Thursday, 2017-10-05

*** sshank has quit IRC00:00
*** fnaval has quit IRC00:56
openstackgerritLingxian Kong proposed openstack/octavia-tempest-plugin master: Create scenario tests for loadbalancers
kongjohnsom, rm_work, i hope the lastest patch will make you happy01:06
kongi will rebase other patches as soon as it's merged01:07
openstackgerritOpenStack Proposal Bot proposed openstack/python-octaviaclient master: Updated from global requirements
*** SumitNaiksatam has joined #openstack-lbaas01:27
*** sanfern has joined #openstack-lbaas01:44
*** SumitNaiksatam has quit IRC01:48
openstackgerritMerged openstack/python-octaviaclient master: Updated from global requirements
*** tongl has joined #openstack-lbaas02:08
*** dayou has quit IRC02:45
*** tongl has quit IRC02:48
*** yamamoto has joined #openstack-lbaas02:48
*** links has joined #openstack-lbaas03:48
*** avtar_ has joined #openstack-lbaas04:09
*** SumitNaiksatam has joined #openstack-lbaas04:10
*** armax has quit IRC04:12
*** armax has joined #openstack-lbaas04:13
*** armax has quit IRC04:13
*** armax has joined #openstack-lbaas04:13
*** armax has quit IRC04:14
*** gcheresh_ has joined #openstack-lbaas04:14
*** armax has joined #openstack-lbaas04:14
*** armax has quit IRC04:14
*** armax has joined #openstack-lbaas04:15
*** armax has quit IRC04:15
*** armax has joined #openstack-lbaas04:16
*** armax has quit IRC04:16
*** eezhova has joined #openstack-lbaas04:41
*** reedip has quit IRC04:42
*** tongl has joined #openstack-lbaas04:50
*** chlong_ has joined #openstack-lbaas04:51
*** gcheresh_ has quit IRC04:57
*** eezhova has quit IRC05:07
*** sanfern has quit IRC05:25
*** sanfern has joined #openstack-lbaas05:30
*** tongl has quit IRC06:02
avtar_i am using haproxy plugin and followed the guide
avtar_but getting this error =>
*** chlong_ has quit IRC06:13
*** rcernin has joined #openstack-lbaas06:24
*** spectr has quit IRC06:34
*** spectr has joined #openstack-lbaas06:34
*** tesseract has joined #openstack-lbaas07:07
*** pcaruana has joined #openstack-lbaas07:08
*** salmankhan has joined #openstack-lbaas08:00
openstackgerritSlawek Kaplonski proposed openstack/neutron-lbaas master: Allow to extend sub_resource attributes
*** salmankhan has quit IRC08:07
*** spectr has quit IRC08:09
*** spectr has joined #openstack-lbaas08:09
*** eezhova has joined #openstack-lbaas08:10
*** spectr has quit IRC08:17
*** spectr has joined #openstack-lbaas08:18
*** daidv has joined #openstack-lbaas08:19
daidvjohnsom, hi08:19
*** daidv has quit IRC08:29
*** daidv has joined #openstack-lbaas08:30
openstackgerritOpenStack Proposal Bot proposed openstack/neutron-lbaas-dashboard master: Imported Translations from Zanata
*** yamamoto has quit IRC09:16
*** yamamoto has joined #openstack-lbaas09:20
*** eezhova has quit IRC09:34
*** eezhova has joined #openstack-lbaas09:38
*** yamamoto has quit IRC10:02
*** yamamoto has joined #openstack-lbaas10:06
openstackgerritVu Cong Tuan proposed openstack/neutron-lbaas master: Remove SCREEN_LOGDIR from devstack
openstackgerritVu Cong Tuan proposed openstack/octavia master: Remove SCREEN_LOGDIR from devstack
*** yamamoto has quit IRC10:12
*** yamamoto has joined #openstack-lbaas10:19
*** yamamoto has quit IRC10:41
*** krypto has joined #openstack-lbaas10:44
*** spectr has quit IRC10:54
*** yamamoto has joined #openstack-lbaas11:01
*** yamamoto has quit IRC11:06
avtar_i am unable to start octavia worker . Here are the logs11:12
*** ianychoi has quit IRC11:46
*** yamamoto has joined #openstack-lbaas12:00
*** atoth has joined #openstack-lbaas12:07
*** leitan has joined #openstack-lbaas12:25
*** yamamoto has quit IRC12:39
*** yamamoto has joined #openstack-lbaas12:41
*** links has quit IRC12:44
*** yamamoto has quit IRC12:46
*** catintheroof has joined #openstack-lbaas12:50
*** catintheroof has quit IRC12:51
*** yamamoto has joined #openstack-lbaas12:51
*** catintheroof has joined #openstack-lbaas12:51
*** spectr has joined #openstack-lbaas13:02
*** yamamoto has quit IRC13:06
*** salmankhan has joined #openstack-lbaas13:25
*** salmankhan has quit IRC13:36
*** leitan has quit IRC13:37
*** spectr has quit IRC13:40
*** krypto has quit IRC13:41
*** salmankhan has joined #openstack-lbaas13:43
*** chlong_ has joined #openstack-lbaas13:46
*** salmankhan has quit IRC13:47
*** spectr has joined #openstack-lbaas13:57
*** yamamoto has joined #openstack-lbaas14:06
*** spectr has quit IRC14:07
*** spectr has joined #openstack-lbaas14:08
*** yamamoto has quit IRC14:13
*** sanfern has quit IRC14:24
*** sanfern has joined #openstack-lbaas14:25
*** tongl has joined #openstack-lbaas14:26
*** tongl has quit IRC14:30
*** fnaval has joined #openstack-lbaas14:35
*** eezhova has quit IRC14:41
*** armax has joined #openstack-lbaas14:41
*** yamamoto has joined #openstack-lbaas14:48
*** yamamoto has quit IRC14:48
johnsomavtar_ You are missing the transport_url setting in the [DEFAULT] block in the octavia.conf for oslo messaging rabbitmq15:09
johnsomI am guessing you are not running with one of the deployment tools or devstack.  Do you need a sample config file from a devstack install?15:09
avtar_My conf has transport_url. Yeah I am not using devstack or packstack or any automatic stack, the deployment is manual.15:11
avtar_Sample config file will be helpful15:11
johnsomOk, give me a minute to do a gist15:12
johnsomSo that is an example config generated by our devstack plugin.15:16
johnsomIt is not necessarily tuned for your environment and assumes that the rest of the OpenStack infrastructure is working (nova, neutron, rabbit, mysql, etc.)15:17
avtar_Okay i will configure this carefully.15:18
johnsomDocs for the config are here:
*** leitan has joined #openstack-lbaas15:20
avtar_for the rest of the files modification from this link is sufficient ?15:22
johnsomThat is a really high level guide, so I can't say it is going to talk to all of the settings.15:22
johnsomIt should get you most of the way there however15:23
johnsomWe have not had time to write a step by step install guide yet.  Unfortunately.  But we are here to help with questions, etc.15:24
*** Alex_Staf has joined #openstack-lbaas15:25
*** eezhova has joined #openstack-lbaas15:31
*** cody-somerville has joined #openstack-lbaas15:32
*** pcaruana has quit IRC15:39
*** yamamoto has joined #openstack-lbaas15:49
*** spectr has quit IRC15:51
*** cody-somerville has quit IRC15:53
*** cody-somerville has joined #openstack-lbaas15:55
*** tongl has joined #openstack-lbaas15:58
*** yamamoto has quit IRC15:58
*** leitan has quit IRC16:09
*** rcernin has quit IRC16:13
*** rcernin has joined #openstack-lbaas16:13
*** rcernin has quit IRC16:14
*** tesseract has quit IRC16:29
*** sanfern has quit IRC16:33
*** sanfern has joined #openstack-lbaas16:34
*** leitan has joined #openstack-lbaas16:36
*** tesseract has joined #openstack-lbaas16:50
*** eezhova has quit IRC16:56
*** csomerville has joined #openstack-lbaas16:57
*** cody-somerville has quit IRC16:59
*** leitan has quit IRC17:04
*** yamamoto has joined #openstack-lbaas17:11
*** sshank has joined #openstack-lbaas17:19
*** tesseract has quit IRC17:22
*** sshank has quit IRC17:24
*** sshank has joined #openstack-lbaas17:24
*** SumitNaiksatam has quit IRC17:24
*** sshank has quit IRC17:28
*** sshank has joined #openstack-lbaas17:39
*** SumitNaiksatam has joined #openstack-lbaas17:51
rm_workaugh i don't understand how we use AZs17:58
rm_workwhich is sad, because i think i'm the one that added the code for it17:59
rm_workbut the real question is, how does *nova* use AZs17:59
rm_workI guess it just ... picks one? when it schedules17:59
rm_workand anti-affinity server groups are per host, not per az17:59
rm_workso we are likely going to provision servers with no guarantees about AZ anti-affinity18:00
*** SumitNaiksatam has quit IRC18:00
*** SumitNaiksatam has joined #openstack-lbaas18:01
*** sanfern has quit IRC18:05
*** yamamoto has quit IRC18:24
*** yamamoto has joined #openstack-lbaas18:27
johnsomRight, but you can add az anti-affinity to nova...  grin18:28
*** yamamoto has quit IRC18:32
rm_workah, can you?18:41
rm_workor you mean, *I* can :P18:41
rm_workso that is the next thing i'm tackling actually...18:42
rm_workbut probably this piece will be downstream unless i can think of a way to make it actually apply to people with less custom nova18:42
*** chlong_ has quit IRC18:44
openstackgerritMichael Johnson proposed openstack/octavia master: Add dependent packages to config reference
*** chlong_ has joined #openstack-lbaas19:00
*** leitan has joined #openstack-lbaas19:03
*** avtar_ has quit IRC19:03
*** sshank has quit IRC19:13
openstackgerritMichael Johnson proposed openstack/octavia master: Adds automatic ERD generation for the Octavia DB
*** leitan has quit IRC19:42
*** chlong_ has quit IRC19:44
*** JudeC has joined #openstack-lbaas19:47
*** tomtomtom has quit IRC19:59
*** sshank has joined #openstack-lbaas20:26
xgerman_rm_work we have anti-affinity groups so you can leverage that… not sure how to make nova to distribute that over AZs thiugh20:39
rm_workyeah that's the point20:42
rm_workthe server-groups with anti-affinity don't do anything with regard to AZs20:42
rm_workthey only have any effect on hosts20:42
rm_workerr, make guarantees with regard to hosts20:42
*** catintheroof has quit IRC20:56
JudeCHey kong, is there anything special I need to do to get these tests to run, I keep getting tenacity.RetryError: RetryError[<Future at 0x7f698191dc10 state=finished raised MismatchError>]21:02
rm_workjohnsom: errr so how do I make a blueprint now? storyboard just has "tasks"?21:07
rm_workis that no longer a thing?21:07
johnsomThe story is the blueprint21:07
johnsomtasks are just the parts to make it happen21:07
rm_workso bugs are also stories, same as blueprints21:08
johnsomYeah, pretty much21:09
openstackgerritLingxian Kong proposed openstack/octavia-tempest-plugin master: Create scenario tests for loadbalancers
kongJudeC: i think you need to set a big value to 'OS_TEST_TIMEOUT' in .testr.conf21:09
johnsomWe can still use the RFE tag too21:09
JudeCYeah I set that to 2000 seems to just want to run and run. maybe I need to rebuild my devstack.21:10
kongJudeC: i set it to 86400 :-)21:11
JudeCoh wow ok haha21:11
kongit may take 10min to wait for the lb to be running21:11
*** rcernin has joined #openstack-lbaas21:22
johnsomkong If it is taking you 10 minutes, you don't have nested virtualization enabled (virtualbox?)21:35
kongjohnsom: yes, maybe, i am running a devstack environment in a vm in our cloud21:36
rm_workyeah probably you don't21:37
rm_workjohnsom: ah i thought of an issue with the option2 maintenance endpoints21:38
rm_work /v2/octavia/amphorae/maintenance DELETE: Unset maintenance mode on a list of amps by ID or all amps by AZ21:38
rm_workDELETE normally doesn't have a body...21:38
rm_workis it OK to have one? or ...21:39
johnsomWhy does it need a body?21:39
rm_workshould we use URL params? and if so, should we also use url_params for the POST?21:39
rm_workto say which amp list, or which AZ name to unset maintenance?21:39
johnsomI thought we would have a maintenance ID in the URL21:40
rm_workah so that'd go back to having a maintenance table and tracking maintenance entities21:40
rm_workwhich is option 121:40
johnsomYeah, I haven't read option 2 today, but I thought it was  just about amps21:41
rm_workamps and AZ21:41
johnsomIn which case it's just a put/post right?  there is no record to delete21:41
rm_workyou said the option2 endpoints made sense, maybe you hadn't read the POST/DELETE21:41
rm_worki thought you had21:41
rm_worki'm like halfway through the spec, lol21:42
johnsomI think you were still working on it as I needed to leave21:42
rm_workk. mind glancing now?21:43
rm_worki've fleshed it out more21:43
johnsomYeah, give me a minute21:43
*** rcernin has quit IRC22:06
*** yamamoto has joined #openstack-lbaas22:14
*** leitan has joined #openstack-lbaas22:17
*** fnaval has quit IRC22:20
johnsomrm_work Yeah, ok, I would just make POST do both enable/disable and forget about DELETE22:22
johnsomProbably want a get on the amp if that isn't in the other amp patch22:23
*** leitan has quit IRC22:25
johnsomOne evil thought I have however, on the mutability(ha) of the LB while one or more amps is in maint...  Flows aren't atomic, so if we aren't locking the whole LB, what happens when a flow touches the amp multiple times and we "enable" the amp half way through...22:26
openstackgerritMerged openstack/octavia-dashboard master: Pull action button right without floating around
rm_workwe can set a var in the flow for MAINT_MODE at the very beginning22:27
rm_workand that would be set for the whole duration of the flow, no?22:28
johnsomYeah, that is probably the answer.  Would the config pushed by exiting maint be "old" or partial?22:28
openstackgerritMerged openstack/octavia-dashboard master: Update health monitor max retries correctly
rm_workprobably we can wait until the LB isn't PENDING to push?22:29
rm_worklike, on the push, try to get a lock on the LB the same way failover does?22:29
rm_workand retry until we can?22:29
*** leitan has joined #openstack-lbaas22:31
xgerman_failover is suppsoed to be a “short” operation - maintenance is probably long22:31
xgerman_so retrying likely not good22:32
johnsomrm_work for the GET should we just use the amp info endpoint?  Is there value in a maint one?22:32
rm_workjohnsom: i was considering that, i am not sure since it is on amphora_health22:33
rm_workso it'd require another lookup22:33
rm_workbut it's doable22:33
johnsomNah, just a DB join22:33
rm_workjohnsom: erm...22:33
rm_workremember we avoid doing that22:33
rm_workstill trying to enforce a hard separation between that table and everything else22:33
*** leitan_ has joined #openstack-lbaas22:33
rm_workat some point we might just give up on that, if we think it really will never be an issue...22:34
rm_workbut I still have dreams of moving the table to something ... faster than mysql22:34
johnsomOh, I know it will be an issue, I just think we will still have more work to do when we get there.22:34
rm_workxgerman_: so, what i'm talking about retrying is something quick, triggered on the END of a maintenance. not for the duration of the maintenance.22:34
xgerman_so why would you need that table for maintenance? Can’t you juts have status=maintenance on teh amp22:34
johnsomSo, yeah, maint endpoint GET makes sense22:34
rm_workxgerman_: it needs to be part of the HM staleness query22:35
rm_workso it needs to be side-by-side with `busy`22:35
*** leitan has quit IRC22:35
xgerman_yes,  but you stil can set the status to MAINT and use that?22:35
rm_workI will just ... post this spec, I guess. I'll make some tweaks, I was already almost done22:35
johnsomYeah, it needs to be in amp-health22:35
xgerman_I undertand that we need to set it busy but that doesn;t mean we can’t have another filed22:36
xgerman_I also dislike “retrying” - what happens if we crash in between…22:37
xgerman_wonder why we can’t “evacuate” an AZ and then repopulate as we see fit22:37
rm_workthat's another option22:38
rm_workbut significantly more complex22:38
rm_work... I think22:38
rm_workI'm listing it under "alternatives"22:38
xgerman_well, you mark the server/AZ unscheduable in nova22:38
johnsomWelll, most people will probably live-migrate using nova22:39
xgerman_then figure out which amps we have their, failover, done22:39
xgerman_so what problem are we solving in that case?22:39
xgerman_I am fine with aiding in migration but reducing the HA for a maintenance seems iffy22:40
xgerman_(or would we take the whole LB offline?)22:40
*** leitan_ has quit IRC22:40
johnsomIt should be in the spec, but his use case is to recover a bunch of amps after maintenance without doing failovers22:41
johnsomIt's a bit use case specific IMO22:41
xgerman_mmh, so his customers are ok with their HA pair being compromised during potentially unscheduled maintenance — sounds wrong but I will give the benefit of the dount22:42
rm_worki'm assuming the actual downtime would be for however long an individual HV was down22:45
rm_workso like... a few minutes hopefully?22:45
rm_workjust for reboot22:45
rm_workwhich is shitty, but don't run in SINGLE22:45
rm_workand in ACTIVE_STANDBY, whatever, it's a blip22:45
xgerman_well, for a few minutes you will run in single22:45
rm_workbut THAT should be fine22:45
xgerman_maybe for GoDaddy ;-) Personally I would failover… what if the machine doesn’t come up after reboot, etc.22:49
xgerman_my view is jaded since I have seen more HW/Kernel issues in the last month then in my complete life22:50
rm_workin that case, when the amp is set out of maintenance mode, the normal failover process would kick in22:51
rm_worksince it would be stale22:51
rm_workmaintenance flag just blocks failover due to old update-time22:51
rm_workif there aren't new heartbeats coming in by the time it is unset, then failover will happen immediately22:52
xgerman_ok, so if the box doesn’t come up quick enough we will failover regardless22:52
openstackgerritMerged openstack/octavia-dashboard master: Being able to change session persistence of pool
rm_workxgerman_: responded on
johnsomI will throw my $0.02 in23:05
rm_workdoing a test render of the spec and then i'll post it23:05
*** sshank has quit IRC23:06
rm_workjohnsom: uhhh i assume you didn't mean to +A23:07
johnsomYep, that was a "ah-C#$@" moment23:07
rm_workk :P23:07
johnsomI don't think it caught it23:07
johnsomI was going to reply again and mention the list full content thing is the norm now too sicnce you can filter columns via the API23:08
johnsomJust missed on the buttons23:08
rm_workOK submitting my draft spec23:18
*** fnaval has joined #openstack-lbaas23:19
rm_worktear it apart plz23:19
rm_work... as soon as i find how to reference the storyboard story23:19
johnsomStory: #23:20
johnsomTask: #23:20
xgerman_ok, +A’d — never noticed the pagination links before23:20
johnsom# = 12345423:20
openstackgerritAdam Harwell proposed openstack/octavia master: Spec: Amphora Maintenance Mode
xgerman_ok, commented…23:43
*** SumitNaiksatam has quit IRC23:44
kongjohnsom, rm_work, could you please take a look at again to make sure all you comments are solved?23:44
rm_worki think not all -- still wanted the server methods in a common util place23:45
rm_workso i can use them in other tests23:45
kongrm_work: it's not here?
kongrm_work: which method do you expect?23:46
rm_workhmm let me see if i am looking at the old patch again >_<23:48
rm_work_copy_file and such23:49
rm_workare all still in test_basic_ops23:49
kongyou need _coyp_file, too?23:49
rm_workbut those are needed if i want to get a backend up and running23:49
rm_workI want to be able to do: "create_backend" or something like that23:49
rm_workin any test23:49
rm_workand have backend server ready for use as a member23:49
kongok, will move them23:50
*** kbyrne has quit IRC23:57
*** kbyrne has joined #openstack-lbaas23:58

Generated by 2.15.3 by Marius Gedminas - find it at!