Wednesday, 2017-07-12

johnsomIt will slow down inserts and updates00:00
rm_workah hmm00:00
johnsomI'm in favor of doing the project_id, we will hit that a bunch for sure00:01
johnsomBut we should be cautious about adding too many00:01
johnsomAh crumb00:05
johnsomYeah, we still have an issue in the API around deleting objects in error.  Some of these controllers are passing in "PENDING_UPDATE" as the target status instead of the true "PENDING_DELETE".00:06
johnsomThis could use a nice refactor too00:06
*** reedip_ has joined #openstack-lbaas00:07
*** tongl has quit IRC00:09
*** reedip_ has quit IRC00:30
*** sticker has quit IRC00:44
*** atoth has quit IRC01:32
*** sanfern has quit IRC03:02
*** aojea has joined #openstack-lbaas03:13
*** aojea has quit IRC03:17
rm_workjohnsom: ah, darn, you fixing that?03:21
rm_workI am happy to do some refactors03:21
rm_workif you point me at the issue you're seeing03:21
rm_workwhere do the *controllers* pass through PENDING_UPDATE/DELETE?03:22
johnsomI don't have the cycles right now to work on it, plus I'm still noodling it.  Let me get you some pointers03:22
rm_workk03:22
johnsomThis is the bug: https://bugs.launchpad.net/octavia/+bug/161395603:23
openstackLaunchpad bug 1613956 in octavia "Can not delete listener from a loadbalancer in ERROR status" [High,Triaged] - Assigned to Michael Johnson (johnsom)03:23
rm_workjohnsom: i don't see where any controllers pass a status?03:24
rm_workthey just use handler.delete()03:24
johnsomSo, the problem is this check: https://github.com/openstack/octavia/blob/master/octavia/db/repositories.py#L65703:25
rm_workI always thought that was odd03:25
johnsomWhich only looks at the target status, which for these is really the right status for LB and listener (pending_update) when you are deleting a pool for example.03:25
rm_workwait what03:26
rm_workwhenever something goes to error, the objects above it do too, no?03:26
johnsomThe refactor is like every controller has a derevitive of this: https://github.com/openstack/octavia/blob/master/octavia/api/v2/controllers/member.py#L9003:26
rm_workwhat stays in PENDING_UPDATE?03:26
johnsomhttps://github.com/openstack/octavia/blob/master/octavia/api/v2/controllers/member.py#L9803:26
johnsomWe are telling it we want those to go to PENDING_UPDATE which is right if we are deleting a member03:27
rm_workoh03:27
rm_worki see, i think03:27
rm_worki'll have to noodle that s well03:27
johnsomYeah03:27
johnsomThere are some weird combinations here.  LB in error, listener is not, LB has pool with members.  Standard delete on LB won't work.  Standard delete on the member to burn up the chain won't work.03:28
johnsomThis week I have a few internal deliverables, so can't take on any big projects.03:30
*** krypto has joined #openstack-lbaas03:52
*** fnaval has quit IRC04:10
*** cpuga has joined #openstack-lbaas04:26
*** yamamoto has joined #openstack-lbaas04:33
*** gongysh has joined #openstack-lbaas04:35
*** pcaruana has joined #openstack-lbaas04:41
*** diltram has quit IRC04:55
*** diltram has joined #openstack-lbaas04:58
*** pcaruana has quit IRC04:58
*** cpuga has quit IRC05:01
*** armax has quit IRC05:14
*** armax has joined #openstack-lbaas05:14
*** armax has quit IRC05:15
*** armax has joined #openstack-lbaas05:15
*** armax has quit IRC05:16
*** armax has joined #openstack-lbaas05:16
*** armax has quit IRC05:16
*** armax has joined #openstack-lbaas05:17
*** armax has quit IRC05:17
*** armax has joined #openstack-lbaas05:17
*** armax has quit IRC05:18
*** armax has joined #openstack-lbaas05:18
*** armax has quit IRC05:19
*** armax has joined #openstack-lbaas05:19
*** armax has quit IRC05:20
*** cpuga has joined #openstack-lbaas05:21
*** gcheresh_ has joined #openstack-lbaas05:23
*** sanfern has joined #openstack-lbaas05:25
*** aojea has joined #openstack-lbaas05:26
*** cpuga has quit IRC05:28
*** rcernin has joined #openstack-lbaas05:33
openstackgerrittuninger proposed openstack/neutron-lbaas master: Update members data based on haproxy stats data.  https://review.openstack.org/47684805:55
*** slaweq has joined #openstack-lbaas06:02
*** slaweq has quit IRC06:06
*** aojea has quit IRC06:07
*** slaweq has joined #openstack-lbaas06:07
*** krypto has quit IRC06:11
*** gtrxcb has joined #openstack-lbaas06:13
*** gtrxcb has quit IRC06:13
*** vlaza has joined #openstack-lbaas06:16
*** gongysh has quit IRC06:34
*** vlaza has left #openstack-lbaas06:37
*** openstackgerrit has quit IRC06:48
-openstackstatus- NOTICE: FYI, ask.openstack.org is down, review.o.o is slow - please have patience until this is fixed06:51
*** kman has joined #openstack-lbaas06:53
*** openstackgerrit has joined #openstack-lbaas07:08
openstackgerritEvgeny Fedoruk proposed openstack/neutron-lbaas master: Enhancements for the the back-end system  https://review.openstack.org/47961407:08
nmagnezirm_work, o/07:10
*** tesseract has joined #openstack-lbaas07:11
*** kobis has joined #openstack-lbaas07:14
*** kman has quit IRC07:17
*** aojea has joined #openstack-lbaas07:25
openstackgerrittuninger proposed openstack/neutron-lbaas master: Update members data based on haproxy stats data.  https://review.openstack.org/47684807:26
*** pcaruana has joined #openstack-lbaas07:31
*** pcaruana has quit IRC07:41
*** amotoki_away is now known as amotoki07:45
*** pcaruana has joined #openstack-lbaas07:45
-openstackstatus- NOTICE: Gerrit is going to be restarted, due to low performance07:49
-openstackstatus- NOTICE: Gerrit has been successfully restarted07:56
*** krypto has joined #openstack-lbaas08:27
*** pcaruana has quit IRC08:33
*** krypto has quit IRC08:35
*** krypto has joined #openstack-lbaas08:35
*** krypto has quit IRC08:35
*** krypto has joined #openstack-lbaas08:35
rm_worknmagnezi: o/08:44
nmagnezirm_work, o/08:44
rm_workah crap today is meeting day isn't it08:45
nmagnezirm_work, may i ask u to have a look at https://bugs.launchpad.net/octavia/+bug/1690812/ , yesterday i discovered that this issue persists08:45
openstackLaunchpad bug 1690812 in octavia "Octavia active/standby config+ pool with sourceip session persistence configuration- Service is not available and LB is not deleted after test" [Critical,Triaged]08:45
rm_workearly morning08:45
rm_workffff08:45
rm_worki probably won't make it08:45
nmagnezirm_work, yeah.. it's becoming an impossible time for me as well08:46
*** pcaruana has joined #openstack-lbaas08:46
nmagnezirm_work, but we had a democratic vote, so that's fair :)08:46
rm_worklol yes08:47
rm_worki did find out many years ago that democratic often means "worst for everyone"08:47
nmagnezirm_work, lol. i tend to agree :)08:48
openstackgerritBernard Cafarelli proposed openstack/octavia master: DIB: drop custom mirror elements  https://review.openstack.org/48258708:48
rm_workwho was it that wanted the new time?08:48
nmagnezii don't remember.. I think it was for folks that reside in India's timezone (no sure which one specifically)08:49
rm_workyou're IST?08:51
rm_workah this is an active/standby bug08:51
rm_workyeah our testing of that is not great at the moment08:51
rm_workbut SOON I'll be running that configuration in my cloud08:51
rm_workso08:52
rm_workit'll get a lot more of a workout :P08:52
nmagnezirm_work, i'm at UTC+03:0008:52
rm_workoh IST isn't what I thought it was08:52
rm_workoh wait it also is08:52
rm_workwtf08:52
rm_workIST is India Standard Time *AND* Israel Standard Time08:52
nmagneziwe should really have a scenario tests job configured with ACTIVE_STANDBY amphoras08:52
rm_workyeah....08:53
rm_workwe need so many gates...08:53
nmagnezimaybe I should put this in the meeting agenda so johnsom will decide is this is sane08:54
nmagnezirm_work, in any case, i'm not sure what's wrong with the haproxy peers config, but if there's something you'd like me to try (manual modifications to haproxy.cfg) I have a live setup for this. and I'd love to submit a patch as soon as we figure out how to change that config08:56
rm_workyeah i'll need to look in the morning i think08:57
openstackgerritBernard Cafarelli proposed openstack/octavia master: DIB: drop custom mirror elements  https://review.openstack.org/48258708:57
nmagnezirm_work, aye :)08:57
*** pcaruana has quit IRC09:32
*** pcaruana has joined #openstack-lbaas09:32
*** sanfern has quit IRC09:55
*** sanfern has joined #openstack-lbaas10:05
*** Alex_Staf has joined #openstack-lbaas10:07
*** sanfern has quit IRC11:08
*** atoth has joined #openstack-lbaas11:18
*** kbyrne has quit IRC12:28
*** kbyrne has joined #openstack-lbaas12:30
openstackgerritJon Schlueter proposed openstack/neutron-lbaas master: removing WebTest from test-deps  https://review.openstack.org/48295412:34
*** kbyrne has quit IRC12:38
*** kbyrne has joined #openstack-lbaas12:39
*** wasmum has quit IRC12:40
*** ctracey_ has joined #openstack-lbaas12:40
*** yamamoto has quit IRC12:41
*** ctracey has quit IRC12:41
*** ctracey_ is now known as ctracey12:41
*** wasmum has joined #openstack-lbaas12:41
*** kbyrne has quit IRC12:44
*** kbyrne has joined #openstack-lbaas12:46
*** sanfern has joined #openstack-lbaas12:48
*** sanfern has quit IRC12:53
*** sanfern has joined #openstack-lbaas12:53
*** yamamoto has joined #openstack-lbaas13:08
*** yamamoto has quit IRC13:13
*** armax has joined #openstack-lbaas13:39
*** cpuga has joined #openstack-lbaas13:49
*** cpuga_ has joined #openstack-lbaas13:50
*** yamamoto has joined #openstack-lbaas13:51
*** cpuga has quit IRC13:54
*** catintheroof has joined #openstack-lbaas13:57
openstackgerritCarlos Puga proposed openstack/octavia master: Spec detailing Octavia service flavors support  https://review.openstack.org/39248514:00
johnsomnmagnezi Commented on that bug14:07
*** catintheroof has quit IRC14:20
*** reedip_ has joined #openstack-lbaas14:56
*** rcernin has quit IRC15:01
*** catintheroof has joined #openstack-lbaas15:03
*** kobis has quit IRC15:12
*** Alex_Staf has quit IRC15:12
*** gcheresh_ has quit IRC15:12
*** kobis has joined #openstack-lbaas15:12
*** kobis has quit IRC15:13
*** fnaval has joined #openstack-lbaas15:22
*** aojea has quit IRC15:40
*** yamamoto has quit IRC15:48
*** yamamoto has joined #openstack-lbaas15:49
*** yamamoto has quit IRC15:51
*** amotoki is now known as amotoki_away15:56
*** evgenyf has joined #openstack-lbaas15:58
*** amotoki_away is now known as amotoki15:59
*** amotoki is now known as amotoki_away15:59
*** tesseract has quit IRC16:05
*** reedip_ has quit IRC16:25
*** catintheroof has quit IRC16:30
*** catintheroof has joined #openstack-lbaas16:31
*** cpuga_ is now known as cpuga16:38
*** evgenyf has quit IRC16:42
nmagnezijohnsom, thanks Michael! I'll recheck the image and report back16:42
johnsomOk16:43
sanfernjohnsom, Do we need to do any specific changes to amphora image to get assigned ipv6 for lb mgmt net?16:50
johnsomNo, it should be all on the controller side16:51
*** yamamoto has joined #openstack-lbaas16:51
sanfernok, my amphora didn't get ipv6 assigned, but nova show reports it16:52
johnsomThis may be of interest to you: https://review.openstack.org/#/c/39120416:52
johnsomIt's for link local, but should give you an idea16:52
johnsomI think it will preference IPv4 if it is available16:53
*** pcaruana has quit IRC16:56
*** gans has joined #openstack-lbaas16:56
sanfernThanks johnsom, let me have a look.16:58
*** fnaval has quit IRC16:58
johnsomOctavia meeting starting soon on #openstack-meeting17:00
*** fnaval has joined #openstack-lbaas17:00
*** evgenyf has joined #openstack-lbaas17:01
*** yamamoto has quit IRC17:02
*** ssmith has joined #openstack-lbaas17:05
ssmithWhich version of openstackclient supports octavia loadbalancer commands?  I'm using pip to install17:06
johnsompip install python-octaviaclient17:07
johnsomThat will install our plugin17:07
ssmithWhat about this? https://docs.openstack.org/python-octaviaclient/latest/usage/osc_cli_plugins.html17:08
johnsomYes, those command will be available after our plugin, python-octaviaclient is installed17:15
*** aojea has joined #openstack-lbaas17:16
*** aojea has quit IRC17:21
ssmithYep, that works.  couldn't find anything referring to that in the documentation though. Thanks.17:22
johnsomYeah, all of OpenStack docs are moving around right now.17:23
johnsomThere is a link from the main page: https://docs.openstack.org/python-octaviaclient/latest/index.html17:23
johnsomFor installation instructions.17:23
johnsomssmith Where would you have expected to find that info?  I might be able to add it to help others17:23
ssmithYes there is but doesn't reffer to anything that it hooks into the openstackclient which led me to beleive that it was a standalone like some others17:24
johnsomAh, gotcha17:24
ssmithWhen the docs reffer to 'openstack loadbalancer list'17:25
johnsomI will update that to be more clear.  I will also be hooking this into our main docs site soon as well.  It's all part of these docs changes17:25
*** gans has quit IRC17:31
ssmithDoes Octavia have a Public endpoint? -->openstack loadbalancer list -->17:32
ssmithpublic endpoint for load-balancer service in RegionOne region not found17:32
johnsomYes, it is load-balancer17:32
johnsomMaybe your keystone is not setup correctly?17:32
johnsom"openstack endpoint list | grep octavia"17:32
ssmithDid that and there's just an internal endpoint.17:33
johnsomHmm, yeah, our devstack plugin creates all three.17:34
ssmithThis is production deployed via ansible17:35
johnsomAh, yeah, I don't think the ansible has been updated for the Octavia v2 API endpoint yet17:35
*** aojea has joined #openstack-lbaas17:37
*** tomtomtom has joined #openstack-lbaas17:39
*** aojea has quit IRC17:42
ssmithIs there an ENV setting to set it to internal?17:43
ssmithWhere could I find a conf setting for public endpoint - I can do it manually for now17:48
xgerman_you would need to change playbooks in ansible which is annoying - or set it your self via keystone17:50
xgerman_but then they won’t be set up in OSA’s load balancer :-(17:50
xgerman_I really need to get around adding some of the Pike features17:51
johnsomI think this is what you are looking for:17:51
johnsom  --os-service-provider-endpoint <auth-service-provider-endpoint>17:51
johnsom                        With v3adfspassword: Service Provider's Endpoint (Env:17:51
johnsom                        OS_SERVICE_PROVIDER_ENDPOINT)17:51
johnsomFrom the openstack command help17:51
johnsomOr maybe this one:17:52
johnsom  --os-url <auth-url>   With token_endpoint: Specific service endpoint to use17:52
johnsom                        (Env: OS_URL)17:52
*** gcheresh_ has joined #openstack-lbaas17:53
*** evgenyf has quit IRC17:55
ssmithI changed the RC file to say export OS_INTERFACE=internal and use the internal URL for keystone but get a 404 error. Does Octavia support Keystone v3?18:00
johnsomYes it does18:01
johnsomCould it be you have an Octavia deployment older than master/Pike?18:01
ssmithProbably it's Ocata18:02
ssmithtomtomtom?18:02
johnsomAh, yeah, you don't have the v2 API in your octavia18:02
ssmithoh boy18:02
johnsomWith Ocata you are running through neutron-lbaas, so you need to use the "neutron" client18:03
*** yamamoto has joined #openstack-lbaas18:04
rm_workssmith: for the record, if you are on an ocata cloud, there is nothing that would prevent you from running pike/master Octavia18:13
rm_workssmith: I'm running octavia from master, on a liberty cloud, with no issues18:13
rm_workin fact I would highly recommend running as close to master as possible18:14
rm_workthe amount of usability work that was done in the Pike cycle is pretty insane18:14
rm_workjohnsom: did I miss anything in the meeting?18:14
ssmithWe're on Newton and think that stable/Newton was installed.  Can we use master (which I'm assuming is Pike)?18:15
*** krypto has quit IRC18:17
*** yamamoto has quit IRC18:20
*** yamamoto has joined #openstack-lbaas18:21
tomtomtomI'm looking at the version we're using18:22
*** sshank has joined #openstack-lbaas18:22
tomtomtomok so is the octavia from about 1 or 2 months ago current enough?  Should I be installing stable/pike to run with newton? or straight up master branch?18:25
tomtomtomis there a way to tell the version I'm using compared to what I should be running?18:25
*** yamamoto has quit IRC18:26
*** jniesz has joined #openstack-lbaas18:29
ssmithCan we use master (which I'm assuming is Pike) on Newton?18:30
xgerman_yes, rm_work does that18:34
rm_workI use master on Liberty :P18:35
rm_worktomtomtom: actually even a couple months back MIGHT be too old18:35
rm_worktomtomtom: we're literally wrapping up the APIv2 stuff this cycle18:35
tomtomtomok then master it is18:36
ssmithHow did you handle the python inconsistences?18:36
johnsomThe other thing to consider is in theory, the python-octaviaclient can be pointed to the the neutron API.  It's un-tested, but for the most part should work.  Though the neutron client is tested, etc.18:37
ssmithThere are other issues we're having too so might as well re-do with the master branch18:38
openstackgerritCarlos Puga proposed openstack/octavia master: Spec detailing Octavia service flavors support  https://review.openstack.org/39248518:38
*** sshank has quit IRC18:47
*** atoth has quit IRC18:59
*** aojea has joined #openstack-lbaas19:03
*** sshank has joined #openstack-lbaas19:03
*** sshank has quit IRC19:05
*** sanfern has quit IRC19:10
*** sanfern has joined #openstack-lbaas19:13
*** yamamoto has joined #openstack-lbaas19:16
*** harlowja has quit IRC19:17
rm_workssmith: python inconsistencies?19:50
rm_workjohnsom: AUGH is there a doc somewhere about how to use this stupid systemctl stuff to restart devstack processes?19:51
rm_workI'm dieing here19:51
rm_workjust want to restart nova and can't remember how, digging through my notes for the last time you told me how19:51
rm_workwhat was wrong with screen T_T19:51
*** cpuga has quit IRC19:52
johnsomhttps://docs.openstack.org/devstack/latest/systemd.html19:52
johnsomThere is that19:52
*** cpuga has joined #openstack-lbaas19:52
*** yamamoto has quit IRC19:54
rm_workperfect, thanks19:54
rm_workbookmarked19:54
*** harlowja has joined #openstack-lbaas20:02
rm_workugh just lost my logs for the past week20:15
rm_workah i guess they're on my ZNC but it's really annoying to get to those20:15
johnsomUgh, yeah, you can also always look back on evesdrop20:16
rm_workah true20:23
rm_workrestarted n-* in devstack and the whole system became non-responsive, eventually rebooted itself20:24
*** gcheresh_ has quit IRC20:24
*** sshank has joined #openstack-lbaas20:24
johnsomOpps20:25
johnsomGeez, 2U servers out there with 3TB of RAM using 128GB DIMMs.  Now we are cooking with gas....20:26
*** sshank has quit IRC20:29
*** sshank has joined #openstack-lbaas20:29
*** jniesz has quit IRC20:46
*** yamamoto has joined #openstack-lbaas20:54
*** yamamoto has quit IRC21:01
nmagnezijohnsom, o/21:20
nmagnezijohnsom, so, should I attempt to add a non-voting active standby amp gate?21:20
*** aojea has quit IRC21:23
*** fnaval has quit IRC21:24
*** cpuga has quit IRC21:30
*** aojea has joined #openstack-lbaas21:32
*** fnaval has joined #openstack-lbaas21:39
*** fnaval has quit IRC21:39
*** fnaval has joined #openstack-lbaas21:40
*** ssmith has quit IRC21:43
johnsomnmagnezi Sorry, was in a meeting.21:44
johnsomWe talked about the need to start spinning up a proper tempest plugin.  I have the repo setup: https://github.com/openstack/octavia-tempest-plugin21:44
johnsomIt's just we need to do the base development for that.21:45
johnsomI fully agree we need that gate, it's just building on good foundations rather than the outdated tempest code we have today21:45
*** catintheroof has quit IRC21:46
*** armax has quit IRC21:51
*** armax has joined #openstack-lbaas21:52
*** aojea has quit IRC21:53
*** armax has quit IRC21:56
*** fnaval has quit IRC22:07
*** rm_work has quit IRC22:32
*** rm_work has joined #openstack-lbaas22:32
*** raginbajin has quit IRC22:35
*** ssmith has joined #openstack-lbaas22:39
*** raginbajin has joined #openstack-lbaas22:42
*** yamamoto has joined #openstack-lbaas22:52
*** armax has joined #openstack-lbaas22:58
*** sshank has quit IRC23:39

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