44
*** ig0r__ has quit IRC21:51
xgermanpothole gave you +2 on one of the patches; rm_work?22:06
rm_workwhich one?22:07
rm_worki was looking at one of blogan's patches then got into a 1.5h meeting22:07
*** jorgem has quit IRC22:08
potholeI gave him no such thing!22:08
bloganpothole will have to push up a new patchset22:09
potholeok, i misread that ><22:10
openstackgerritPhillip Toohill proposed openstack/octavia: Hooking TLS together
*** jorgem has joined #openstack-lbaas22:11
minwang2"debuginfo": "Traceback (most recent call last):\n\n  File \"/usr/local/lib/python2.7/dist-packages/wsmeext/\", line 84, in callfunction\n    result = f(self, *args, **kwargs)\n\n  File \"/opt/stack/octavia/octavia/api/v1/controllers/\", line 62, in get_all\n    [lb_types.LoadBalancerResponse])\n\n  File \"/opt/stack/octavia/octavia/api/v1/controllers/\", line 57, in _convert_db_to_type\n    converted =22:12
minwang2[_convert(db_obj) for db_obj in db_entity]\n\n  File \"/opt/stack/octavia/octavia/api/v1/controllers/\", line 49, in _convert\n = lb_types.VIP.from_data_model(\n\n  File \"/opt/stack/octavia/octavia/api/v1/types/\", line 48, in from_data_model\n    return cls(**data_model.to_dict())\n\nAttributeError: 'NoneType' object has no attribute 'to_dict'\n",22:12
minwang2    "faultcode": "Server",22:12
minwang2    "faultstring": "'NoneType' object has no attribute 'to_dict'"22:12
minwang2this when i tried to retrieve the latest lb that i created22:12
minwang2from the db table, this new lb’s provisioing status is active22:12
rm_workxgerman: which patchset? the one he just submitted?22:20
rm_worki am doing reviews right now22:20
rm_worki can add another to my queue22:21
*** jorgem has quit IRC22:25
*** jorgem has joined #openstack-lbaas22:26
*** jorgem has quit IRC22:27
xgermanand he overwrote my +2 so never mind :-)22:28
potholeForgot things have merged with other migrations. Had to update revisions22:28
rm_workwas gonna say, i don't see you ever giving that first one a +222:29
xgermanyeah, need to +2 some more22:30
*** apuimedo has quit IRC22:31
potholebloogan is saying that might be wrong also, so give me a sec. i got distracted with other things22:31
xgermanyep, I was holding my horses...22:32
*** apuimedo has joined #openstack-lbaas22:36
*** harlowja has quit IRC22:44
*** harlowja has joined #openstack-lbaas22:44
rm_workfinally finished reviewing blogan's22:46
rm_workbut on an old patchset >_<22:46
rm_workwill try to move them over22:46
*** abdelwas has quit IRC22:48
rm_workgerman and i had the same issue with call_count22:48
rm_workalmost identical comments22:48
rm_workall moved to new patchset22:49
rm_workblogan: ^^22:49
rm_workusing call_count is almost always just lazy22:51
rm_workI guess blogan has been in a hurry recently :P22:51
bloganha it was laziness22:51
rm_workyou lazy bastard22:52
bloganbut i would still use it in addition22:52
bloganbc i'd like to know if it is ever called more than i expect22:52
rm_workif you check the specific calls, it implicitly checks call_count22:52
rm_workbecause it will fail the assert if there are non-listed calls22:52
rm_worki think22:53
rm_workand a list of
bloganthat doesn't exist22:53
rm_workyes it does22:53
rm_workassert_called_once_with() doesn't22:53
xgermanI usually do assert_any_call_with22:53
bloganthat one does!22:53
xgermanof course order is hard to test22:53
bloganassert_called_with exists too damnit22:53
*** minwang2 has quit IRC22:53
rm_workah yeah assert_has_calls22:54
rm_workis what you want i guess22:54
bloganno bc that won't fail if there is other calls22:54
rm_workbut the ones i mentioned do exist22:54
rm_workit should22:54
bloganneed to use any_order argument then22:55
rm_workah yeah22:55
blogannever looked at that one22:55
rm_work" There can be extra calls before or after the specified calls."22:55
rm_workso i guess you do22:55
rm_workyeah unless "any_order"22:55
bloganbut what i fi want to test order and call count?22:55
*** minwang2 has joined #openstack-lbaas22:55
rm_workactually i think any_order=True still doesn't guarantee call count22:56
rm_workthe wording is weird22:56
blogansounds like it does22:56
bloganbut it woudl be weird that any_order=True and False has different behavior with that22:56
rm_workit has the same behavior22:56
rm_workso, no guarantee22:56
bloganoh yeah it doesn't sound like it would22:56
bloganyeah i dont think it would assert call count either22:57
bloganso i stand by my original statement22:57
rm_workdo both22:57
openstackgerritPhillip Toohill proposed openstack/octavia: Hooking TLS together
rm_workdamnit pothole i'm trying to review that22:57
rm_workare you done? :P22:57
bloganif he tested the migration ran he is22:57
bloganpothole: did you test the migration this time?22:58
rm_workheh k23:08
*** apuimedo has quit IRC23:12
bloganalright responded to comments, will fix later tonight, im going home23:13
rm_workpothole: some issues with the tests on yours, but otherwise giving it a +223:15
rm_workkk blogan23:15
bloganrm_work: did you test out functionality too wtih barbican?23:15
rm_worki haven't run it in devstack, if that's what you mean23:15
rm_workbut it should work23:15
bloganoh should it eh?23:15
rm_workyes :P23:15
rm_workI am assuming pothole has run it in devstack23:16
bloganim going to test it out, if yall want to merge it fine by me actually, bc we can do follow up bug fixes23:16
rm_workwell, it's more than an assumption, since i have seen him running this code IIRC23:16
rm_workpothole: comments on that?23:16
potholeYes, ive had this running, not since the migration update23:17
openstackgerritGerman Eichberger proposed openstack/octavia: Adds method to update server.pem to Amphora REST spec
potholebut, that shouldnt really matter and nothing else has changed since it was last running23:18
potholein other words, i have not ran it today23:18
bloganpothole, rm_work: they still haven't fixed the issue with teh barbican client not allowing a normal use to store a secret?23:20
bloganactually sounds like a barbican api problem23:20
potholelol, comments on stuff that's been there and i didnt touch ><23:20
rm_workmy only comments were on changed code? I thought23:21
rm_workit was green23:21
rm_worki'll check again23:21
potholeno :P23:21
potholenone were on changed code23:21
rm_workoh prolly it came in as part of a rebase23:21
rm_worki was viewing changed from 4:623:21
bloganpothole 14:16 sounds like a chapter:verse i'd like to read23:22
blogani really should go home23:22
bloganand i really am23:22
potholehit a pothole on the way23:22
bloganonly if you're walking across the street23:23
rm_workpothole: well I gave you a +2 anyway <_<23:23
rm_workso no complaining23:23
pothole:P lol, thank you23:23
rm_workany others need review?23:24

Generated by 2.14.0 by Marius Gedminas - find it at!