*** slaweq has joined #openstack-lbaas | 00:39 | |
*** dims has quit IRC | 00:42 | |
*** slaweq has quit IRC | 00:44 | |
*** tzumainn has joined #openstack-lbaas | 02:30 | |
*** tzumainn has quit IRC | 02:58 | |
xgerman_ | mnaser: johnsom os_octavia does not deploy the legacy policy by default: | 03:19 |
---|---|---|
xgerman_ | https://www.irccloud.com/pastebin/ylGLi8fk/ | 03:19 |
*** sanfern has joined #openstack-lbaas | 04:24 | |
*** slaweq has joined #openstack-lbaas | 04:41 | |
*** slaweq has quit IRC | 04:45 | |
openstackgerrit | Ganpat Agarwal proposed openstack/octavia master: Active-Active: ExaBGP amphora L3 distributor driver https://review.openstack.org/537842 | 05:42 |
*** sanfern has quit IRC | 06:50 | |
*** slaweq has joined #openstack-lbaas | 06:51 | |
*** rcernin has quit IRC | 06:56 | |
*** velizarx has joined #openstack-lbaas | 06:56 | |
*** pcaruana has joined #openstack-lbaas | 07:13 | |
*** velizarx has quit IRC | 07:20 | |
*** tesseract has joined #openstack-lbaas | 07:21 | |
*** tesseract has quit IRC | 07:33 | |
*** tesseract has joined #openstack-lbaas | 07:35 | |
*** velizarx has joined #openstack-lbaas | 07:40 | |
*** threestrands has quit IRC | 07:47 | |
*** Alex_Staf has joined #openstack-lbaas | 07:49 | |
*** AlexeyAbashkin has joined #openstack-lbaas | 07:58 | |
openstackgerrit | Jacky Hu proposed openstack/octavia-dashboard master: Allow detail pages to auto refresh upon action https://review.openstack.org/561458 | 08:14 |
openstackgerrit | Jacky Hu proposed openstack/octavia-dashboard master: Allow detail pages to auto refresh upon action https://review.openstack.org/561458 | 08:18 |
openstackgerrit | Jacky Hu proposed openstack/octavia-dashboard master: Allow detail pages to auto refresh upon action https://review.openstack.org/561458 | 08:35 |
*** gans has joined #openstack-lbaas | 08:39 | |
openstackgerrit | Ganpat Agarwal proposed openstack/octavia master: Active-Active: ExaBGP amphora L3 distributor driver https://review.openstack.org/537842 | 08:42 |
nmagnezi | dayou, thanks for the answer :) I'm re-stacking so will review it later today | 08:54 |
dayou | nmagnezi, thanks :) | 08:57 |
nmagnezi | dayou, np. let me know when you think the auto refresh patch is ready | 08:58 |
cgoncalves | dayou, thanks a zillion for the auto-refresh! | 08:59 |
dayou | nmagnezi, https://review.openstack.org/#/c/561457/, you needs some patching on horizon side | 08:59 |
dayou | cgoncalves, my pleasure :) | 09:00 |
*** yboaron has joined #openstack-lbaas | 09:08 | |
*** ipsecguy has quit IRC | 09:13 | |
*** ipsecguy has joined #openstack-lbaas | 09:14 | |
*** gans has quit IRC | 09:23 | |
*** gans has joined #openstack-lbaas | 09:25 | |
*** salmankhan has joined #openstack-lbaas | 09:44 | |
*** gans has quit IRC | 09:48 | |
*** salmankhan has quit IRC | 10:36 | |
*** salmankhan has joined #openstack-lbaas | 10:37 | |
*** salmankhan has quit IRC | 10:44 | |
*** salmankhan has joined #openstack-lbaas | 10:47 | |
*** salmankhan has quit IRC | 10:54 | |
*** salmankhan has joined #openstack-lbaas | 10:54 | |
*** salmankhan has quit IRC | 11:02 | |
openstackgerrit | Jacky Hu proposed openstack/octavia-dashboard master: Allow detail pages to auto refresh upon action https://review.openstack.org/561458 | 11:04 |
*** salmankhan has joined #openstack-lbaas | 11:10 | |
*** salmankhan has quit IRC | 11:23 | |
*** atoth has joined #openstack-lbaas | 11:59 | |
openstackgerrit | Jacky Hu proposed openstack/octavia master: Allow default pool without a name on single create https://review.openstack.org/561600 | 12:27 |
dayou | nmagnezi, auto refresh patch should be ready for review | 12:29 |
*** dims has joined #openstack-lbaas | 12:30 | |
*** yamamoto has quit IRC | 12:38 | |
mnaser | xgerman_: awesome so i'm using the newer one :) | 12:47 |
*** oanson has quit IRC | 12:49 | |
*** irenab has quit IRC | 12:49 | |
*** irenab has joined #openstack-lbaas | 12:57 | |
*** oanson has joined #openstack-lbaas | 12:57 | |
*** ArchiFleKs has quit IRC | 13:00 | |
*** velizarx has quit IRC | 13:01 | |
*** phuoc_ has joined #openstack-lbaas | 13:01 | |
*** yamamoto has joined #openstack-lbaas | 13:01 | |
*** tzumainn has joined #openstack-lbaas | 13:05 | |
*** velizarx has joined #openstack-lbaas | 13:20 | |
*** KeithMnemonic has joined #openstack-lbaas | 13:22 | |
gokhan | hi folks, I installed octavia with openstack ansible . when I try to create loadbalancer ı am getting amphora not found instance couldn't be found error. This is error log: http://paste.openstack.org/show/719298/ . How can ı solve this problem what is your opinions?? | 13:33 |
*** velizarx has quit IRC | 13:36 | |
*** velizarx has joined #openstack-lbaas | 13:43 | |
*** yboaron has quit IRC | 14:39 | |
*** mugsie has quit IRC | 14:53 | |
*** mugsie has joined #openstack-lbaas | 14:53 | |
*** mugsie has quit IRC | 14:53 | |
*** mugsie has joined #openstack-lbaas | 14:53 | |
openstackgerrit | Jacky Hu proposed openstack/octavia-dashboard master: Allow detail pages to auto refresh upon action https://review.openstack.org/561458 | 15:03 |
*** pcaruana has quit IRC | 15:09 | |
*** Alex_Staf has quit IRC | 15:10 | |
*** velizarx has quit IRC | 15:14 | |
xgerman_ | gokhan: the kog before that error should have it tryign to reach the amphora agent until it gives up… when I looked at your network config I didn’t see the veth pair between br-vlan and the octavia conatiner. Wonde rif you cna post an outout of `brctl show` | 15:42 |
xgerman_ | johnsom: rm_work my periodic job for osa borks on the diskimage build (need to replace that part with the download eventually) http://logs.openstack.org/periodic/git.openstack.org/openstack/openstack-ansible/master/openstack-ansible-deploy-translations-ubuntu-xenial/a8d4a5e/job-output.txt.gz#_2018-04-16_06_15_38_965062 | 15:44 |
xgerman_ | anything we know about DOB breakage? | 15:45 |
xgerman_ | DIB | 15:45 |
johnsom | Yeah, DIB and devstack were both impacted by the pip 10 release over the weekend | 16:13 |
xgerman_ | k | 16:20 |
*** yamamoto has quit IRC | 16:41 | |
*** velizarx has joined #openstack-lbaas | 16:45 | |
*** yamamoto has joined #openstack-lbaas | 16:46 | |
*** yamamoto has quit IRC | 16:51 | |
*** samccann has joined #openstack-lbaas | 16:53 | |
*** samccann has quit IRC | 16:53 | |
*** samccann has joined #openstack-lbaas | 16:53 | |
*** winircuser-713 has joined #openstack-lbaas | 16:56 | |
*** winircuser-713 has left #openstack-lbaas | 16:58 | |
*** yamamoto has joined #openstack-lbaas | 17:01 | |
*** gokhan27 has joined #openstack-lbaas | 17:01 | |
*** yamamoto has quit IRC | 17:06 | |
*** yamamoto has joined #openstack-lbaas | 17:11 | |
*** yamamoto has quit IRC | 17:11 | |
gokhan27 | hi xgerman_ again, I created veth pairs. this is my last network config: http://paste.openstack.org/show/719312/ and this is my openstack_user_config.yml file for provider networks: http://paste.openstack.org/show/719313/ | 17:14 |
*** velizarx has quit IRC | 17:15 | |
gokhan27 | do you think which I may miss ? | 17:15 |
*** AlexeyAbashkin has quit IRC | 17:17 | |
xgerman_ | you have the `brctl show` outout? | 17:17 |
xgerman_ | also some `ip a ` on the octavia cintainer is helpul | 17:18 |
gokhan27 | xgerman_: and this output of brctl Show: http://paste.openstack.org/show/719315/ | 17:18 |
xgerman_ | ok, that looks good | 17:18 |
gokhan27 | xgerman_: and output of ip a in Octavia container: http://paste.openstack.org/show/719316/ | 17:20 |
gokhan27 | xgerman_: ı see that Octavia management network is created by service Project not admin Project, is that be problem ? | 17:21 |
xgerman_ | no, shouldn;t be | 17:21 |
xgerman_ | that all looks fine | 17:21 |
xgerman_ | johnsom: and I had out fair share of network debugging :-) | 17:22 |
xgerman_ | there are two possible next steps: | 17:23 |
xgerman_ | 1) enable some more advanced debug settings in Octavia OR | 17:23 |
xgerman_ | 2) boot up a vm on the octavia management net and ping it from an octavia container (after the security groups are not blocking ICMP) | 17:23 |
gokhan27 | xgerman_: I made Octavia network share and public and created an instance with amphora agent image. I can reach this instance and ping it from Octavia container | 17:25 |
xgerman_ | ok, that’s good ;-) | 17:25 |
gokhan27 | xgerman_: when run 'openstack loadbalancer create lb --vip-subnet Octavia-subnet ' | 17:26 |
gokhan27 | instance is created but it is shutdown :( | 17:26 |
xgerman_ | yes, if it can’t reach it’s REST API it will do that | 17:26 |
xgerman_ | mmh, we usually use not the Octavia management net for the VIP net but we have people for whom that works… | 17:28 |
gokhan27 | xgerman_: and ampora instances are taking 2 ips from neutron | 17:29 |
gokhan27 | may be it can not create communicate with neutron correctly | 17:30 |
xgerman_ | yep, one for the menagement net and one for the VIP net | 17:30 |
gokhan27 | xgerman_: hımm can ı use any private subnet for VIP ? | 17:30 |
xgerman_ | yes | 17:31 |
gokhan27 | xgerman_: sorry ı am new to Octavia and ı am asking a lot of questions :( | 17:31 |
gokhan27 | xgerman_: ok ı am trying now with private subnet | 17:31 |
xgerman_ | sounds good | 17:32 |
gokhan27 | xgerman_: after ı source my demo Project and try create loadbalancer , ı am getting http 403 error. ı can not reach Octavia ? What is the reason of this ? | 17:37 |
johnsom | gokhan27 https://developer.openstack.org/api-ref/load-balancer/v2/index.html#faults | 17:38 |
johnsom | 403 means the Role Based Access Control (RBAC) is not allowing you to do that. | 17:38 |
johnsom | gokhan27 https://docs.openstack.org/octavia/latest/configuration/policy.html | 17:39 |
johnsom | If you want the simpler RBAC config (owner or admin) copy https://github.com/openstack/octavia/blob/master/etc/policy/admin_or_owner-policy.json into your /etc/octavia directory as policy.json and restart the API processes | 17:40 |
gokhan27 | johnsom: ok but openstack-ansible-os_octavia does not install policy.yaml with default. I will add it now. | 17:41 |
johnsom | gokhan27: right, it OSA runs with the default advanced RBAC enabled which does not require a policy file | 17:43 |
xgerman_ | yep, but we will create the roles ;-) | 17:43 |
gokhan27 | xgerman_: I really thank you for Octavia role :) it is good work. I have problems because ı do not have enough information about Octavia :( | 17:47 |
xgerman_ | no worries - make sure to submit a patch updating the docs so the next perosn has an easier time | 17:48 |
gokhan27 | xgerman_: yep if ı achieve create loadbalancers, I will submit a patch. where do I need define policy.yaml file in Octavia.conf? | 17:53 |
xgerman_ | https://github.com/openstack/openstack-ansible-os_octavia/blob/master/defaults/main.yml#L33 | 17:54 |
xgerman_ | or you can add the loadbalancer_member role to the user you are using | 17:54 |
*** yamamoto has joined #openstack-lbaas | 17:57 | |
gokhan27 | xgerman_: ok thanks | 17:59 |
*** yamamoto has quit IRC | 18:02 | |
*** sanfern has joined #openstack-lbaas | 18:10 | |
*** tesseract has quit IRC | 18:36 | |
*** harlowja has joined #openstack-lbaas | 18:40 | |
*** AlexeyAbashkin has joined #openstack-lbaas | 18:44 | |
*** atoth has quit IRC | 18:46 | |
*** AlexeyAbashkin has quit IRC | 18:49 | |
*** AlexeyAbashkin has joined #openstack-lbaas | 18:49 | |
gokhan27 | xgerman_: ı am getting errors on health manager http://paste.openstack.org/show/719320/. this is weird | 18:50 |
xgerman_ | yep, so if we recive haleth message from an amphora we don;t have in our DB we will log things like that — so this is ‘ok’ and the system will work/ignore them — but question is why are there amphoras which are not accounted for? | 18:53 |
xgerman_ | (we just added that so maybe need some “nicer” error message) | 18:53 |
gokhan27 | xgerman_: ı can see three amphora on db. I have HA installation and have 3 controllers. may be beacouse of that | 18:57 |
*** yamamoto has joined #openstack-lbaas | 18:58 | |
xgerman_ | yeha, ther emigth be a decrepancy between what’s in the DN and what’s in nova list | 18:58 |
*** AlexeyAbashkin has quit IRC | 18:59 | |
gokhan27 | xgerman_: ok if ı can ignore them, it is not problem :) | 19:00 |
xgerman_ | well, if it’s not too many it’s fine — if it gets out of hand it will chew up resources (we have like mnore then a 100 once and that was a drag) | 19:01 |
openstackgerrit | Adam Harwell proposed openstack/neutron-lbaas master: WIP: Test l7 proxy to octavia https://review.openstack.org/561049 | 19:01 |
*** yamamoto has quit IRC | 19:02 | |
*** openstackgerrit has quit IRC | 19:05 | |
rm_work | xgerman_ / gokhan27 err, i don't think you should see *tracebacks*. That's wonky | 19:07 |
rm_work | and, if you have amps that aren't in your system, that's generally a bad sign, as it would be a symptom of something broken | 19:08 |
xgerman_ | rm_work: yep, we clearly need to improve those error messages | 19:09 |
rm_work | err but why is that happening | 19:11 |
rm_work | eugh | 19:11 |
xgerman_ | 1) If you start amphora by hand | 19:12 |
xgerman_ | 2) I have seen amps related to failing over an ERROR LB | 19:13 |
xgerman_ | or maybe not | 19:13 |
xgerman_ | 3) If we send a nova dleete and nova doesn’t do it | 19:14 |
xgerman_ | (2) is speculation since that might also be part of (3) | 19:14 |
rm_work | yeah 3 is not great if that's happening | 19:18 |
rm_work | something is busted in your env and it's only going to get worse | 19:18 |
rm_work | i had another "big hammer" change for that | 19:18 |
rm_work | but dunno what approach we want to take | 19:19 |
gokhan27 | xgerman_: I create load balancer with private subnet and it is now active :) | 19:20 |
xgerman_ | yeah, you are good | 19:20 |
xgerman_ | rm_work: I am on the fence of hammering #3 - on the one hand people should run an HA nova and if they do planned maintenance stop Octavia - on the other hand we should be able to handle short API outages | 19:21 |
johnsom | I have a patch in flight to fix that traceback error. Now how you got an amp in that state is a different question | 19:24 |
xgerman_ | he said earlier he started one amp by hand | 19:25 |
johnsom | Must have configured it too | 19:26 |
johnsom | https://review.openstack.org/#/c/561369/ | 19:26 |
gokhan27 | xgerman_: I started more than one :) | 19:29 |
*** sanfern has quit IRC | 19:31 | |
*** imacdonn has quit IRC | 19:35 | |
*** imacdonn has joined #openstack-lbaas | 19:35 | |
rm_work | xgerman_: i mean, api outages is one thing, but nova consistently saying "deleted successfully!" and then leaving servers around, is another issue altogether | 19:56 |
rm_work | brb | 19:56 |
rm_work | johnsom: oh, do we have a plan for PIP? is that "not our problem" since it's primarily a DIB issue? is someone working it? | 19:56 |
johnsom | Well, I think devstack is getting back into good shape, there are a few patches there, mine included. | 19:57 |
johnsom | I have been poking at DIB this morning. | 19:57 |
johnsom | It's a bit complicated as cloud-init has a depends-on for the ubuntu pyyaml package. | 19:58 |
johnsom | So, still working on that one | 19:58 |
*** yamamoto has joined #openstack-lbaas | 19:58 | |
*** yamamoto has quit IRC | 20:04 | |
*** gokhan27 has quit IRC | 20:22 | |
*** AlexeyAbashkin has joined #openstack-lbaas | 20:41 | |
*** kbyrne has quit IRC | 20:42 | |
*** kbyrne has joined #openstack-lbaas | 20:44 | |
*** cgoncalves has quit IRC | 20:45 | |
*** AlexeyAbashkin has quit IRC | 20:45 | |
*** harlowja has quit IRC | 20:55 | |
*** yamamoto has joined #openstack-lbaas | 21:00 | |
*** yamamoto has quit IRC | 21:06 | |
rm_work | johnsom: ok, i'm focusing internally until gate issues are resolved then, as it seems it's handled, and i can't do much upstream | 21:16 |
rm_work | unless there's stuff that needs eyes | 21:16 |
* rm_work checks reviews etherpad | 21:16 | |
johnsom | Yeah, everything is stuck at the moment, plus I'm getting pulled internal too | 21:17 |
*** samccann has quit IRC | 21:17 | |
rm_work | k | 21:18 |
nmagnezi | johnsom, rm_work, yikes. and I thought that this breakage only occurs in my node and I should just deploy a clean one | 21:20 |
johnsom | No, it's the pip version 10 issue | 21:21 |
johnsom | It doesn't like distribution packages of python modules | 21:21 |
nmagnezi | :< | 21:21 |
nmagnezi | can't we force it back to pip 9.x until it gets fixed? where is that bump coming from? | 21:22 |
johnsom | Yeah, that is what they did for devstack. | 21:22 |
johnsom | I'm poking at DIB to see if there is another fix or not | 21:22 |
nmagnezi | anything I can do to help? | 21:23 |
nmagnezi | johnsom, ? | 21:26 |
johnsom | Well, I don't want to keep you up all night.... | 21:26 |
nmagnezi | haha | 21:26 |
nmagnezi | it's morning somewhere in the world I guess.. :) | 21:27 |
johnsom | If you are highly motivated: somewhere in here: https://github.com/openstack/diskimage-builder/tree/master/diskimage_builder/elements/pip-and-virtualenv | 21:27 |
johnsom | We need to pin the version of pip like they did for devstack https://review.openstack.org/#/c/561526/1/tools/cap-pip.txt | 21:28 |
johnsom | This is the script https://github.com/openstack-dev/devstack/blob/master/tools/install_pip.sh | 21:28 |
johnsom | devstack uses. | 21:28 |
johnsom | Someone could propose that patch while I keep playing with the elements to see if I can work around the problem package. | 21:29 |
nmagnezi | will try to find my way around there | 21:30 |
nmagnezi | brb | 21:30 |
*** openstackgerrit has joined #openstack-lbaas | 21:34 | |
openstackgerrit | Adam Harwell proposed openstack/octavia master: When SG delete fails on vip deallocate, try harder https://review.openstack.org/549263 | 21:34 |
*** kong is now known as lxkong | 21:45 | |
*** rcernin has joined #openstack-lbaas | 21:51 | |
nmagnezi | johnsom, https://review.openstack.org/#/c/561740/ (I still need to test it) | 21:53 |
johnsom | Cool | 21:54 |
nmagnezi | johnsom,I'm building an image locally, will report soon | 21:58 |
*** yamamoto has joined #openstack-lbaas | 22:02 | |
nmagnezi | johnsom, still not working :< | 22:03 |
nmagnezi | 2018-04-16 22:02:22.513 | Found existing installation: PyYAML 3.11 | 22:03 |
nmagnezi | 2018-04-16 22:02:23.173 | Cannot uninstall 'PyYAML'. It is a distutils installed project and thus we cannot accurately determine which files belong to it which would lead to only a partial uninstall. | 22:03 |
nmagnezi | johnsom, another fix is on the way https://review.openstack.org/#/c/561479/ | 22:08 |
*** yamamoto has quit IRC | 22:08 | |
*** slaweq has quit IRC | 22:16 | |
*** slaweq has joined #openstack-lbaas | 22:16 | |
nmagnezi | johnsom, the log does not explain why it does not work | 22:18 |
nmagnezi | but maybe it's just worth waiting for https://review.openstack.org/#/c/561479/ ? | 22:18 |
johnsom | nmagnezi Ian claims his fix will solve this. I didn't understand that from looking at the patch, but that is cool | 22:19 |
* nmagnezi crossing fingers for Ian | 22:19 | |
johnsom | I'm not sure how he is doing it.... I'm running into nasty cross-package issues | 22:19 |
nmagnezi | maybe worth commenting this in the patch | 22:19 |
*** slaweq has quit IRC | 22:21 | |
johnsom | nmagnezi http://logs.openstack.org/40/561740/1/check/dib-dsvm-functests-python2-ubuntu-trusty/e35bcb7/job-output.txt.gz#_2018-04-16_22_17_33_592998 | 22:27 |
*** yamamoto has joined #openstack-lbaas | 23:04 | |
*** yamamoto has quit IRC | 23:09 | |
*** ianychoi has quit IRC | 23:18 | |
*** rm_mobile has joined #openstack-lbaas | 23:20 | |
rm_mobile | #txgaming | 23:20 |
rm_mobile | Errrrrr | 23:20 |
rm_mobile | That was not where I meant to type that | 23:20 |
rm_work | lol | 23:22 |
xgerman_ | it’s almost rm_work and rm_mobile are different people | 23:22 |
rm_work | almost :P | 23:22 |
*** rm_mobile has quit IRC | 23:22 | |
johnsom | We only worry if they start talking to each other | 23:23 |
rm_work | multi-sleeving is punishable by Real Death | 23:23 |
rm_work | I wouldn't dare | 23:23 |
rm_work | (referencing Altered Carbon, for those who haven't seen it yet) | 23:25 |
*** ianychoi has joined #openstack-lbaas | 23:25 | |
rm_work | wait.. wut | 23:29 |
rm_work | how did... wut | 23:29 |
*** tzumainn has left #openstack-lbaas | 23:30 | |
rm_work | ah nm | 23:31 |
rm_work | johnsom: err, this didn't set correctly: https://review.openstack.org/#/c/561049/7/zuul.d/jobs.yaml@72 | 23:35 |
rm_work | http://logs.openstack.org/49/561049/7/check/neutron-lbaasv2-dsvm-api-l7octavia/381bff3/logs/devstacklog.txt.gz#_2018-04-16_19_24_26_908 | 23:36 |
rm_work | did I do it wrong? | 23:36 |
johnsom | Looks fine there, let me check where you use it | 23:38 |
rm_work | https://review.openstack.org/#/c/561049/7/devstack/plugin.sh@159 | 23:46 |
rm_work | that was the log-line i linked | 23:46 |
rm_work | it was just blank | 23:46 |
rm_work | umm | 23:48 |
rm_work | it didn't get set in the localrc | 23:48 |
rm_work | http://logs.openstack.org/49/561049/7/check/neutron-lbaasv2-dsvm-api-l7octavia/381bff3/logs/local.conf.txt.gz | 23:48 |
rm_work | nothing about PROXY | 23:49 |
rm_work | so obviously I set it wrong | 23:49 |
*** ianychoi has quit IRC | 23:55 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!