Thursday, 2017-06-08

« Wednesday, 2017-06-07 Index Friday, 2017-06-09 »
*** leitan has quit IRC00:04
*** leitan has joined #openstack-lbaas00:07
*** fnaval has quit IRC00:34
*** fnaval has joined #openstack-lbaas00:36
*** leitan has quit IRC00:42
*** armax has quit IRC00:48
*** armax has joined #openstack-lbaas00:56
*** armax has quit IRC00:59
*** ipsecguy has quit IRC01:07
*** csomerville has joined #openstack-lbaas01:19
*** cody-somerville has quit IRC01:20
*** leitan has joined #openstack-lbaas01:21
openstackgerritMerged openstack/python-octaviaclient master: Add member commands to client  https://review.openstack.org/46303501:28
*** ipsecguy has joined #openstack-lbaas01:28
*** sanfern has quit IRC01:32
rm_workeandersson: still failing i guess? https://review.openstack.org/#/c/471611/01:33
*** JudeC has joined #openstack-lbaas01:34
*** dayou has quit IRC01:36
openstackgerritJude Cross proposed openstack/python-octaviaclient master: Fix listener command to include PROXY  https://review.openstack.org/47199101:40
*** dayou has joined #openstack-lbaas01:56
*** csomerville has quit IRC01:57
*** fnaval has quit IRC02:01
*** armax has joined #openstack-lbaas02:02
*** fnaval has joined #openstack-lbaas02:04
*** yamamoto_ has quit IRC02:14
*** eanderson_ has joined #openstack-lbaas02:15
*** eanderson_ is now known as eandersson_02:15
eandersson_rm_work, yea, but it isn't using the new patch02:15
eandersson_> 2017-06-07 21:39:33.472349 |     inst_method = types.MethodType(go_to_method, None, Navigation)02:16
*** JudeC has quit IRC02:16
*** yamamoto_ has joined #openstack-lbaas02:17
eandersson_Since it's just getting the latest tar.gz file and installing that02:18
eandersson_>  Collecting horizon from http://tarballs.openstack.org/horizon/horizon-master.tar.gz#egg=horizon02:18
eandersson_We will need to wait until the patch has been merged02:18
openstackgerritMerged openstack/python-octaviaclient master: Add l7policy commands to client  https://review.openstack.org/46243102:22
openstackgerritMerged openstack/python-octaviaclient master: Add l7rule commands to client  https://review.openstack.org/46280802:22
openstackgerritMerged openstack/python-octaviaclient master: Add healthmonitor commands to client  https://review.openstack.org/46383002:22
openstackgerritMerged openstack/python-octaviaclient master: Update help text for all commands  https://review.openstack.org/46644102:23
*** sanfern has joined #openstack-lbaas02:54
*** yamamoto_ has quit IRC03:04
*** leitan has quit IRC03:23
*** links has joined #openstack-lbaas03:34
*** gans has joined #openstack-lbaas03:48
*** gcheresh_ has joined #openstack-lbaas04:17
*** gcheresh_ has quit IRC04:24
*** pcaruana has joined #openstack-lbaas04:31
*** kobis has joined #openstack-lbaas04:34
*** kobis has quit IRC04:39
*** kobis has joined #openstack-lbaas04:43
*** kobis has quit IRC04:45
*** yamamoto has joined #openstack-lbaas04:48
*** reedip is now known as reedip_afk04:54
*** gans has quit IRC04:55
*** pcaruana has quit IRC04:57
*** leitan has joined #openstack-lbaas05:00
*** gans has joined #openstack-lbaas05:01
*** leitan has quit IRC05:04
*** gcheresh_ has joined #openstack-lbaas05:13
openstackgerritMerged openstack/octavia master: Handle log message interpolation by the logger  https://review.openstack.org/47194705:21
*** aojea has joined #openstack-lbaas05:26
*** reedip_afk is now known as reedip05:39
*** belharar has joined #openstack-lbaas05:45
openstackgerritOpenStack Proposal Bot proposed openstack/neutron-lbaas-dashboard master: Updated from global requirements  https://review.openstack.org/47206805:50
openstackgerritOpenStack Proposal Bot proposed openstack/neutron-lbaas-dashboard master: Updated from global requirements  https://review.openstack.org/47209406:06
*** aojea has quit IRC06:07
*** armax has quit IRC06:08
*** armax has joined #openstack-lbaas06:08
*** armax has quit IRC06:09
*** rcernin has joined #openstack-lbaas06:15
*** belharar has quit IRC06:54
*** pcaruana has joined #openstack-lbaas06:57
*** JudeC has joined #openstack-lbaas06:58
*** gans has quit IRC07:08
*** tesseract has joined #openstack-lbaas07:12
*** belharar has joined #openstack-lbaas07:15
*** JudeC has quit IRC07:21
*** aojea has joined #openstack-lbaas07:22
*** leitan has joined #openstack-lbaas07:24
*** leitan has quit IRC07:28
eandersson_rm_work:  > The old method of installing a version of _() in the builtins namespace is deprecated07:28
eandersson_I was just browsing the olso.i18n documentation and noticed the above dislaimar07:29
eandersson_https://docs.openstack.org/developer/oslo.i18n/usage.html#creating-an-integration-module07:29
*** eandersson_ has quit IRC07:32
*** links has quit IRC08:25
*** links has joined #openstack-lbaas08:38
*** yamamoto has quit IRC08:58
*** belharar has quit IRC09:08
*** gans has joined #openstack-lbaas09:16
*** gans has quit IRC09:25
*** gans has joined #openstack-lbaas09:30
*** gongysh has joined #openstack-lbaas09:32
*** gcheresh_ has quit IRC10:03
*** gcheresh_ has joined #openstack-lbaas10:03
*** med_ has quit IRC10:31
*** Guest72612 has joined #openstack-lbaas10:35
*** sanfern has quit IRC10:52
*** aojea has quit IRC11:15
*** gongysh has quit IRC11:37
*** gans has quit IRC11:41
*** kobis has joined #openstack-lbaas11:43
*** atoth has joined #openstack-lbaas12:03
*** ianychoi has quit IRC12:27
*** ianychoi has joined #openstack-lbaas12:28
*** catintheroof has joined #openstack-lbaas12:32
*** sanfern has joined #openstack-lbaas12:41
*** yamamoto has joined #openstack-lbaas12:55
*** aojea has joined #openstack-lbaas13:08
*** yamamoto has quit IRC13:10
*** yamamoto has joined #openstack-lbaas13:14
*** links has quit IRC13:15
*** sanfern has quit IRC13:20
*** sanfern has joined #openstack-lbaas13:21
*** cpuga has joined #openstack-lbaas13:23
*** cpuga has quit IRC13:25
*** cpuga has joined #openstack-lbaas13:25
*** leitan has joined #openstack-lbaas13:35
openstackgerritMerged openstack/octavia-tempest-plugin master: Updated from global requirements  https://review.openstack.org/47118013:35
*** leitan has quit IRC13:35
*** leitan has joined #openstack-lbaas13:35
*** fnaval_ has joined #openstack-lbaas13:49
*** fnaval has quit IRC13:52
*** ianychoi_ has joined #openstack-lbaas13:54
*** ianychoi has quit IRC13:57
*** ianychoi_ is now known as ianychoi13:57
*** matt-borland has joined #openstack-lbaas13:57
*** chlong has joined #openstack-lbaas14:02
*** aojea has quit IRC14:12
openstackgerritMerged openstack/octavia master: Add check O346 to disallow backslash line continuation  https://review.openstack.org/47150314:20
*** kobis has quit IRC14:26
openstackgerritMerged openstack/octavia master: Minor code cleanup in l7policy controller  https://review.openstack.org/47117614:33
*** aojea has joined #openstack-lbaas14:34
*** aojea has quit IRC14:36
*** aojea has joined #openstack-lbaas14:40
*** armax has joined #openstack-lbaas14:47
*** gcheresh_ has quit IRC14:54
*** fnaval_ has quit IRC15:01
*** aojea has quit IRC15:12
*** rcernin has quit IRC15:16
*** fnaval has joined #openstack-lbaas15:18
*** cpuga has quit IRC15:28
*** cpuga has joined #openstack-lbaas15:28
*** yamamoto has quit IRC15:30
*** aojea has joined #openstack-lbaas15:32
*** cpuga_ has joined #openstack-lbaas15:32
*** yamamoto has joined #openstack-lbaas15:35
*** cpuga has quit IRC15:35
*** bzhao_ has joined #openstack-lbaas15:35
*** gans has joined #openstack-lbaas15:37
*** yamamoto has quit IRC15:39
*** armax has quit IRC15:40
*** leitan has quit IRC15:42
*** cpuga_ has quit IRC15:42
*** cpuga has joined #openstack-lbaas15:43
*** aojea has quit IRC15:53
*** reedip_ has joined #openstack-lbaas16:06
openstackgerritMichael Johnson proposed openstack/octavia-dashboard master: Replace SortedDict with OrderedDict and fixing Python 3.5 test  https://review.openstack.org/47161116:08
johnsomLet's see what we get with that16:09
*** cpuga has quit IRC16:12
*** cpuga has joined #openstack-lbaas16:13
*** leitan has joined #openstack-lbaas16:18
openstackgerritMichael Johnson proposed openstack/octavia-dashboard master: Replace SortedDict with OrderedDict and fixing Python 3.5 test  https://review.openstack.org/47161116:23
openstackgerritOpenStack Proposal Bot proposed openstack/octavia master: Updated from global requirements  https://review.openstack.org/47232516:23
openstackgerritOpenStack Proposal Bot proposed openstack/neutron-lbaas-dashboard master: Updated from global requirements  https://review.openstack.org/47232616:23
*** tesseract has quit IRC16:32
*** yamamoto has joined #openstack-lbaas16:37
*** armax has joined #openstack-lbaas16:43
*** yamamoto has quit IRC16:43
*** fnaval has quit IRC16:49
*** gans has quit IRC16:53
*** pcaruana has quit IRC17:00
*** aojea has joined #openstack-lbaas17:14
*** eanderson_ has joined #openstack-lbaas17:16
*** eanderson_ is now known as eandersson_17:17
*** aojea has quit IRC17:19
openstackgerritErik Olof Gunnar Andersson proposed openstack/octavia-dashboard master: Replace SortedDict with OrderedDict and fixing Python 3.5 test  https://review.openstack.org/47161117:21
eandersson_johnsom: Added horizon back now as the upstream fix was merged17:22
*** amotoki is now known as amotoki_away17:28
johnsomWell, I was waiting for the tar to actually be built, it is still in the queue17:30
eandersson_Ah17:30
eandersson_I  was just looking at the status of that17:30
eandersson_I assumed it was done already as I slept in :D17:31
johnsomYeah, it is sitting in the "post" queue to build the tar17:31
*** amotoki_away is now known as amotoki17:35
*** sanfern has quit IRC17:39
leitanhi guys17:42
leitantrying to test the brand new generated amphora image17:42
leitanto test the keepalived patches17:42
leitangetting on the log "Cannot connect to instance retrying"17:43
johnsomHi ther17:43
johnsomthere17:43
leitanbut i can connect to the instances just fine, via the mgmt network17:43
leitanill try rollabking to the old image, to check if something else got broken in the meantime17:43
leitanbut any thoughts ?17:43
johnsomWell, if you can connect over the mgmt net, take a look at the amphora-agent logs.  /var/log/amphora-agent.log and /var/log/syslog17:44
eandersson_The tar is finally there... fingers crossed :D17:45
johnsomYeah, waiting to force a recheck17:45
eandersson_bah didn't realize it was still running -_-17:45
*** bzhao_ has quit IRC17:45
leitanjohnsom: great, ill take a look there and let you know17:45
johnsomThat last run isn't finish yet17:45
leitanjohnsom: guess nothing weird on the agent http://paste.openstack.org/show/612037/17:47
*** eandersson_ has quit IRC17:49
johnsomHow about the syslog?17:49
leitanjohnsom: checking17:50
leitanhmm, inside the amphora image, the api on the port 9443 but if i curl it inside the same VM i get curl: (56) Recv failure: Connection reset by peer17:58
leitandoing it to localhost or the actual mgmt ip17:58
leitaninside or outside the amphorae vms same result17:59
*** aojea has joined #openstack-lbaas18:02
leitanhmm, maybe coause its SSL , ill try with https instead18:03
leitanyep, was that, getting a 404 now, curling the amphorae vm, from the controller18:05
leitanso its reaching it ...18:05
leitani dont understand why the rest_api.driver says that connot reach the vm18:05
*** reedip_ has quit IRC18:10
*** fnaval has joined #openstack-lbaas18:15
eanderssonSuccess :D18:17
johnsomOh good!18:18
eanderssonor wel... npm-run-test is probably gonna fail18:18
eanderssonSo we still need that project update18:19
johnsomWhich?18:19
*** cpuga has quit IRC18:19
eanderssonhttps://review.openstack.org/#/c/471887/18:19
*** cpuga has joined #openstack-lbaas18:19
johnsomAh, ok.18:19
leitan9443 answering same 404 inside the container where the api is running, im going to debug there to see what "connection error" is returning18:22
*** aojea has quit IRC18:22
*** ipsecguy_ has joined #openstack-lbaas18:23
*** ipsecguy has quit IRC18:24
eanderssonjohnsom, I'll switch to chrome and update the dependencies to match horizon18:28
eanderssonand see if that fixes the issue18:28
johnsomCool, thanks!18:29
rm_workeandersson: ah fix it so the test is actually PASSING? :P18:29
johnsomI like passing tests....  Grin18:30
eanderssonYea - I mean sounds crazy I know18:31
robcresswelleandersson: I'd try just the deps first before switching18:33
robcresswellI think its just a fix to karma.conf, at a guess...18:33
eanderssonI managed to get the tests running, but hit a new issue18:47
leitanwell18:55
leitani managed to debug the issue18:55
leitanand i think its a bug18:55
*** amotoki is now known as amotoki_away18:55
leitanfor some reason, the amphorae image that is spawned first (maybe the master?) get assigned two or more ips from the management network18:57
leitanif the first ip, is assigned as a port first to the VM, its reacheable and the balancer gets configured ok18:58
leitanbut, if for some reason the rest_api driver tries to contact the other binded ip address that are not the one that the VM actually HAS inside18:58
leitanfails18:58
xgerman_sounds weird18:58
leitanso the real issue here is why the amphorae that is launched first, gets assigned more than one ip from the management network18:59
leitanim looking at all the balancers i created18:59
leitanand for all is the same18:59
leitanthe first one, has two or more ip addresses from the mgmt network18:59
xgerman_huh, they should only get one IP but two on the VIP network18:59
leitanthe second one, has one and one on the tunneled (OK)18:59
leitanxgerman_: thats correct, for the tunneled network the behaviour is correct, 1 ip for each and the vip as another address on the allowed_address pairs19:00
leitanbut the issue here, is the first amphorae getting more than one ip on the management network19:00
xgerman_yeah, we ask nova to spin it up in that network and assign IP19:01
leitanthis causes that my loadbalancers, are created ok ACTIVE only if the ip that uses the rest_api driver to contact the balancer is the same as the one that finally got inside the vm from the two or more that got assigned19:01
xgerman_yep, if the REST can’t talk to the VM it will be shot and result in ERROR19:01
leitanxgerman_: thats correrct and thats whats happening, but i realized the behaviour adding some print debugs where the rest_api was trying to connect, and saw the multiple ips stuff19:02
leitanxgerman_: to be more graphic, here is a pic: http://i66.tinypic.com/10nb7u0.png19:04
xgerman_yeah, that is definitely wrong19:04
leitanas you can see the first LB got the .200 and .198 ip address, the one that is on the vm and working is the .200, but the rest_api_driver is trying to contact the .198, without success of course19:04
xgerman_mmh, I don’t know where that seconf ip could come from… we allocate/plug all our IP on tenant networks only (unless you passed in mgmt network for some subnet)19:06
rm_workleitan: does your nova add a default vif?19:07
rm_worksome deployments do19:07
rm_workfor example, mine does :P19:07
leitanrm_work: no19:07
rm_workok19:07
*** cpuga has quit IRC19:07
leitanrm_work xgerman_ if i manually launch many instances, on the lb_mgmt_net, the issue is not present19:09
*** cpuga has joined #openstack-lbaas19:09
leitanso im discarding the nova -> neutron circuit19:09
rm_workhmm19:09
leitanbut this happen with every single balancer pair i create with octavia, with the first vm, two or more ip address on the mgmt net, some LBs has 3 ips  hahaha19:11
*** gcheresh_ has joined #openstack-lbaas19:13
rm_workwhat does your config look like for "lb_mgmt_networks"19:14
leitanrm_work: one sec19:15
leitanamp_boot_network_list = 6e668339-70ef-49c6-82ad-ad3f2ad432c519:17
leitanrm_work: thats the only place when i refer to a network19:19
rm_workhmmm19:19
rm_workset debug=True and post your whole worker log maybe for one create19:19
rm_worki can scan through and see what's going on maybe...19:19
johnsommaybe do a "neutron net-list | grep 6e668339-70ef-49c6-82ad-ad3f2ad432c5"19:19
johnsomSee if there is multiple matching?19:20
leitanrm_work: doing it right now19:20
leitanjohnsom: no, | 6e668339-70ef-49c6-82ad-ad3f2ad432c5 | lb-mgmt-net                                        | 122d58d6-ffa5-49ed-99b0-be71aa899f3e 172.30.3.0/24    |19:20
rm_workok19:20
johnsomAnd "neutron net-show 6e668339-70ef-49c6-82ad-ad3f2ad432c5" just to make sure it doesn't have multiple subnets19:20
rm_workunfortunately i'm about to head to lunch, so19:20
rm_workafter i get back i can look19:20
leitanrm_work: k19:21
leitanjohnsom: | subnets                   | 122d58d6-ffa5-49ed-99b0-be71aa899f3e |19:21
leitanonly one19:21
johnsomAnd a "nova show" on the amp with two IPs?19:22
leitanjohnsom: buuuuuut, the allocation pool is not contiguos19:22
leitanjohnsom: maybe something with that ?19:22
leitanjohnsom: allocation_pools  | {"start": "172.30.3.165", "end": "172.30.3.250"}19:22
leitan{"start": "172.30.3.10", "end": "172.30.3.160"}19:22
leitanthat cant be it ... but just saying19:22
johnsomHow about the nova show?19:23
leitanjohnsom: http://paste.openstack.org/show/612050/19:25
johnsomSo that one actually has three mgmt Ips, contiguous19:27
leitanjohnsom: yes, that discard the non-contigous alloc pool malefic theory19:28
johnsomCheck the flavor, since you are using a different flavor than we create, it might be something in there, but otherwise a run with debug=true and a copy of that o-cw log would be my next step.19:28
johnsomVery unusual.  FYI, we take that network ID from the config and pass it to nova, we don't hot plug it or even ask for anything special, just pass it in as the boot network19:29
leitanjohnsom: only an anti-affinity spec on the flavor http://paste.openstack.org/show/612051/19:30
leitanjohnsom: yes, but in that case, the fact that i have spinned 50 vms in the lb-mgmt-network without getting dup ips ... make it weirder19:32
leitanand happens always with the first amphorae, always19:32
leitanill do the full debbuged run now19:33
*** blogan has joined #openstack-lbaas19:34
johnsomWhat version of Octavia are you running?19:37
leitanshould be ocata, its a kolla container19:40
leitanoctavia-0.10.1.dev10619:41
johnsomleitan Can you check this: https://review.openstack.org/#/c/358928/2/octavia/controller/worker/tasks/compute_tasks.py19:51
johnsomIt should be in that 0.10.1 build, but I am wondering if it is missing19:51
*** JudeC has joined #openstack-lbaas19:52
*** leitan has quit IRC20:01
eanderssonhttps://review.openstack.org/#/c/471611/720:04
eanderssonhttps://review.openstack.org/#/c/471611/20:04
eandersson^ It's passing finally20:04
*** harlowja has quit IRC20:08
johnsomeandersson That doesn't seem to have the schemaform fix, are you working on another patch for that?20:12
*** matt-borland has quit IRC20:17
*** matt-borland_ has joined #openstack-lbaas20:17
*** gcheresh_ has quit IRC20:18
*** borland__ has joined #openstack-lbaas20:27
*** matt-borland_ has quit IRC20:28
openstackgerritMerged openstack/python-octaviaclient master: Add hacking checks  https://review.openstack.org/47116820:29
johnsomJudeC Any chance you can address this comment so we can do a client release soon?  https://review.openstack.org/#/c/471991/20:33
JudeCHa literally 2 sec away from pushing that20:33
johnsomAwesome20:34
JudeCI noticed some other crap that needed fixing so I added it to that.20:34
johnsomOk20:34
JudeCjust documentation stuff20:34
openstackgerritJude Cross proposed openstack/python-octaviaclient master: Fix command to include PROXY and update metavars  https://review.openstack.org/47199120:34
JudeCBoom20:34
johnsomYeah, I skimmed the docs.  So there might be a few things I missed.  I did have a fail on not noticing the proxy proto though..  Sigh20:35
JudeCHey its my code, if anyone should be bummed it me. :P20:35
johnsomAh, found an issue....20:36
eanderssonYea - johnsom - I rather keep that separate as well20:37
johnsomOk, NP.  If you weren't going to do it, I was going to try Rob's advice, but if you have it I can move on to other issues...20:38
eanderssonI managed to get it past the schema thing, but ran into a new issue20:38
johnsomJudeC PROXY is a backend protocol only.  It sends extra info on backend requests about the front end client20:38
*** leitan has joined #openstack-lbaas20:39
eanderssonbut I probably won't have much more time today, so if you want have a go at it20:39
eanderssonI say we keep them separate though as that isn't blocking anymore20:39
JudeCjohnsom: ohhhh ok fixing now.20:39
johnsomMaybe, I need to check on some bugs that were filed, so not sure if I would have a lot of time either20:40
johnsomYeah, separate is great with me.20:40
johnsomJudeC http://logs.openstack.org/99/471499/1/gate/gate-python-octaviaclient-pep8-ubuntu-xenial/eb227be/console.html#_2017-06-08_20_29_04_19154920:42
johnsomWhile you are there....20:42
JudeCoh weird why does my pep8 pass hmm20:43
johnsomBecause that patch adds the check for them.20:43
*** leitan has quit IRC20:44
*** kbyrne has quit IRC20:48
*** kbyrne has joined #openstack-lbaas20:50
openstackgerritJude Cross proposed openstack/python-octaviaclient master: Fix command to include PROXY and update metavars  https://review.openstack.org/47199120:54
*** borland__ has quit IRC21:03
rm_workjohnsom / JudeC: OK if Proxy is backend only then it was right before and I told him wrongly -- but I thought it was not21:07
rm_workjohnsom: that means I need to change MY patch here: https://review.openstack.org/#/c/463851/5/octavia/common/constants.py21:08
rm_workPROXY should not be a LISTENER protocol?21:08
rm_workonly POOL?21:08
johnsomAh, ok.  Yeah, it takes info from the frontend connection (client IP, etc.) and sendings it to the backend before the main TCP connection.21:08
rm_workJudeC: ok my bad, sorry about that :P21:08
rm_workI thought proxy was both21:08
johnsomWow, bummer, ok, somewhere we introduced a regression21:24
johnsomfailovers with public IP vips don't work.21:24
johnsomElena reported them here: https://bugs.launchpad.net/octavia/+bug/169444021:24
openstackLaunchpad bug 1694440 in octavia "Spares pool failover errors for a load balancer created in public subnet" [Undecided,New]21:24
johnsomWell, with spares pool or act/stdby21:24
johnsomSo, I'm going to dig into this....21:25
johnsomI hate that the gates don't have the horse power to test this...  We need to get containers working....21:27
*** blogan has quit IRC21:31
*** leitan has joined #openstack-lbaas21:39
leitanguys21:42
leitansorry got in a meeting21:42
leitanlonger than expected21:42
leitanrm_work johnsom here is the full debug run: https://pastebin.com/uz07YwcC21:42
leitanjohnsom: ill check what you send me when i got home21:43
rm_workjohnsom: yeah I'm tracking the lxd stuff in the background21:43
rm_workjohnsom: seems nova-lxd devstack plugin broke, it no longer builds a working nova+lxd21:43
rm_workwas trying to get someone to talk to / work with us21:43
*** leitan has quit IRC21:44
rm_workbut also busy with a bunch of other stuff T_T21:44
*** catintheroof has quit IRC21:50
*** chlong has quit IRC21:53
*** bzhao_ has joined #openstack-lbaas22:06
*** cpuga has quit IRC22:32
*** marcin123 has joined #openstack-lbaas22:41
marcin123What are the plans of implementing in Neutron/Octavia dynamically updated of "neutron lbaas-loadbalancer-status <lb_name>" ? To actually see offline member in API manner22:47
johnsommarcin123 What driver are you using?  This works fine with the octavia driver22:48
*** tonygunk has quit IRC22:53
marcin123I am using Octavia driver I am checking my config because I saw this not working.22:54
johnsomleitan There is something in your nova config adding the extra IPs.  grep that log for "a3c64b9d-79eb-48b0-b2fd-fb6e57ab0824".  You will see that we poll nova to see when the status goes active so there are a number of debug messages while nova is building the instance, noted by "status": "BUILD".  As the build progresses nova is adding those interfaces to the instance.  If you grep for REQ you can see how we are22:55
johnsomasking nova to build it with one network.22:55
johnsommarcin123 Did you enable the event streamer?   octavia.conf event_streamer_driver = queue_event_streamer22:55
johnsommarcin123 That tells octavia to send update events to neutron22:56
johnsomhttps://docs.openstack.org/developer/octavia/main/configref.html#health_manager.event_streamer_driver22:59
eanderssonI just had him configure it23:00
johnsomYeah it's an evil required to keep neutron up to date.  That mess goes away with the new Octavia v2 API23:00
johnsomThough we haven't finished the status API in that yet23:01
*** tonygunk has joined #openstack-lbaas23:01
openstackgerritOpenStack Proposal Bot proposed openstack/neutron-lbaas-dashboard master: Updated from global requirements  https://review.openstack.org/47247023:02
johnsomOh, this hasn't merged yet...  https://review.openstack.org/#/c/471611/  rm_work xgerman_ if you have a minute23:03
marcin123so I was missing that "event_streamer_driver = queue_event_streamer" in /etc/octavia/octavia.conf23:09
rm_worksounds like that'd be the problem, yep! working now?23:11
rm_workjohnsom: +A'd23:11
*** armax_ has joined #openstack-lbaas23:12
johnsomCool, that should fix our proposal bot spam23:12
*** armax has quit IRC23:14
*** armax_ is now known as armax23:14
openstackgerritMerged openstack/octavia-dashboard master: Replace SortedDict with OrderedDict and fixing Python 3.5 test  https://review.openstack.org/47161123:15
*** armax_ has joined #openstack-lbaas23:33
*** armax has quit IRC23:33
*** armax_ is now known as armax23:33
openstackgerritMichael Johnson proposed openstack/octavia master: Fix an issue with failover on VIP net without DHCP  https://review.openstack.org/47248623:49
*** JudeC has quit IRC23:49
johnsomThat seems to fix the issue, but I want to do a bit more testing.  WIP23:49
*** tonygunk has quit IRC23:53
rm_workAHH I have a similar fix in my own flow johnsom23:53
rm_workhttps://review.openstack.org/#/c/435612/25/octavia/controller/worker/flows/amphora_flows.py23:54
rm_workI found the default one to be a bit broken23:54
rm_work(for our network, without DHCP)23:54
rm_workI found I always needed to do a ReloadLoadBalancer then ReloadAmphora or stuff wasn't working right23:55
rm_workjohnsom: you need to rebase your other octavia-dashboard commit?23:56
johnsomOh at a minimum.  I kind of started that and didn't finish because of the horizon bugs we just squished.  It's marked WIP.23:57
johnsomJuggling priorities.  I will probably get back to that soon.23:57
*** harlowja has joined #openstack-lbaas23:57
« Wednesday, 2017-06-07 Index Friday, 2017-06-09 »

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!