| *** spotz_zzz is now known as spotz | 00:15 | |
| *** Ger-chervyak has joined #openstack-ansible | 00:22 | |
| *** Ger-chervyak has quit IRC | 00:25 | |
| *** abitha has quit IRC | 00:31 | |
| *** Ger-chervyak has joined #openstack-ansible | 00:32 | |
| *** portante has joined #openstack-ansible | 00:44 | |
| *** ajo has joined #openstack-ansible | 00:44 | |
| *** pabelanger has joined #openstack-ansible | 00:45 | |
| *** dmellado has joined #openstack-ansible | 00:49 | |
| *** Ger-chervyak has quit IRC | 00:51 | |
| *** iceyao has joined #openstack-ansible | 01:01 | |
| *** sdake has joined #openstack-ansible | 01:02 | |
| *** iceyao_ has joined #openstack-ansible | 01:05 | |
| *** fishcried has joined #openstack-ansible | 01:05 | |
| spotz | cloudnull: I think I've been working on that one, how'd you replace build-essentials? | 01:06 |
|---|---|---|
| *** Nepoc has quit IRC | 01:06 | |
| *** fishcried has quit IRC | 01:07 | |
| *** Nepoc has joined #openstack-ansible | 01:07 | |
| *** iceyao has quit IRC | 01:08 | |
| *** Brew has quit IRC | 01:19 | |
| *** Ger-chervyak has joined #openstack-ansible | 01:24 | |
| *** kiranv_ has joined #openstack-ansible | 01:27 | |
| *** mummer has quit IRC | 01:28 | |
| *** thorst_ has joined #openstack-ansible | 01:28 | |
| *** kiranv_ has quit IRC | 01:29 | |
| *** kiranv has quit IRC | 01:30 | |
| *** chandanc has quit IRC | 01:37 | |
| *** sacharya has joined #openstack-ansible | 01:39 | |
| *** sacharya has quit IRC | 01:44 | |
| *** BjoernT has joined #openstack-ansible | 01:49 | |
| *** Ger-chervyak has quit IRC | 01:50 | |
| *** saneax is now known as saneax_AFK | 01:51 | |
| *** klamath has quit IRC | 01:59 | |
| *** klamath has joined #openstack-ansible | 02:00 | |
| cloudnull | spotz: in RHEL that is '@Development Tools' | 02:02 |
| *** sacharya has joined #openstack-ansible | 02:02 | |
| *** klamath has quit IRC | 02:03 | |
| *** klamath has joined #openstack-ansible | 02:03 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-openstack_hosts: Updated role using the Multi-Distro framework https://review.openstack.org/295613 | 02:07 |
| *** fishcried has joined #openstack-ansible | 02:09 | |
| *** niknakpaddywak has joined #openstack-ansible | 02:10 | |
| spotz | cloudnull In the men time learning to hate galera | 02:10 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-openstack_hosts: Updated role using the Multi-Distro framework https://review.openstack.org/295613 | 02:11 |
| *** clickboom has joined #openstack-ansible | 02:13 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-openstack_hosts: Updated role using the Multi-Distro framework https://review.openstack.org/295613 | 02:15 |
| *** jorge_munoz has quit IRC | 02:16 | |
| *** jorge_munoz has joined #openstack-ansible | 02:19 | |
| errr | cloudnull: with config_template that would even add new sections to a config right, not just options to existing sections? | 02:22 |
| cloudnull | yes | 02:22 |
| errr | that rules. | 02:22 |
| cloudnull | you can add new sections | 02:22 |
| cloudnull | I think its pretty useful | 02:23 |
| errr | I hope they merge your p/r into core ansible soon with that | 02:23 |
| cloudnull | me too | 02:25 |
| cloudnull | https://github.com/ansible/ansible/pull/12555 <- thats the change | 02:25 |
| cloudnull | more reviews are always helpful :) | 02:25 |
| errr | yeah I read that a bit ago. Its like lost in libo or something | 02:25 |
| errr | limbo* | 02:25 |
| cloudnull | I've poked at few times | 02:26 |
| errr | do we not have core ansible people working for us? | 02:26 |
| cloudnull | no | 02:26 |
| errr | weak | 02:26 |
| cloudnull | yup | 02:26 |
| cloudnull | its a pita | 02:27 |
| cloudnull | i hope we can make it go. | 02:28 |
| errr | I bet it will happen | 02:28 |
| cloudnull | the ansible-ceph folks have adopted it | 02:28 |
| *** elo has quit IRC | 02:28 | |
| cloudnull | so it'd be helpful to lots of people if it was in core | 02:28 |
| errr | is Matt Martz able o help? | 02:29 |
| errr | to* | 02:29 |
| errr | he is still a racker I think and in the ansible org group on github at least | 02:29 |
| errr | https://github.com/sivel | 02:29 |
| *** ericlopez has quit IRC | 02:30 | |
| cloudnull | maybe? | 02:30 |
| errr | I bet Major knows him :) | 02:30 |
| *** clickboom has quit IRC | 02:30 | |
| errr | they used to public cloud together | 02:30 |
| errr | and slicehost | 02:30 |
| *** elo has joined #openstack-ansible | 02:31 | |
| cloudnull | I've tried not to go around process and be a good netizen so I've only worked though official channels. but it might be good to get some other Rackers involved to help thing along. | 02:31 |
| cloudnull | **help things along | 02:32 |
| errr | cant hurt, plus that is super helpful outside of osa for sure | 02:32 |
| cloudnull | +1 i agree | 02:32 |
| *** furlongm has quit IRC | 02:37 | |
| *** chandanc has joined #openstack-ansible | 02:39 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-openstack_hosts: Updated role using the Multi-Distro framework https://review.openstack.org/295613 | 02:40 |
| *** saneax_AFK is now known as saneax | 02:40 | |
| *** chandanc_ has joined #openstack-ansible | 02:41 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-openstack_hosts: Updated role using the Multi-Distro framework https://review.openstack.org/295613 | 02:45 |
| cloudnull | ^ i think that will finally make cent7 happy | 02:45 |
| *** ericlopez has joined #openstack-ansible | 02:46 | |
| *** kiranv_ has joined #openstack-ansible | 02:49 | |
| *** winggundamth has quit IRC | 02:50 | |
| *** BjoernT has quit IRC | 02:53 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-openstack_hosts: Updated role using the Multi-Distro framework https://review.openstack.org/295613 | 02:56 |
| *** mummer has joined #openstack-ansible | 02:58 | |
| *** klamath has quit IRC | 03:00 | |
| *** sdake has quit IRC | 03:07 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-openstack_hosts: Updated role using the Multi-Distro framework https://review.openstack.org/295613 | 03:08 |
| *** mummer has quit IRC | 03:32 | |
| *** thorst_ has quit IRC | 03:33 | |
| *** thorst_ has joined #openstack-ansible | 03:34 | |
| *** thorst_ has quit IRC | 03:38 | |
| *** thorst_ has joined #openstack-ansible | 03:42 | |
| *** kiranv_ has quit IRC | 04:02 | |
| *** chhavi has joined #openstack-ansible | 04:07 | |
| *** saneax is now known as saneax_AFK | 04:11 | |
| *** sdake has joined #openstack-ansible | 04:13 | |
| *** kiranv_ has joined #openstack-ansible | 04:16 | |
| *** iceyao_ has quit IRC | 04:35 | |
| *** iceyao has joined #openstack-ansible | 04:35 | |
| *** woodard has quit IRC | 04:38 | |
| *** kiranv_ has quit IRC | 04:44 | |
| *** saneax_AFK is now known as saneax | 04:47 | |
| *** javeriak has joined #openstack-ansible | 05:06 | |
| *** sdake has quit IRC | 05:06 | |
| *** winggundamth has joined #openstack-ansible | 05:07 | |
| *** kiranv_ has joined #openstack-ansible | 05:20 | |
| *** shausy has joined #openstack-ansible | 05:24 | |
| *** iceyao_ has joined #openstack-ansible | 05:28 | |
| *** chhavi has quit IRC | 05:28 | |
| *** iceyao has quit IRC | 05:31 | |
| *** saneax is now known as saneax_AFK | 05:37 | |
| *** woodard has joined #openstack-ansible | 05:39 | |
| *** chhavi has joined #openstack-ansible | 05:41 | |
| *** woodard has quit IRC | 05:44 | |
| *** javeriak has quit IRC | 05:46 | |
| openstackgerrit | Merged openstack/openstack-ansible: Add neutron_openvswitch_agent to env.d files https://review.openstack.org/300605 | 05:52 |
| *** thorst_ has quit IRC | 06:06 | |
| *** thorst_ has joined #openstack-ansible | 06:06 | |
| *** thorst_ has quit IRC | 06:11 | |
| *** Oku_OS-away is now known as Oku_OS | 06:13 | |
| *** kiranv_ has quit IRC | 06:18 | |
| *** javeriak has joined #openstack-ansible | 06:31 | |
| *** neilus has joined #openstack-ansible | 06:40 | |
| *** chhavi has quit IRC | 06:43 | |
| *** chhavi has joined #openstack-ansible | 06:44 | |
| *** sacharya has quit IRC | 06:53 | |
| *** iceyao_ has quit IRC | 07:07 | |
| *** iceyao has joined #openstack-ansible | 07:07 | |
| *** jamielennox is now known as jamielennox|away | 07:08 | |
| *** fishcried has quit IRC | 07:09 | |
| *** Bogdan has quit IRC | 07:18 | |
| *** karimb has joined #openstack-ansible | 07:19 | |
| *** mikelk has joined #openstack-ansible | 07:34 | |
| *** Adri2000 has joined #openstack-ansible | 07:41 | |
| Adri2000 | hi | 07:41 |
| Adri2000 | I've got an issue when trying inventory/dynamic_inventory.py --config /etc/openstack_deploy/ after making the initial configuration: the scripts errors out with "RuntimeError: maximum recursion depth exceeded while calling a Python object" | 07:42 |
| Adri2000 | turns out the "culprit" line in openstack_user_config.yml is "- all_containers" in provider_networks: - network: group_binds: | 07:43 |
| Adri2000 | this is the mgmt network, exactly as mentioned at http://docs.openstack.org/developer/openstack-ansible/install-guide/configure-networking.html | 07:44 |
| Adri2000 | any idea? | 07:44 |
| *** javeriak has quit IRC | 07:47 | |
| *** javeriak has joined #openstack-ansible | 07:47 | |
| *** jamielennox|away is now known as jamielennox | 07:49 | |
| mattt | Adri2000: can you gist your config ? | 07:50 |
| *** admin0 has joined #openstack-ansible | 07:50 | |
| *** sacharya has joined #openstack-ansible | 07:54 | |
| Adri2000 | mattt: http://pastebin.com/dCKxNWFZ this is the output of grep -v ^# | 07:57 |
| mattt | Adri2000: k, having a poke now | 07:58 |
| *** sacharya has quit IRC | 07:58 | |
| mattt | Adri2000: indentation under used_ips is wrong, for starters | 07:59 |
| mattt | Adri2000: and under global_overrides -> provider_networks -> group_binds | 07:59 |
| admin0 | good morning all | 08:03 |
| mattt | Adri2000: do you see the issue? | 08:03 |
| mattt | Adri2000: make sure your spacing is consistent when you're dealing w/ the list items | 08:04 |
| mattt | morning adm0 | 08:04 |
| Adri2000 | mattt: I see that on pastebin indeed, but if you look at the raw text (bottom of the page) it seems correct, and it does as well in my vim | 08:06 |
| Adri2000 | http://pastebin.com/raw/dCKxNWFZ here even | 08:06 |
| openstackgerrit | Matt Thompson proposed openstack/openstack-ansible-os_tempest: Standardise tempest role https://review.openstack.org/310766 | 08:08 |
| mattt | Adri2000: oh, derp | 08:08 |
| winggundamth | hi mattt, cloudnull. I'm back from Austin now :) | 08:10 |
| *** javeriak has quit IRC | 08:12 | |
| openstackgerrit | Darren Birkett proposed openstack/openstack-ansible: install rabbitmq-server in serial https://review.openstack.org/314457 | 08:19 |
| mattt | winggundamth: welcome back -- how was your trip ? | 08:21 |
| mattt | Adri2000: i remember seeing this error if you had a host w/ the same name as a host group, but that doesn't seem to be the case here (and pretty sure that was fixed also) | 08:22 |
| winggundamth | mattt: it's great trip I ever had. with airbnb host. he is so nice. | 08:22 |
| *** iceyao has quit IRC | 08:23 | |
| mattt | Adri2000: can you gist your python error also? | 08:26 |
| *** iceyao has joined #openstack-ansible | 08:28 | |
| Adri2000 | mattt: http://pastebin.com/raw/hBtpS8zW | 08:29 |
| *** tiagogomes has joined #openstack-ansible | 08:41 | |
| mattt | Adri2000: i'm really not sure :( i think i'm missing something obvious | 08:46 |
| odyssey4me | Adri2000 typically that error comes up when you're using a host name that's the same as one of the group names | 08:48 |
| odyssey4me | o/ all | 08:48 |
| mattt | odyssey4me: http://pastebin.com/raw/dCKxNWFZ -- controller is called 'controller' | 08:51 |
| mattt | i don't know if that is what is causing the issue as i see no reference to 'controller' in my own inventory | 08:51 |
| Adri2000 | ahhhh, compute and network | 08:54 |
| Adri2000 | changed them to compute1 and network1 | 08:54 |
| mattt | Adri2000: great! now i could have sworn this issue was addressed, let me look through commits to see | 08:55 |
| Adri2000 | that was it | 08:55 |
| Adri2000 | no problem with "controller" though | 08:55 |
| mattt | excellent! i overlooked network and compute, need to go back to sleep :P | 08:56 |
| *** sacharya has joined #openstack-ansible | 09:00 | |
| Adri2000 | thanks anyway because I wouldn't have found this myself :) | 09:00 |
| *** sdake has joined #openstack-ansible | 09:01 | |
| *** sacharya has quit IRC | 09:04 | |
| *** saneax_AFK is now known as saneax | 09:20 | |
| admin0 | if i have 4 swift proxy servers and 2 load balancers, can I round-robin between 2 load balancers ? or is swift stateful ? | 09:34 |
| openstackgerrit | Merged openstack/openstack-ansible: Isolate Ansible bootstrap from repo servers https://review.openstack.org/314294 | 09:35 |
| *** iceyao_ has joined #openstack-ansible | 09:51 | |
| *** iceyao has quit IRC | 09:51 | |
| odyssey4me | admin0 I think https://github.com/openstack/openstack-ansible/blob/master/playbooks/vars/configs/haproxy_config.yml#L175 answers your question? | 09:56 |
| admin0 | odyssey4me: thinking of having 2 haproxies | 10:01 |
| admin0 | on different routers, differnet network .. but Ips can still falover each other | 10:02 |
| odyssey4me | admin0 for the moment our implementation uses keepalived and an active/passive model | 10:02 |
| odyssey4me | you can choose to do something different, but it's then up to you to carry the testing for it | 10:03 |
| *** javeriak has joined #openstack-ansible | 10:06 | |
| *** iceyao_ has quit IRC | 10:09 | |
| *** smatzek has joined #openstack-ansible | 10:25 | |
| *** johnmilton has quit IRC | 10:29 | |
| *** smatzek has quit IRC | 10:30 | |
| *** javeriak has quit IRC | 10:31 | |
| *** smatzek has joined #openstack-ansible | 10:33 | |
| *** javeriak has joined #openstack-ansible | 10:35 | |
| *** javeriak has quit IRC | 10:35 | |
| *** javeriak has joined #openstack-ansible | 10:36 | |
| *** spotz is now known as spotz_zzz | 10:39 | |
| *** sdake has quit IRC | 10:42 | |
| *** sacharya has joined #openstack-ansible | 11:01 | |
| *** sacharya has quit IRC | 11:05 | |
| *** chandanc has quit IRC | 11:06 | |
| *** sdake has joined #openstack-ansible | 11:07 | |
| *** chandanc_ has quit IRC | 11:07 | |
| *** javeriak has quit IRC | 11:20 | |
| *** johnmilton has joined #openstack-ansible | 11:24 | |
| *** deadnull_ has joined #openstack-ansible | 11:26 | |
| *** javeriak has joined #openstack-ansible | 11:30 | |
| *** javeriak has quit IRC | 11:34 | |
| *** javeriak has joined #openstack-ansible | 11:34 | |
| *** retreved has joined #openstack-ansible | 11:39 | |
| *** javeriak has quit IRC | 11:40 | |
| *** javeriak_ has joined #openstack-ansible | 11:40 | |
| *** woodard has joined #openstack-ansible | 11:44 | |
| *** weshay has joined #openstack-ansible | 11:48 | |
| *** woodard_ has joined #openstack-ansible | 11:50 | |
| *** woodard has quit IRC | 11:50 | |
| *** sdake has quit IRC | 11:56 | |
| *** saneax is now known as saneax_AFK | 11:56 | |
| *** sdake has joined #openstack-ansible | 12:02 | |
| *** psilvad has joined #openstack-ansible | 12:03 | |
| *** saneax_AFK is now known as saneax | 12:05 | |
| *** ozialien10 has quit IRC | 12:08 | |
| *** sdake_ has joined #openstack-ansible | 12:08 | |
| *** ozialien10 has joined #openstack-ansible | 12:09 | |
| *** sdake has quit IRC | 12:11 | |
| *** psilvad has quit IRC | 12:11 | |
| *** javeriak_ has quit IRC | 12:18 | |
| *** psilvad has joined #openstack-ansible | 12:21 | |
| *** tiagogomes has quit IRC | 12:29 | |
| mhayden | morning folks | 12:34 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-security: Add CentOS 7 and Ubuntu 16.04 support https://review.openstack.org/313697 | 12:35 |
| *** chandanc has joined #openstack-ansible | 12:36 | |
| *** chandanc_ has joined #openstack-ansible | 12:36 | |
| *** kstev has joined #openstack-ansible | 12:37 | |
| *** saneax is now known as saneax_AFK | 12:44 | |
| *** tiagogomes has joined #openstack-ansible | 12:44 | |
| *** thorst_ has joined #openstack-ansible | 12:45 | |
| *** spotz_zzz is now known as spotz | 12:45 | |
| *** thorst_ has quit IRC | 12:49 | |
| *** thorst_ has joined #openstack-ansible | 12:49 | |
| *** saneax_AFK is now known as saneax | 12:51 | |
| *** Nepoc has quit IRC | 12:53 | |
| *** Nepoc has joined #openstack-ansible | 12:54 | |
| cloudnull | morning | 12:55 |
| cloudnull | o/ winggundamth great seeing at the summit ! :) | 12:55 |
| *** spotz is now known as spotz_zzz | 12:56 | |
| admin0 | morning mhayden | 12:57 |
| admin0 | morning cloudnull | 12:57 |
| cloudnull | o/ admin0 | 12:57 |
| admin0 | couldn’t come to the summit, so took a vacation to barcelona ( where the next summit will be ) | 12:58 |
| admin0 | :D | 12:58 |
| *** sacharya has joined #openstack-ansible | 13:02 | |
| *** keedya has joined #openstack-ansible | 13:02 | |
| *** asettle has joined #openstack-ansible | 13:03 | |
| cloudnull | that is a good vacation spot for sure! | 13:03 |
| admin0 | i liked it | 13:05 |
| admin0 | driving felt like same as driving in Nepal ( twists, turns, ups and downs ) | 13:05 |
| admin0 | netherlands is kind of straight and flat | 13:05 |
| *** sacharya has quit IRC | 13:06 | |
| *** Andrew_jedi has joined #openstack-ansible | 13:06 | |
| *** neilus1 has joined #openstack-ansible | 13:07 | |
| Andrew_jedi | hello foljs, Is it possible to integrate the Ceph Object Gateway with Keystone via OSA ? something like this, http://docs.ceph.com/docs/hammer/radosgw/keystone/ | 13:08 |
| Andrew_jedi | s/foljs/folks | 13:08 |
| *** neilus has quit IRC | 13:11 | |
| cloudnull | can we get a few folks to weigh in on https://review.openstack.org/#/c/313890/ -- the pip_lock_down and pip_install roles have become one role and jmccrory has a few patches to clean up that dependecy tree and it'd be great to work out the pattern we want to use so we can get those patches merged. | 13:13 |
| cloudnull | on https://review.openstack.org/#/c/313890/ there's a convo regarding pattern which should determine the direction of the other patche.s | 13:13 |
| cloudnull | Andrew_jedi: IDK. mattt RE: ceph + keystone gw ? | 13:14 |
| cloudnull | Andrew_jedi: I would think it work . | 13:15 |
| cloudnull | the [client.radosgw.gateway] section is in your ceph conf | 13:15 |
| mattt | cloudnull: i don't see why not ? | 13:15 |
| Andrew_jedi | cloudnull: But how to configure it | 13:15 |
| cloudnull | then all you need to do is create the endpoint | 13:15 |
| mattt | nothing has to go into osa itself for this | 13:15 |
| mattt | it's all client configurations | 13:15 |
| *** asettle has quit IRC | 13:15 | |
| Andrew_jedi | matt: Ok so all this has to be done manually | 13:16 |
| mattt | our deploys still support keystone v2 right, because rgw on hammer doesn't support v3 | 13:16 |
| mattt | sorry, that was a question :) | 13:16 |
| cloudnull | yes they do | 13:16 |
| Andrew_jedi | cloudnull: I am on kilo btw | 13:16 |
| cloudnull | I assume you've not already deployed swift ? | 13:17 |
| Andrew_jedi | cloudnull: Nope, no swift | 13:17 |
| mattt | Andrew_jedi: you the drizzle guy ? | 13:17 |
| *** tiagogomes has quit IRC | 13:17 | |
| *** jamesden_ has joined #openstack-ansible | 13:17 | |
| Andrew_jedi | matt: Sorry i am not aware of the reference here | 13:18 |
| Andrew_jedi | :) | 13:18 |
| mattt | Andrew_jedi: no worries, not you then :) | 13:18 |
| mattt | Andrew_jedi: yeah, just follow that doc, from what i can tell nothing needs to be implemented from an OSA perspective | 13:18 |
| mattt | but not tried it myself tbh | 13:18 |
| cloudnull | mattt so to make it go all Andrew_jedi would need to do is update his client config? | 13:19 |
| cloudnull | seems like you can laydown the needed config with the ceph_conf var | 13:19 |
| cloudnull | https://github.com/openstack/openstack-ansible/blob/master/playbooks/roles/ceph_client/templates/ceph.conf.j2 | 13:19 |
| *** thorst_ has quit IRC | 13:20 | |
| cloudnull | Andrew_jedi: once you have the config in place you run the one keystone/openstack command and youre off to the races | 13:20 |
| *** chandanc has quit IRC | 13:22 | |
| Andrew_jedi | cloudnull: Which playbooks do i have to run after making changes to ceph.conf? | 13:22 |
| mattt | cloudnull: i think you would have to manually put that ceph.conf on the keystone containers, iirc ceph_client is tied into glance,cinder,nova | 13:22 |
| *** chandanc_ has quit IRC | 13:22 | |
| cloudnull | Andrew_jedi: i dont think you have to run any playbooks | 13:23 |
| cloudnull | once your ceph.conf is in place. | 13:23 |
| Andrew_jedi | matt cloudnull : Okay let me try this and i will come back to you with the results. thanks guys :) | 13:24 |
| cloudnull | just use the admin user creds and execute the couple of openstack commands to make the new endpoint | 13:24 |
| cloudnull | mattt: does that ceph conf need to be in the keystone containers? | 13:24 |
| * cloudnull is ceph ignorant | 13:25 | |
| mattt | cloudnull: actually i don't think so tbh | 13:26 |
| mattt | cloudnull: but it's been many years since we played with rgw | 13:26 |
| cloudnull | yea. good old Joe of the city :) | 13:27 |
| *** Andrew_jedi has quit IRC | 13:29 | |
| *** git-harry has quit IRC | 13:29 | |
| *** git-harry has joined #openstack-ansible | 13:30 | |
| mhayden | jmccrory: i like what you've done in https://review.openstack.org/#/c/286282/ -- i wonder if we should get some developer docs for the var_files_var pattern | 13:30 |
| mhayden | so that subsequent roles could use it | 13:30 |
| *** tiagogomes has joined #openstack-ansible | 13:30 | |
| *** thorst_ has joined #openstack-ansible | 13:32 | |
| *** Ger-chervyak has joined #openstack-ansible | 13:34 | |
| openstackgerrit | Samuel Matzek proposed openstack/openstack-ansible: Refactor ceph_client for multi-OS and ceph https://review.openstack.org/312692 | 13:36 |
| cloudnull | mattt: RE: https://review.openstack.org/#/c/313278/ is there an issue with the ceph libs on pypi that hughsaunders is maintaining ? | 13:37 |
| *** klamath has joined #openstack-ansible | 13:37 | |
| openstackgerrit | Merged openstack/openstack-ansible-rabbitmq_server: install rabbitmq-server in serial https://review.openstack.org/310542 | 13:37 |
| mhayden | andymccr: that single curly brace stuff in the swift rsync patch made me do a double take :P | 13:39 |
| *** thorst__ has joined #openstack-ansible | 13:40 | |
| *** thorst__ has quit IRC | 13:40 | |
| *** thorst_ has quit IRC | 13:41 | |
| smatzek | cloudnull: the cephlibs on pypi pins OSA to only support ceph hammer from ceph.com. It was discussed here: http://eavesdrop.openstack.org/irclogs/%23openstack-ansible/%23openstack-ansible.2016-05-05.log.html#t2016-05-05T12:47:56 | 13:43 |
| smatzek | cloudnull: This becomes an issue if you want to use the latest Ceph LTS, Ceph Jewel which comes by default in Xenial. | 13:44 |
| cloudnull | ah. | 13:44 |
| cloudnull | this makes sense | 13:44 |
| cloudnull | thanks smatzek | 13:44 |
| smatzek | cloudnull: no problem | 13:45 |
| mattt | smatzek: yeah still not sure we arrived at a decision here | 13:45 |
| *** maestro1 has joined #openstack-ansible | 13:45 | |
| *** thorst_ has joined #openstack-ansible | 13:46 | |
| smatzek | mattt: maybe not final, but I thought we were leaning toward going back to the linking and maintaining a dict like this: http://paste.openstack.org/show/496362/ The lesser evil between trying to maintain separate pypis for the different Ceph releases and sources (ceph.com, Canoncial distro, UCA, and in the future RDO?) | 13:47 |
| *** kiranv has joined #openstack-ansible | 13:49 | |
| mattt | smatzek: yeah, but it starts to get pretty ugly as we have people deploying trusty, xenial, ceph repos, centos, etc. | 13:49 |
| *** kiranv has quit IRC | 13:49 | |
| *** thorst_ has quit IRC | 13:50 | |
| *** thorst_ has joined #openstack-ansible | 13:50 | |
| *** thorst_ has quit IRC | 13:51 | |
| *** thorst_ has joined #openstack-ansible | 13:52 | |
| *** Mudpuppy has joined #openstack-ansible | 13:53 | |
| *** Min_Cai has joined #openstack-ansible | 13:53 | |
| Min_Cai | hi, I'd like to run tempest and add a new testcases for openstack-ansible, how can I do? Thanks | 13:54 |
| *** thorst_ has quit IRC | 13:56 | |
| *** thorst_ has joined #openstack-ansible | 13:58 | |
| *** ametts has joined #openstack-ansible | 13:58 | |
| *** michaelgugino has joined #openstack-ansible | 14:00 | |
| *** jthorne has joined #openstack-ansible | 14:02 | |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_hosts: Implement Ubuntu support on ppc64le https://review.openstack.org/313807 | 14:02 |
| *** sigmavirus24_awa is now known as sigmavirus24 | 14:04 | |
| *** jayc has joined #openstack-ansible | 14:04 | |
| *** jthorne has quit IRC | 14:04 | |
| *** jthorne has joined #openstack-ansible | 14:05 | |
| *** iceyao has joined #openstack-ansible | 14:06 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_swift: [WIP] Covert swift python to pypy https://review.openstack.org/314587 | 14:09 |
| cloudnull | andymccr: ^ idk if i'm insane but its working for me and is seems to be crazy fast | 14:11 |
| cloudnull | Min_Cai: install the tempest role | 14:12 |
| cloudnull | once installed | 14:12 |
| cloudnull | tempest will be available from within the utility container | 14:12 |
| Min_Cai | yes | 14:13 |
| Min_Cai | I can see the /opt/openstack_tempest_gate.sh in utility container | 14:13 |
| Min_Cai | and a folder: tempest_ac94d0640cbcdbccf90d1da343dad0f01539779f/ | 14:13 |
| cloudnull | if you want to add additional test cases that is the script to modify | 14:13 |
| Min_Cai | you mean change /opt/openstack_tempest_gate.sh? | 14:15 |
| cloudnull | yes. | 14:17 |
| cloudnull | https://github.com/openstack/openstack-ansible-os_tempest/blob/master/templates/openstack_tempest_gate.sh.j2 | 14:17 |
| cloudnull | thats the script that runs our scenarios | 14:17 |
| *** clickboom has joined #openstack-ansible | 14:17 | |
| *** neilus1 has quit IRC | 14:18 | |
| mattt | Min_Cai: what do you mean add new test cases for openstack-ansible ? | 14:19 |
| odyssey4me | cloudnull mattt automagically jmccrory with regards to the discussion in https://review.openstack.org/#/c/313890/1/meta/main.yml - cloudnull which approach is it that you like - it's not clear? is it basing the inclusion on developer_mode as jmccrory has proposed, or is it basing it on pip_links as automagically has proposed in https://review.openstack.org/#/c/313681/1/meta/main.yml ? | 14:19 |
| cloudnull | I like the approach that jmccrory has created | 14:20 |
| cloudnull | we don't lockdown on developer mode currently so it makes sense to keep that API the same | 14:21 |
| Min_Cai | thanks. yes, matt, we'd like to add our compute-driver to ansible support, and also need new cases for it. Just investigate how to run and how to add cases. | 14:21 |
| odyssey4me | Min_Cai is this for the ppc compute driver support? | 14:22 |
| *** phalmos has joined #openstack-ansible | 14:22 | |
| Min_Cai | yes | 14:22 |
| Min_Cai | power platform | 14:22 |
| odyssey4me | Min_Cai the convention we're using is generally to have the tempest test cases run - so new test cases should be implemented in tempest or tempest plugins... mattt correct me if I'm wrong there | 14:23 |
| cloudnull | odyssey4me: so yes, I'd like to see us follow this pattern -- https://review.openstack.org/#/c/313890 for all of the other roles | 14:23 |
| odyssey4me | cloudnull I guess that developer_mode is an in-role var so it makes better sense than doing it using pip_links which is not in the role | 14:24 |
| Min_Cai | we're now working on this, still picking up ansible skills so we walk slowly but keep moving. | 14:24 |
| cloudnull | odyssey4me: +1 | 14:24 |
| Min_Cai | so we should add new things on openstack-ansible-os_tempest project rather than openstack-ansible project right? | 14:25 |
| Min_Cai | openstack-ansible -->calls--> openstack-ansible-os_tempest | 14:26 |
| odyssey4me | cloudnull I find myself wondering why we need the var pip_lock_internal_repo at all... why not just include the pip_install role and then in the role key off whether there are any pip links configured? | 14:26 |
| odyssey4me | r perhaps each playbook can feed the value for pip_lock_internal_repo | 14:27 |
| odyssey4me | *or | 14:27 |
| mattt | my concern there is that we're adding weird logic to the role that should be elsewhere | 14:27 |
| cloudnull | maybe both should be present as a constraint | 14:27 |
| mattt | like the role shouldn't care about having pip being locked down or not, that's irrelevant | 14:27 |
| cloudnull | IE lockdown if there are links set and not in developer mode | 14:28 |
| mattt | and i'm just unsure about starting to hide variables in meta files, eventually it's going to be hard to track things down when we start this | 14:28 |
| odyssey4me | cloudnull my point is that the lockdown is false by default, so the role meta doesn't need to have a conditional | 14:28 |
| odyssey4me | not conditional - the var | 14:28 |
| cloudnull | thats a fair point too | 14:29 |
| odyssey4me | the var should perhaps be handled in the role, or from the play | 14:29 |
| cloudnull | maybe all that logic goes into the pip_install role regardless | 14:29 |
| *** gregfaust has joined #openstack-ansible | 14:30 | |
| *** Ger-chervyak has quit IRC | 14:31 | |
| *** Ger-chervyak has joined #openstack-ansible | 14:31 | |
| *** sigmavirus24 is now known as sigmavirus24_awa | 14:32 | |
| *** sigmavirus24_awa is now known as sigmavirus24 | 14:33 | |
| *** saneax is now known as saneax_AFK | 14:33 | |
| *** saneax_AFK is now known as saneax | 14:33 | |
| *** keedya has quit IRC | 14:35 | |
| *** sigmavirus24 is now known as sigmavirus24_awa | 14:35 | |
| odyssey4me | cloudnull I would go as far as to say that pip_install doesn't even have to be a dependency for the service role - the os_magnum role should not have to care about how pip got there, just like it shouldn't have to care how the database sb/user/password and rabbit vhost/user/password got there | 14:36 |
| odyssey4me | s/sb/db/ | 14:36 |
| *** galstrom_zzz is now known as galstrom | 14:36 | |
| *** keedya has joined #openstack-ansible | 14:36 | |
| cloudnull | i disagree. it cant function without pip so it has to be a dep. | 14:36 |
| odyssey4me | it can't function without a db, and can't function without a rabbit vhost either | 14:37 |
| cloudnull | its an application installation not a connection string | 14:37 |
| cloudnull | right but those are conenction strings | 14:37 |
| cloudnull | not a binary that must be present | 14:37 |
| *** sigmavirus24_awa is now known as sigmavirus24 | 14:37 | |
| *** TxGirlGeek has joined #openstack-ansible | 14:37 | |
| cloudnull | also we need to go through the os_* roles and remove the database sb/user/password bits from the roles and put them into the plays | 14:38 |
| cloudnull | it seems that work got partially lost in the last cycle. | 14:38 |
| *** thorst_ has quit IRC | 14:39 | |
| cloudnull | IE: https://github.com/openstack/openstack-ansible-os_nova/blob/master/tasks/nova_db_setup.yml | 14:39 |
| odyssey4me | oh, we should clean that up | 14:39 |
| *** thorst_ has joined #openstack-ansible | 14:39 | |
| cloudnull | yea. | 14:40 |
| cloudnull | :'( | 14:40 |
| odyssey4me | with regards to the meta deps - I'm not too fussed either way on whether it should be a dep or not... functionally it results in the same thing so it's not really something I care too much about either way at this point | 14:40 |
| cloudnull | so anyway, i'm +1 on removing the var to lock down or not, im -1 on removing the dep. | 14:40 |
| cloudnull | odyssey4me: what you said :) | 14:41 |
| odyssey4me | however I really don't think we need to add a ternary into the dep because it will have the same functional result without it if we just include the role dep and have the pip_lock_to_internal_repo activated using the play or some other mechanism | 14:41 |
| *** thorst_ has quit IRC | 14:44 | |
| odyssey4me | would it make sense to have https://github.com/openstack/openstack-ansible-pip_install/blob/master/tasks/main.yml#L20-L22 changed to react to the number of pip_links defined? | 14:49 |
| odyssey4me | ie instead of a new var, we react based on an existing var | 14:49 |
| *** weezS has joined #openstack-ansible | 14:50 | |
| gregfaust | howdy guys... I'm running into an issue with latest OSA and trying to figure out if I need to open a bug or if it's my fault. Basically, when I attempt to create a cinder volume (backed by ceph) from an image, it fails and logs inside the cinder-volumens container "qemu-img executable not found" (and a bunch of python stack dumps). searching through the code I can't find anywhere that the qemu-utils package | 14:50 |
| gregfaust | would have been installed. Any thougts? If I manually install qemu-utils inside the cinder-volumes container my issue goes away and it works fine. | 14:50 |
| *** pcaruana has joined #openstack-ansible | 14:53 | |
| *** jorge_munoz_ has joined #openstack-ansible | 14:53 | |
| odyssey4me | gregfaust what branch/tag are you running, because we do already install qemu-utils on the cinder-volume host/container: https://github.com/openstack/openstack-ansible-os_cinder/blob/master/vars/ubuntu-14.04.yml#L33 | 14:54 |
| *** thorst_ has joined #openstack-ansible | 14:54 | |
| odyssey4me | oh hang on - you're not using lvm and it seems that's only installed on a host which is lvm-backed cinder | 14:54 |
| *** jorge_munoz has quit IRC | 14:54 | |
| *** jorge_munoz_ is now known as jorge_munoz | 14:54 | |
| odyssey4me | https://github.com/openstack/openstack-ansible-os_cinder/blob/master/tasks/cinder_install_apt.yml#L39-L50 | 14:55 |
| gregfaust | thanks guys, I'm running 13.0.1, but I'm not using cinder lvm backend, I'm using ceph backend | 14:56 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-galera_server: [WIP] Implement Xenial Support https://review.openstack.org/312602 | 14:56 |
| odyssey4me | gregfaust that's a bug - would you mind registering a bug for that and providing the details? | 14:58 |
| *** thorst_ has quit IRC | 14:58 | |
| openstackgerrit | Merged openstack/openstack-ansible: install rabbitmq-server in serial https://review.openstack.org/314457 | 15:02 |
| *** thorst_ has joined #openstack-ansible | 15:03 | |
| *** saneax is now known as saneax_AFK | 15:05 | |
| *** TxGirlGeek has quit IRC | 15:05 | |
| *** Brew has joined #openstack-ansible | 15:05 | |
| gregfaust | odyssey4me, will do | 15:06 |
| *** Brew has quit IRC | 15:06 | |
| *** Andrew_jedi has joined #openstack-ansible | 15:06 | |
| *** Ger-chervyak has quit IRC | 15:07 | |
| *** Brew has joined #openstack-ansible | 15:07 | |
| gregfaust | https://bugs.launchpad.net/openstack-ansible/+bug/1580222 | 15:09 |
| openstack | Launchpad bug 1580222 in openstack-ansible "volume creation based on image source fails when using ceph backend" [Undecided,New] | 15:09 |
| *** sacharya has joined #openstack-ansible | 15:09 | |
| mattt | gregfaust: i think in our testing we've always deployed an LVM backend along side ceph, so never caught this | 15:11 |
| *** mageshgv has joined #openstack-ansible | 15:13 | |
| *** chandanc_ has joined #openstack-ansible | 15:14 | |
| *** chandanc has joined #openstack-ansible | 15:14 | |
| gregfaust | yep, lvm is working fine for me, and I could see the docs were steering me towards multi-backend which appears would work fine, but I didn't test that. In my prod environment the cinder backends are Ceph and SolidFire so I need to figure out how to add SolidFire too. automagically is in a class this week so maybe I can discuss with him next week. | 15:20 |
| mattt | gregfaust: here's the regression : https://review.openstack.org/#/c/293836/ | 15:20 |
| mattt | you can thank him as he approved that one :) | 15:21 |
| gregfaust | lol | 15:21 |
| gregfaust | :) | 15:21 |
| mattt | hehe | 15:21 |
| mattt | i think openstack-ansible needs to make ceph a first class citizen | 15:21 |
| mattt | so we can do better gating on it | 15:22 |
| mattt | seems like more and more people are using it :( | 15:22 |
| *** shausy has quit IRC | 15:22 | |
| jmccrory | morning | 15:22 |
| gregfaust | matt, why the sad face, ceph is awesome | 15:23 |
| mattt | gregfaust: don't disagree, but openstack-ansible doesn't gate using it so some code paths aren't exercised in the gate (like this one, otherwise we'd have caught it) | 15:23 |
| *** keedya has quit IRC | 15:23 | |
| jmccrory | odyssey4me: keying off of pip_links instead of the new var in pip_install makes sense. pip_lock_down could just removed/replaced with pip_install everywhere else without any other changes or logic | 15:23 |
| odyssey4me | jmccrory I thought a little more about using pip_links... and don't think it's a good idea | 15:24 |
| jmccrory | hmm what issues would come up? | 15:24 |
| odyssey4me | unless we change how the repo_build works we would end up with the repo server locked down to itself | 15:24 |
| odyssey4me | so it's probably best to specifically make the lock down happen through the lock_down var in the playbooks | 15:25 |
| odyssey4me | not in group_vars | 15:25 |
| jmccrory | https://review.openstack.org/#/c/310475 ? | 15:25 |
| odyssey4me | jmccrory yeah, so that could help - but then we break how we build the repo in OpenStack-CI too because that relies on a global pip conf... the user pip conf would override the global one and we'd end up moving back to the stone age | 15:27 |
| jmccrory | ah | 15:27 |
| odyssey4me | we would have to figure out a way to slurp in the host's /etc/pip.conf and to feed the pypi index and wheel mirror info through to the repo build process | 15:28 |
| *** iceyao has quit IRC | 15:28 | |
| *** keedya has joined #openstack-ansible | 15:28 | |
| *** sacharya_ has joined #openstack-ansible | 15:29 | |
| *** Ger-chervyak has joined #openstack-ansible | 15:31 | |
| *** sacharya has quit IRC | 15:31 | |
| *** spotz_zzz is now known as spotz | 15:31 | |
| *** admin0 has quit IRC | 15:31 | |
| *** TxGirlGeek has joined #openstack-ansible | 15:34 | |
| *** Ger-chervyak has quit IRC | 15:35 | |
| *** flwang has quit IRC | 15:39 | |
| *** deadnull_ is now known as _deadnull | 15:40 | |
| mattt | wut, just got a +2 from mhayden | 15:43 |
| * mattt runs away screaming | 15:43 | |
| *** dweaver has quit IRC | 15:44 | |
| *** javeriak has joined #openstack-ansible | 15:45 | |
| *** mummer has joined #openstack-ansible | 15:45 | |
| *** TxGirlGeek has quit IRC | 15:45 | |
| *** mikelk has quit IRC | 15:45 | |
| *** TxGirlGeek has joined #openstack-ansible | 15:45 | |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible-rabbitmq_server: Multi-distro framework for rabbitmq_server role https://review.openstack.org/286282 | 15:47 |
| *** sdake_ has quit IRC | 15:47 | |
| *** javeriak_ has joined #openstack-ansible | 15:48 | |
| *** Oku_OS is now known as Oku_OS-away | 15:49 | |
| mhayden | prometheanfire: https://review.openstack.org/#/c/313697/ | 15:49 |
| *** javeriak has quit IRC | 15:49 | |
| *** Ger-chervyak has joined #openstack-ansible | 15:50 | |
| *** karimb_ has joined #openstack-ansible | 15:50 | |
| *** sdake has joined #openstack-ansible | 15:50 | |
| *** maestro1 has quit IRC | 15:50 | |
| * mhayden patpats mattt | 15:51 | |
| *** karimb has quit IRC | 15:52 | |
| *** sdake_ has joined #openstack-ansible | 15:54 | |
| *** stelouie has joined #openstack-ansible | 15:57 | |
| *** sdake has quit IRC | 15:57 | |
| *** flwang has joined #openstack-ansible | 15:58 | |
| *** jmccrory_ has joined #openstack-ansible | 15:58 | |
| *** thorst_ has quit IRC | 15:59 | |
| *** stelouie has quit IRC | 16:01 | |
| *** thorst_ has joined #openstack-ansible | 16:01 | |
| *** Brew has quit IRC | 16:01 | |
| *** Brew1 has joined #openstack-ansible | 16:01 | |
| *** Brew1 is now known as Brew | 16:01 | |
| spotz | real work swamped today ping if needed | 16:01 |
| *** TxGirlGeek has quit IRC | 16:02 | |
| *** keedya has quit IRC | 16:04 | |
| *** joker_ has quit IRC | 16:05 | |
| odyssey4me | whoops, anyone up for bug triage? | 16:06 |
| odyssey4me | bug triage here cloudnull, mattt, andymccr, d34dh0r53, hughsaunders, b3rnard0, palendae, Sam-I-Am, odyssey4me, serverascode, rromans, erikmwilson, mancdaz, _shaps_, BjoernT, claco, echiu, dstanek, jwagner, ayoung, prometheanfire, evrardjp, arbrandes, mhayden, scarlisle, luckyinva, ntt, javeriak, automagically, spotz, vdo, jmccrory, alextricity25, jasondotstar, KLevenstein, admin0, michaelgugino, ametts, v1k0d3n, | 16:06 |
| odyssey4me | severion, bgmccollum, darrenc, JRobinson__, asettle | 16:06 |
| michaelgugino | here | 16:07 |
| spotz | o/-ish | 16:07 |
| cloudnull | o/ | 16:08 |
| odyssey4me | https://bugs.launchpad.net/openstack-ansible/+bug/1576724 | 16:08 |
| openstack | Launchpad bug 1576724 in openstack-ansible "OSA doesn't handle hosts file in compute nodes for live migration" [Undecided,New] - Assigned to Matt Thompson (mattt416) | 16:08 |
| *** eil397 has joined #openstack-ansible | 16:08 | |
| odyssey4me | alright, it looks like mattt is covering that | 16:09 |
| odyssey4me | winggundamth did you see that mattt's looking for a response? | 16:09 |
| cloudnull | mattt: one thing to mention regarding the host names | 16:09 |
| odyssey4me | cloudnull mattt is out, so perhaps comment in the bug | 16:09 |
| cloudnull | if you set the host names in the user_config file to the actual hostname | 16:09 |
| cloudnull | and its longer than 62 ssh will break | 16:10 |
| cloudnull | ok | 16:10 |
| cloudnull | odyssey4me: | 16:10 |
| *** thorst_ has quit IRC | 16:10 | |
| cloudnull | ^ thats an issue for container create which is $HOSTNAME-CONTAINER-HASH | 16:11 |
| *** thorst_ has joined #openstack-ansible | 16:11 | |
| *** thorst_ has quit IRC | 16:11 | |
| *** TxGirlGeek has joined #openstack-ansible | 16:11 | |
| *** dweaver has joined #openstack-ansible | 16:12 | |
| odyssey4me | oh boo :( | 16:12 |
| *** sacharya_ has quit IRC | 16:12 | |
| odyssey4me | cloudnull are you going to make a note about that in the bug? | 16:12 |
| *** Min_Cai has quit IRC | 16:13 | |
| eil397 | does it mean that server hostname should be less then 42 ? | 16:13 |
| *** Ger-chervyak has quit IRC | 16:13 | |
| cloudnull | odyssey4me: im following up in the bug | 16:13 |
| eil397 | host server | 16:13 |
| prometheanfire | ohai | 16:13 |
| *** Ger-chervyak has joined #openstack-ansible | 16:14 | |
| cloudnull | we have this https://github.com/openstack/openstack-ansible/blob/master/playbooks/inventory/dynamic_inventory.py#L333-L339 | 16:14 |
| cloudnull | to guard against it | 16:14 |
| cloudnull | sorry 52 not 62 | 16:14 |
| *** karimb has joined #openstack-ansible | 16:15 | |
| eil397 | : - ) "Ensures that container names are not longer than 63" | 16:15 |
| *** karimb_ has quit IRC | 16:16 | |
| *** chandanc has quit IRC | 16:17 | |
| odyssey4me | https://bugs.launchpad.net/openstack-ansible/+bug/1578680 | 16:17 |
| openstack | Launchpad bug 1578680 in openstack-ansible "Running an update to /etc/hosts takes 2.5hrs in large environments" [Undecided,New] | 16:17 |
| *** chandanc_ has quit IRC | 16:17 | |
| *** thorst_ has joined #openstack-ansible | 16:17 | |
| cloudnull | i think we're good there. | 16:18 |
| cloudnull | unless we want to backport that | 16:18 |
| eil397 | wow "nova_api_os_compute_container-58184e45" - 39 chars | 16:18 |
| odyssey4me | cloudnull I'll leave the backport requesting to the bug reporter | 16:18 |
| odyssey4me | next up https://bugs.launchpad.net/openstack-ansible/+bug/1579091 | 16:18 |
| openstack | Launchpad bug 1579091 in openstack-ansible "There are no html docs for the plugins" [Undecided,New] | 16:18 |
| cloudnull | eil397: yup. | 16:18 |
| *** asettle has joined #openstack-ansible | 16:18 | |
| odyssey4me | clearly a wishlist item, and it would be neat | 16:19 |
| odyssey4me | let me take a look into that\ | 16:20 |
| cloudnull | +1 | 16:20 |
| cloudnull | wishlist | 16:20 |
| odyssey4me | next up https://bugs.launchpad.net/openstack-ansible/+bug/1579170 | 16:20 |
| openstack | Launchpad bug 1579170 in openstack-ansible "Disable or Enable a balanced node from haproxy via command line" [Undecided,New] - Assigned to Tom (tomjosekal) | 16:20 |
| odyssey4me | clearly wishlist too | 16:21 |
| logan- | that sounds like something that would be perfect for the osa ops repo discussed at the summit | 16:21 |
| cloudnull | do we need to install soimething for access to "haproxyctl" ? | 16:21 |
| odyssey4me | dunno | 16:21 |
| odyssey4me | logan- yeah, you're right - I need to get on that | 16:21 |
| *** thorst_ has quit IRC | 16:22 | |
| cloudnull | looks like we'd need to grabn https://github.com/flores/haproxyctl | 16:23 |
| cloudnull | *grab | 16:23 |
| *** asettle has quit IRC | 16:23 | |
| odyssey4me | cloudnull apparently there's a deb package | 16:24 |
| cloudnull | i just did a search in ubuntu 14.04 its not there. | 16:24 |
| cloudnull | at least not by default | 16:24 |
| cloudnull | maybe debian proper | 16:24 |
| odyssey4me | https://launchpad.net/ubuntu/+source/haproxyctl | 16:24 |
| d34dh0r53 | yeah, didn't see it on 14.04 either | 16:25 |
| odyssey4me | not in trusty, but it is in xenial | 16:25 |
| odyssey4me | anyway, moving on | 16:27 |
| odyssey4me | https://bugs.launchpad.net/openstack-ansible/+bug/1579612 | 16:27 |
| openstack | Launchpad bug 1579612 in openstack-ansible "heat domain setup fails with keystone using self signed certs" [Undecided,New] | 16:27 |
| odyssey4me | ah, that makes sense | 16:28 |
| jmccrory | has the keystone library been updated for that TODO? | 16:28 |
| odyssey4me | hmm, I vaguely recall that it might have | 16:28 |
| odyssey4me | jmccrory do you want to look into that? obviously the big win is to convert that to use the module instead of the CLI | 16:29 |
| jmccrory | sure | 16:29 |
| odyssey4me | clearly a bug - importance? | 16:29 |
| odyssey4me | I'm thinking medium | 16:29 |
| jmccrory | think medium too | 16:30 |
| odyssey4me | alright, next up https://bugs.launchpad.net/openstack-ansible/+bug/1579919 | 16:30 |
| openstack | Launchpad bug 1579919 in openstack-ansible "msgpack pip issue" [Undecided,New] | 16:30 |
| cloudnull | installing the package works | 16:32 |
| cloudnull | http://cdn.pasteraw.com/7imnz23vtbq1wdxmu6gj8ln9fsakljv | 16:32 |
| odyssey4me | it more than likely relates to a mismatch of pip/wheel versions between the rpc-repo and the host being built | 16:33 |
| odyssey4me | I've commented asking for more info | 16:33 |
| *** _deadnull is now known as deadnull_ | 16:33 | |
| cloudnull | kk | 16:33 |
| *** chandanc_ has joined #openstack-ansible | 16:33 | |
| *** chandanc has joined #openstack-ansible | 16:33 | |
| odyssey4me | https://bugs.launchpad.net/openstack-ansible/+bug/1580222 | 16:34 |
| openstack | Launchpad bug 1580222 in openstack-ansible "volume creation based on image source fails when using ceph backend" [Undecided,New] | 16:34 |
| *** flaviodsr has joined #openstack-ansible | 16:34 | |
| odyssey4me | gregfaust reported this earlier today | 16:34 |
| odyssey4me | mattt did some digging and found that https://review.openstack.org/#/c/293836/ caused this... it didn't take into account the non lvm back-ends | 16:35 |
| odyssey4me | so we need to do something like implement qemu-utils for all cinder-volume group members, or perhaps introduce some sort of package list per back-end? | 16:35 |
| *** electrofelix has joined #openstack-ansible | 16:36 | |
| *** abitha has joined #openstack-ansible | 16:36 | |
| odyssey4me | anyone want to pick that up? | 16:37 |
| odyssey4me | I would say High importance. | 16:37 |
| *** sdake_ is now known as sdake | 16:38 | |
| *** thorst_ has joined #openstack-ansible | 16:38 | |
| eil397 | just find right place to add deb package dependency to install qemu-utils | 16:38 |
| *** chhavi has quit IRC | 16:39 | |
| eil397 | : - ) | 16:39 |
| gregfaust | i'd be happy to submit a patch and get ATC for barcelnoa :), i was thinking that it's basiclaly a two line change to move the qemu-utils package from cinder_lvm_volume_apt_packages: to the common cinder_apt_packages: | 16:40 |
| gregfaust | I didn't want to suggest a solution earlier though until a few more folks looked at it | 16:40 |
| gregfaust | https://github.com/openstack/openstack-ansible-os_cinder/blob/master/vars/ubuntu-14.04.yml#L33 | 16:41 |
| odyssey4me | gregfaust it'll be a small change, for sure, but I don't think that's the right way because then the package will end up on all the api servers too | 16:41 |
| eil397 | +1 | 16:41 |
| *** thorst_ has quit IRC | 16:42 | |
| odyssey4me | I would suggest changing the task name for https://github.com/openstack/openstack-ansible-os_cinder/blob/master/tasks/cinder_install_apt.yml#L39-L50 to be 'Install apt packages for LVM backend' | 16:42 |
| eil397 | am I right that we should have in osa some examples to resolve this kind of situations ? when we need to isntall something on backend service and should not install it on frontend (api) | 16:42 |
| *** arbrandes1 has quit IRC | 16:43 | |
| odyssey4me | And adding another task with a corresponding var to 'Install apt packages commonly used for cinder-volume backends' | 16:43 |
| *** karimb has quit IRC | 16:43 | |
| odyssey4me | the conditional should just be 'when: inventory_hostname in groups['cinder_volume']' | 16:44 |
| odyssey4me | does that make sense? | 16:44 |
| *** thorst_ has joined #openstack-ansible | 16:44 | |
| evrardjp | it does to me | 16:44 |
| gregfaust | I'll mention that in the future I'm planning to do some work (or have NetApp do it), to add a cinder backend for SolidFire. It would be nice if there were an easy way to add new backends. | 16:45 |
| odyssey4me | gregfaust I expect that most back-ends just need a python installable, but it's hard to tell without a specific case | 16:45 |
| odyssey4me | the neutron role has what is probably the closest sort of pattern you can look at | 16:46 |
| *** chhavi has joined #openstack-ansible | 16:46 | |
| odyssey4me | it includes patterns for Nuage (completely properietary) and PLUMgrid (freely available python bits) | 16:46 |
| *** sacharya has joined #openstack-ansible | 16:47 | |
| evrardjp | also it has the easy switch facility that could also be used | 16:49 |
| *** thorst_ has quit IRC | 16:49 | |
| *** sacharya has quit IRC | 16:50 | |
| *** sacharya has joined #openstack-ansible | 16:50 | |
| kencjohnston | hi osa team, is there a place where service scaling (adding additional control plane hosts, breaking out services into dedicated hosts, etc) is documented? | 16:52 |
| flaviodsr | hey odyssey4me, as we have talked I have the sahara role ready for sharing. How do I proceed? | 16:53 |
| evrardjp | kencjohnston it should be :D | 16:54 |
| odyssey4me | alrightbug triage is done for today | 16:54 |
| odyssey4me | thanks all | 16:54 |
| odyssey4me | kencjohnston adding additional hosts is fairly obvious once you've done an install because the process isn't much different... we have called out how to expand your compute hosts, but perhaps we need more on the different profiles | 16:55 |
| *** eil397 has quit IRC | 16:55 | |
| odyssey4me | kencjohnston we have http://docs.openstack.org/developer/openstack-ansible/install-guide/ops-addcomputehost.html for now, that is all | 16:55 |
| evrardjp | kencjohnston if it's on install you have this http://docs.openstack.org/developer/openstack-ansible/install-guide/configure-hostlist.html | 16:55 |
| *** arbrandes has joined #openstack-ansible | 16:55 | |
| kencjohnston | OK thanks odyssey4me and evrardjp | 16:56 |
| kencjohnston | I found a rackspace doc - http://docs.rackspace.com/rpc/api/v11/bk-rpc-ops/content/containers-scaling-services-deployment-groups.html | 16:56 |
| *** eil397 has joined #openstack-ansible | 16:56 | |
| palendae | Has anyone seen the gate failing on openssh-client installations? http://logs.openstack.org/52/308052/16/check/gate-openstack-ansible-dsvm-commit/fb754fa/console.html | 16:56 |
| odyssey4me | kencjohnston would you mind registering a bug with the title 'Doc: <short title>' - preferably one for each procedure, then describe in the bug what you'd like to see documented | 16:56 |
| *** deadnull_ has quit IRC | 16:56 | |
| kencjohnston | odyssey4me will do. | 16:57 |
| *** kiranv has joined #openstack-ansible | 16:57 | |
| evrardjp | kencjohnston http://docs.openstack.org/developer/openstack-ansible/install-guide/index.html would be your general documentation, and the chapter 4 is what you are looking for | 16:57 |
| evrardjp | (at least for the install) | 16:57 |
| kencjohnston | evrardjp I was mostly interested in post installation scaling. | 16:58 |
| kencjohnston | but thank you, I'll review and add bugs | 16:58 |
| *** karl907 has quit IRC | 16:58 | |
| palendae | I don't think we have anything about scaling infrastructure servers - I'd imagine we need notes about adding mariadb/galera nodes in such a way that you keep odd numbers | 16:59 |
| odyssey4me | thanks kencjohnston - that RPC documentation is OK, but there are simpler ways to do that and we should add that to our docs | 16:59 |
| *** kiranv has quit IRC | 16:59 | |
| odyssey4me | palendae ++ | 16:59 |
| evrardjp | adding nodes on infra should follow the same behavior | 16:59 |
| odyssey4me | palendae konichi-wa? | 16:59 |
| evrardjp | adding nodes in inventory, running playbooks, etc | 16:59 |
| palendae | And I'm not sure how replication for galera would account for moving the data | 16:59 |
| palendae | odyssey4me: Not yet :) One more day in the US | 16:59 |
| gregfaust | kencjohnston, I was just starting to thinkg about the same thing, for example multiple rabbitmqs with different replication policies to speed up some of the really noisy queues like celiometer notifications by making them non-HA and eliminating the replication overhead. | 17:00 |
| *** clickboo_ has joined #openstack-ansible | 17:00 | |
| palendae | evrardjp: Yeah, adding the host to the inventory and running playbooks shouldn't be too bad..but making sure replication for galera and rabbitmq are questions for me. Mostly because I'm not as familiar with them | 17:00 |
| *** clickboom has quit IRC | 17:00 | |
| evrardjp | good question for the galera and rabbit | 17:00 |
| palendae | They may handle it themselves, but I'd have to research | 17:01 |
| evrardjp | I remember having issues with galera on kilo with 5 nodes | 17:01 |
| palendae | Load balancers would need updating, etc | 17:01 |
| evrardjp | palendae haproxy should be fine :p | 17:01 |
| palendae | evrardjp: Sure, but not every install uses that :) | 17:01 |
| evrardjp | I know :D | 17:01 |
| odyssey4me | gregfaust sooo, if someone could submit a patch for that based on the Fuel policy changes then it would probably work quite nicely.... see https://etherpad.openstack.org/p/openstack-ansible-newton-roadmap-brainstorm near the bottom under 'Performance' | 17:02 |
| *** Andrew_jedi has quit IRC | 17:02 | |
| evrardjp | according to my basic knowledge of galera, I think if we have a problem it should be filed as a bug, because it's probably an issue with the way we are doing it :p | 17:02 |
| odyssey4me | palendae I would think that going beyond 3 Galera nodes should need some special config to implement multicast data replication to help ensure that performance doesn't degrade | 17:03 |
| *** clickboo_ has quit IRC | 17:03 | |
| palendae | Yeah, something like that | 17:03 |
| evrardjp | that's a good "plus" feature | 17:04 |
| kencjohnston | bug submitted - https://bugs.launchpad.net/openstack-ansible/+bug/1580264 | 17:04 |
| openstack | Launchpad bug 1580264 in openstack-ansible "Doc: Operations - Infrastructure Scaling" [Undecided,New] | 17:04 |
| stevelle | gregfaust: your mention of ceilo notifications is interesting to me | 17:04 |
| stevelle | if you had a separate rabbit cluster I'm not sure why you would need a diff policy | 17:05 |
| stevelle | but I don't know what you're using ceilo for | 17:05 |
| evrardjp | my concern was more about the initial sync to a new node after having lots of data in the db already, which should be slow but working (galera is handling that IIRC) | 17:05 |
| *** karimb has joined #openstack-ansible | 17:06 | |
| evrardjp | but I'm tired ... didn't get the conversation :p | 17:06 |
| palendae | evrardjp: Yeah, my concern's basically the same | 17:06 |
| palendae | Should function, but may experience slowness til sync is ciomplete | 17:07 |
| palendae | complete* | 17:07 |
| evrardjp | my concern would be the ansible behaviour in these cases too | 17:07 |
| *** sacharya has quit IRC | 17:08 | |
| palendae | kencjohnston, evrardjp: I'd think answering some of these questions may be a good case for checking out an OSIC cluster :) | 17:08 |
| odyssey4me | stevelle Fuel does this: https://github.com/openstack/fuel-library/blob/686bed1b4f090d7f6fd368b94a5ced12c8e28744/files/fuel-ha-utils/ocf/set_rabbitmq_policy.sh | 17:08 |
| *** sacharya has joined #openstack-ansible | 17:08 | |
| evrardjp | palendae that's true | 17:08 |
| kencjohnston | palendae +1, teh OSIC team actually has a test like this on our roadmap | 17:08 |
| palendae | kencjohnston: Awesome | 17:09 |
| stevelle | odyssey4me: again, not sure if that is needed or the best solution | 17:09 |
| *** Andrew_jedi has joined #openstack-ansible | 17:10 | |
| *** thorst_ has joined #openstack-ansible | 17:11 | |
| stevelle | I don't want to just copy what someone else did without understanding why | 17:11 |
| odyssey4me | sure, as I understand it it's pretty much that there are messages that need to endure (like notifications) and messages that don't (like agent status messages) | 17:14 |
| odyssey4me | so if the cluster is configured to not bother doing the ha mirroring for the queues that hold messages that don't need ot endure, then it can improve performance | 17:15 |
| *** thorst_ has quit IRC | 17:15 | |
| *** thorst_ has joined #openstack-ansible | 17:16 | |
| stevelle | to exaggerate a little, would be easier to just send messages from ceilo agents to /dev/null | 17:16 |
| *** arbrandes has quit IRC | 17:16 | |
| odyssey4me | the ha queues make a significant difference in message processing speed - see http://docs.openstack.org/developer/performance-docs/test_results/mq/rabbitmq/index.html | 17:16 |
| odyssey4me | and http://docs.openstack.org/developer/performance-docs/test_results/mq_ha/index.html | 17:17 |
| gregfaust | odyssey4me, that's my understanding too, I think the lead dev for rabbitmq gave a presentation about tuning queues (and erlang) in austin, but I don't have the link to the video handy. he said that mirroring queues is one of the most expensive operations in rabbitmq. | 17:17 |
| *** arbrandes has joined #openstack-ansible | 17:17 | |
| *** karl907 has joined #openstack-ansible | 17:19 | |
| *** Andrew_jedi has quit IRC | 17:19 | |
| *** karl907_ has joined #openstack-ansible | 17:19 | |
| *** weezS has quit IRC | 17:20 | |
| *** quantumshadow has joined #openstack-ansible | 17:20 | |
| stevelle | I think the idea here is /why/ are ceilo messages wanted but yet they don't need to have ensured delivery | 17:22 |
| *** thorst_ has quit IRC | 17:23 | |
| *** karl907 has quit IRC | 17:24 | |
| *** karl907_ has quit IRC | 17:24 | |
| *** karl907 has joined #openstack-ansible | 17:24 | |
| odyssey4me | stevelle good question | 17:25 |
| odyssey4me | Time for me to head home. Have a great day everyone! | 17:25 |
| cloudnull | cheers odyssey4me | 17:29 |
| cloudnull | have a good one | 17:29 |
| *** thorst_ has joined #openstack-ansible | 17:30 | |
| *** chandanc_ has quit IRC | 17:30 | |
| *** chandanc has quit IRC | 17:31 | |
| *** thorst_ has quit IRC | 17:34 | |
| *** karimb has quit IRC | 17:36 | |
| *** TxGirlGeek has quit IRC | 17:36 | |
| openstackgerrit | Samuel Matzek proposed openstack/openstack-ansible: Refactor ceph_client for multi-OS and ceph https://review.openstack.org/312692 | 17:38 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_swift: Covert swift python to pypy https://review.openstack.org/314587 | 17:39 |
| *** chhavi has quit IRC | 17:41 | |
| *** thorst_ has joined #openstack-ansible | 17:43 | |
| *** keedya has joined #openstack-ansible | 17:46 | |
| mhayden | antonym: thanks for that libe migration bug | 17:46 |
| *** thorst_ has quit IRC | 17:47 | |
| *** asettle has joined #openstack-ansible | 17:49 | |
| *** asettle has quit IRC | 17:50 | |
| *** chhavi has joined #openstack-ansible | 17:54 | |
| *** quantumshadow has left #openstack-ansible | 17:56 | |
| *** quantumshadow has joined #openstack-ansible | 17:59 | |
| *** thorst_ has joined #openstack-ansible | 18:02 | |
| *** chhavi has quit IRC | 18:03 | |
| *** mkrish004c has joined #openstack-ansible | 18:03 | |
| *** ben___ has joined #openstack-ansible | 18:08 | |
| mkrish004c | Hi Guys, @jmccrory, sday we discussed on an issue, that compute node status is not listing under service list. the issue was due to different nova version in conductor and nova-compute. Thanks | 18:09 |
| *** Ger-chervyak has quit IRC | 18:09 | |
| jmccrory | mkrish004c: thanks, good to know. were different release versions of openstack deployed over each other? | 18:10 |
| *** jorge_munoz has quit IRC | 18:13 | |
| mkrish004c | @jmccrory, yeah it was 2.12 in compute node and 2.2 in conductor | 18:14 |
| *** chhavi has joined #openstack-ansible | 18:15 | |
| *** jorge_munoz has joined #openstack-ansible | 18:16 | |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible-rabbitmq_server: Multi-distro framework for rabbitmq_server role https://review.openstack.org/286282 | 18:18 |
| *** Ger-chervyak has joined #openstack-ansible | 18:18 | |
| *** jmccrory_ has quit IRC | 18:19 | |
| *** electrofelix has quit IRC | 18:19 | |
| *** ben___ is now known as bengentil | 18:23 | |
| *** TxGirlGeek has joined #openstack-ansible | 18:27 | |
| *** karl907 has quit IRC | 18:28 | |
| *** sacharya has quit IRC | 18:31 | |
| *** mkrish004c has quit IRC | 18:42 | |
| *** mkrish004c has joined #openstack-ansible | 18:42 | |
| *** bengentil has quit IRC | 18:42 | |
| *** bengentil has joined #openstack-ansible | 18:42 | |
| *** TxGirlGeek has quit IRC | 18:44 | |
| *** bengentil has quit IRC | 18:45 | |
| *** thorst_ has quit IRC | 18:48 | |
| *** thorst_ has joined #openstack-ansible | 18:49 | |
| *** TxGirlGeek has joined #openstack-ansible | 18:50 | |
| *** bengentil has joined #openstack-ansible | 18:51 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-galera_client: Implement multi-distro support https://review.openstack.org/314717 | 18:51 |
| *** sdake has quit IRC | 18:52 | |
| *** weezS has joined #openstack-ansible | 18:53 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-galera_client: Implement multi-distro support https://review.openstack.org/314717 | 18:55 |
| *** thorst_ has quit IRC | 18:57 | |
| *** TxGirlGeek has quit IRC | 18:59 | |
| *** TxGirlGeek has joined #openstack-ansible | 18:59 | |
| *** TxGirlGeek has quit IRC | 19:01 | |
| *** TxGirlGeek has joined #openstack-ansible | 19:02 | |
| *** thorst_ has joined #openstack-ansible | 19:03 | |
| *** TxGirlGeek has quit IRC | 19:04 | |
| *** TxGirlGeek has joined #openstack-ansible | 19:04 | |
| *** thorst_ has quit IRC | 19:08 | |
| *** thorst_ has joined #openstack-ansible | 19:09 | |
| *** pcaruana has quit IRC | 19:12 | |
| *** karimb has joined #openstack-ansible | 19:12 | |
| *** jorge_munoz has quit IRC | 19:13 | |
| *** Ger-cher_ has joined #openstack-ansible | 19:14 | |
| *** thorst_ has quit IRC | 19:14 | |
| *** Ger-cher_ has quit IRC | 19:14 | |
| *** Ger-chervyak has quit IRC | 19:15 | |
| *** Ger-chervyak has joined #openstack-ansible | 19:15 | |
| *** thorst_ has joined #openstack-ansible | 19:15 | |
| *** jorge_munoz has joined #openstack-ansible | 19:16 | |
| spotz | oh odyssey4me cloudnull was actually working on the n1 conversion I had taken so I'll look for another one. Need to update the etherpad when I remember | 19:16 |
| *** thorst_ has quit IRC | 19:20 | |
| *** jthorne has quit IRC | 19:21 | |
| *** jthorne has joined #openstack-ansible | 19:21 | |
| *** thorst_ has joined #openstack-ansible | 19:22 | |
| *** sacharya has joined #openstack-ansible | 19:22 | |
| *** jthorne has quit IRC | 19:23 | |
| *** thorst_ has quit IRC | 19:23 | |
| *** thorst_ has joined #openstack-ansible | 19:23 | |
| *** jthorne has joined #openstack-ansible | 19:24 | |
| *** thorst_ has quit IRC | 19:30 | |
| *** thorst_ has joined #openstack-ansible | 19:30 | |
| mattt | cloudnull: i was hoping there was nothing needing to be done short of a doc update for that review :P | 19:31 |
| cloudnull | which one? | 19:32 |
| mattt | cloudnull: the live migration hostname bug (sorry, not review) | 19:32 |
| cloudnull | ah yea. | 19:33 |
| cloudnull | im not sure what we need to do to make that happy... | 19:33 |
| cloudnull | so long as no containers are on the compute nodes it should be safe to just use the host name | 19:34 |
| cloudnull | but if its a compute node w/ containers on it we'll have to make sure the container rendered names are <63 | 19:34 |
| *** thorst_ has quit IRC | 19:36 | |
| jmccrory | cloudnull: have you come across this? http://logs.openstack.org/82/286282/36/experimental/gate-openstack-ansible-rabbitmq_server-ansible-func-centos-7-nv/e80a7d4/console.html#_2016-05-10_18_26_27_692 | 19:37 |
| *** thorst_ has joined #openstack-ansible | 19:37 | |
| *** jthorne has quit IRC | 19:38 | |
| cloudnull | not w/ that domain | 19:39 |
| *** jthorne has joined #openstack-ansible | 19:39 | |
| cloudnull | i've seen it when they update the keys which requires an upgrade to happen first | 19:39 |
| cloudnull | but not for https://bootstrap.pypa.io | 19:40 |
| *** jthorne has quit IRC | 19:41 | |
| *** TxGirlGeek has quit IRC | 19:41 | |
| jmccrory | weird thing is the first container goes through that task without a problem earlier in the log | 19:41 |
| *** thorst_ has quit IRC | 19:42 | |
| *** jayc has quit IRC | 19:43 | |
| odyssey4me | jmccrory it fails on both pypa and github - that's suspicious | 19:46 |
| *** jthorne has joined #openstack-ansible | 19:51 | |
| *** albertcard has joined #openstack-ansible | 19:53 | |
| *** chhavi has quit IRC | 19:56 | |
| *** jayc has joined #openstack-ansible | 19:58 | |
| *** weezS has quit IRC | 20:00 | |
| *** javeriak_ has quit IRC | 20:05 | |
| *** sacharya_ has joined #openstack-ansible | 20:13 | |
| openstackgerrit | Nolan Brubaker proposed openstack/openstack-ansible: Check for two IP addresses assigned to same host https://review.openstack.org/308052 | 20:15 |
| *** sacharya has quit IRC | 20:16 | |
| *** Ger-chervyak has quit IRC | 20:22 | |
| *** jthorne has quit IRC | 20:22 | |
| *** jthorne has joined #openstack-ansible | 20:23 | |
| *** weezS has joined #openstack-ansible | 20:24 | |
| *** abitha has quit IRC | 20:26 | |
| *** keedya has quit IRC | 20:27 | |
| *** jthorne has quit IRC | 20:27 | |
| spotz | sigmavirus24: DOC is so I can find it on the list to review it:) | 20:31 |
| *** thorst_ has joined #openstack-ansible | 20:32 | |
| sigmavirus24 | spotz: the rest of the title can still be rewritten to be less redundant ;) | 20:33 |
| spotz | yeah, but don't blame him for following our ask:) | 20:33 |
| openstackgerrit | Jacob Wagner proposed openstack/openstack-ansible-os_tempest: Allow tempest network overrides for public and private net https://review.openstack.org/314760 | 20:34 |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible-plugins: Support users without projects in keystone library https://review.openstack.org/314761 | 20:34 |
| *** johnmilton has quit IRC | 20:34 | |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible-plugins: Support users without projects in keystone library https://review.openstack.org/314761 | 20:35 |
| *** Ger-chervyak has joined #openstack-ansible | 20:37 | |
| openstackgerrit | Michael Gugino proposed openstack/openstack-ansible-galera_client: Implement Xenial Support https://review.openstack.org/314765 | 20:40 |
| mhayden | okay, this gertty thing is handy | 20:42 |
| mhayden | thanks prometheanfire ;) | 20:42 |
| michaelgugino | cloudnull: https://review.openstack.org/#/c/314765/ | 20:42 |
| cloudnull | ah cool . | 20:43 |
| cloudnull | ill rebase | 20:43 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-security: Add CentOS 7 and Ubuntu 16.04 support https://review.openstack.org/313697 | 20:43 |
| prometheanfire | michaelgugino: :D | 20:45 |
| prometheanfire | mhayden: :D | 20:45 |
| prometheanfire | michaelgugino: you too though | 20:45 |
| michaelgugino | prometheanfire: eh? | 20:46 |
| prometheanfire | michaelgugino: mistab | 20:46 |
| michaelgugino | oh, gotcha ;) | 20:46 |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible-os_heat: Use keystone library for heat_stack_domain_admin https://review.openstack.org/314769 | 20:46 |
| mhayden | jmccrory: i was looking over 314761 -- in what situation would a project == None and a project_name != None ? | 20:47 |
| jmccrory | if someone wants to create a user within a given project_name and it doesn't exist | 20:47 |
| mhayden | jmccrory: okay, i think i see the logic now | 20:48 |
| michaelgugino | cloudnull: for the tests in galera client, I don't see a repo server being used. Should the lxc containers have build-essential installed? I was building MySQL-python by hand inside the container | 20:49 |
| *** kstev has quit IRC | 20:50 | |
| cloudnull | yes. it needs to be there. in the 14.04 base image we kinda force it to be there but for MySQL-python it needs build essential | 20:51 |
| jmccrory | mhayden: if they don't give a project_name (domain only user), both are None and that's just fine | 20:52 |
| cloudnull | and a few other devel libs | 20:52 |
| mhayden | jmccrory: nice -- just learned something new about keystone i didn't know before :) | 20:52 |
| mhayden | i had assumed a project was required | 20:53 |
| michaelgugino | cloudnull: needs python-dev and build-essential | 20:53 |
| jmccrory | heh i didn't really know it either before working on this | 20:53 |
| *** stelouie has joined #openstack-ansible | 20:53 | |
| michaelgugino | should be include some logic to install those packages if the repo server is not available? We don't need it if we're pulling wheels from a repo server. | 20:53 |
| cloudnull | maybe they should only be installed in developer mode? | 20:55 |
| *** Ger-chervyak has quit IRC | 20:56 | |
| michaelgugino | cloudnull: osa/playbooks/galera-install.yml: | 20:56 |
| michaelgugino | roles: - role: "pip_lock_down" when: "(groups['repo_all'] | length) >= 1" | 20:56 |
| mhayden | jmccrory: tested! works well (and my mind is blown) | 20:57 |
| *** Ger-chervyak has joined #openstack-ansible | 20:57 | |
| michaelgugino | I think we should key off that, because if someone chooses not to deploy a repo server, it's going to be broken. | 20:57 |
| jmccrory | haha thanks for testing | 20:57 |
| michaelgugino | this is true if someone has a pip mirror and they only have sdists. | 20:58 |
| michaelgugino | cloudnull: check experimental passed for 16.04 on galera client. Somehow the bdists were made without error. I'm guessing the host we're deploying to in the ci environment has the necessary packages installed already | 21:05 |
| cloudnull | +1 I think the infra hosts have all the things needed. | 21:05 |
| *** mkrish004c has quit IRC | 21:07 | |
| *** thorst_ has quit IRC | 21:08 | |
| *** thorst_ has joined #openstack-ansible | 21:09 | |
| *** psilvad has quit IRC | 21:09 | |
| *** johnmilton has joined #openstack-ansible | 21:11 | |
| errr | what playbook do you need to rerun if your container didnt get a virtualenv created in it? I have the dir but it only added a bin fir inside of it and nothing else | 21:15 |
| errr | s,fir,dir, | 21:16 |
| *** retreved has quit IRC | 21:16 | |
| *** ametts has quit IRC | 21:17 | |
| *** thorst_ has quit IRC | 21:17 | |
| *** johnmilton has quit IRC | 21:17 | |
| *** galstrom is now known as galstrom_zzz | 21:18 | |
| *** karimb has quit IRC | 21:20 | |
| *** phalmos has quit IRC | 21:21 | |
| *** Mudpuppy has quit IRC | 21:21 | |
| *** bengentil has quit IRC | 21:23 | |
| *** smatzek has quit IRC | 21:23 | |
| errr | never mind, I think this is caused by me failing to ansible correctly | 21:24 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-galera_client: Update Multi-distro support for CentOS7 https://review.openstack.org/314717 | 21:25 |
| *** karimb has joined #openstack-ansible | 21:26 | |
| openstackgerrit | Michael Gugino proposed openstack/openstack-ansible-galera_server: [WIP] Implement Xenial Support https://review.openstack.org/312602 | 21:28 |
| *** johnmilton has joined #openstack-ansible | 21:29 | |
| openstackgerrit | Michael Gugino proposed openstack/openstack-ansible-galera_server: Implement Xenial Support https://review.openstack.org/312602 | 21:29 |
| openstackgerrit | Michael Gugino proposed openstack/openstack-ansible-galera_server: Implement Xenial Support https://review.openstack.org/312602 | 21:34 |
| *** jamesden_ has quit IRC | 21:34 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_swift: Covert swift python to pypy https://review.openstack.org/314587 | 21:35 |
| errr | if my new role Im making requires packages from pip not currently part of the frozen pip stuff how do I add the packages I need to the local mirror? | 21:37 |
| *** thorst_ has joined #openstack-ansible | 21:38 | |
| cloudnull | errr: if your role is in /etc/ansible/roles or playbooks/roles the repo-build play will index and add your new packages to the repo-server | 21:39 |
| errr | cloudnull: It is, but Im sure Ive done something wrong because I get this: https://gist.github.com/michaelrice/4133194b4f0b8fc79115d72e1bf38175 | 21:40 |
| cloudnull | just make sure your variable names end in pip_packages | 21:40 |
| errr | which Im sure is caused by what Im needing not being in there, but thats just a guess | 21:40 |
| cloudnull | so something like <ROLE>_pip_packages: [1,2,3,4] | 21:40 |
| cloudnull | I assume you have msgpack-python in your variable list | 21:41 |
| errr | ah ok, Ill try that, what playbook will I need to rerun to make that rebuild? | 21:41 |
| cloudnull | yes | 21:41 |
| errr | I dont | 21:41 |
| cloudnull | its a dep coming from tooz | 21:41 |
| cloudnull | and as long as you see that added to the constaints it should work | 21:42 |
| errr | I didnt add it, Ill have to go add it now, once its added which playbook will rebuild the pip mirror? | 21:42 |
| *** thorst_ has quit IRC | 21:42 | |
| cloudnull | you may need to add a section on your role defaults for the following to pull in all of the deps https://github.com/openstack/openstack-ansible-os_glance/blob/master/defaults/main.yml#L21-L27 | 21:43 |
| cloudnull | once youve added the requirements / constraints run repo-build.yml | 21:43 |
| cloudnull | that will update the centralized repo | 21:44 |
| errr | sweet thanks. Ill give that a whirl | 21:44 |
| cloudnull | cool | 21:46 |
| cloudnull | stevelle: thanks for the review on the swift bits earlier | 21:46 |
| stevelle | cloudnull: yaya, I'm enthusiastic about the potential. | 21:46 |
| cloudnull | its really no joke | 21:47 |
| cloudnull | i was skeptical at first | 21:47 |
| cloudnull | but swift bench sails w/ pypy vs cpython | 21:47 |
| cloudnull | and its not a big change for us to take advantage of it | 21:47 |
| stevelle | Also, totally want to run gnocchi on pypy | 21:47 |
| cloudnull | yea. the pattern should be simple to implement on services we want it to be default. | 21:48 |
| stevelle | I'm sure most are I/O bound but anything that might be CPU intensive... | 21:48 |
| cloudnull | its all 1 dep and a single task to make it go | 21:48 |
| cloudnull | and the dep is available on all of our "supported" os's so it should be fairly universal . | 21:49 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-galera_client: Update Multi-distro support for CentOS7 https://review.openstack.org/314717 | 21:50 |
| *** sacharya has joined #openstack-ansible | 21:52 | |
| *** jthorne has joined #openstack-ansible | 21:53 | |
| mrda | Morning OSA | 21:53 |
| spotz | hey mrda | 21:53 |
| stevelle | I expect some projects will encounter issues or the pypy gates wouldn't have been disabled. | 21:54 |
| *** sacharya_ has quit IRC | 21:54 | |
| *** stelouie_ has joined #openstack-ansible | 21:55 | |
| *** stelouie has quit IRC | 21:55 | |
| cloudnull | stevelle: yea projects that do a lot of monkey patching are have trouble with pypy. | 21:56 |
| *** weezS has quit IRC | 21:56 | |
| cloudnull | i think we can selectively enable it | 21:56 |
| cloudnull | and if we do it earlier enough in in the cycle we can test its usefulness (or not) and find issues that need to be looked at elsewhere. | 21:57 |
| cloudnull | morning mrda | 21:57 |
| mrda | hey spotz and cloudnull! | 21:57 |
| *** jthorne has quit IRC | 22:00 | |
| stelouie_ | Hi all, I'm new to openstack-ansible but am studying its organization to apply a similar structure to kubernetes/contrib/ansible. | 22:00 |
| stelouie_ | I wanted to ask if I'm understanding this right, I'm just looking at openstack-ansible and openstack-ansible-rabbitmq-server as an example. Does openstack-ansible run with any tags for rabbitmq here: https://github.com/openstack/openstack-ansible/blob/master/scripts/run-playbooks.sh#L100 I see that everything in the rabbitmq role is tagged but cannot find any use of it. Following main.yml I saw that everything in the rabbitmq rol | 22:00 |
| stelouie_ | e is called eventually, and I believe you'd filter out using tags. Is there any example or spec on how you set up roles/tags. | 22:00 |
| stelouie_ | My understanding is that openstack-ansible includes the role, main.yml sets the given flow through all plays included in that role and tags are used to selectively run tasks. So one has to be certain that the play order works regardless of tag sets included the role. | 22:00 |
| michaelgugino | cloudnull: I'd like to get some reviews on https://review.openstack.org/#/c/298765/ I think it's pretty near completion. | 22:01 |
| cloudnull | I'll give that another go this evening | 22:01 |
| michaelgugino | stelouie_: I believe tags are used for development and upgrades | 22:01 |
| cloudnull | michaelgugino: +1 | 22:02 |
| *** stelouie_ has quit IRC | 22:04 | |
| cloudnull | chao stelouie_ :) | 22:04 |
| errr | cloudnull: Im still getting the same error and I installed links2 and went to the pip mirror and saw tooz and msgpack-python in the mirror. Any idea what could be causing it now? Is there a chance that pip is using some cache or something and isnt actually hitting the mirror? | 22:05 |
| cloudnull | check the pip.conf file | 22:06 |
| cloudnull | maybe its locked down to the wrong mirror URL ? | 22:06 |
| errr | I used the mirror from the pip.conf to verify with links2 | 22:07 |
| *** mageshgv has quit IRC | 22:07 | |
| *** stelouie has joined #openstack-ansible | 22:09 | |
| cloudnull | errr: can you run `pip install msgpack-python` | 22:09 |
| *** markvoelker_ has joined #openstack-ansible | 22:10 | |
| *** sigmavirus24 is now known as sigmavirus24_awa | 22:10 | |
| errr | cloudnull: no, but msgpack_python-0.4.7-cp27-cp27mu-manylinux1_x86_64.whl is on the mirror, I am able to wget it | 22:10 |
| stelouie | michaelgugino: whats the choice to use tags instead of making a "upgrade" playbook that would call things like stop, install, start for a restart playbook | 22:11 |
| *** thorst_ has joined #openstack-ansible | 22:12 | |
| errr | cloudnull: I get: Could not find a version that satisfies the requirement msgpack-python (from versions: ) No matching distribution found for msgpack-python | 22:12 |
| cloudnull | errr: can you wget and install it ? | 22:12 |
| cloudnull | or just use the url? | 22:13 |
| stelouie | https://github.com/hpe-helion-os/helion-ansible/tree/master/hp_ceph_ansible is set up very differently as compared to openstack-ansible, just curious why go tags instead of something like hpe, here | 22:13 |
| cloudnull | pip install http://<URL> | 22:13 |
| errr | cloudnull: msgpack_python-0.4.7-cp27-cp27mu-manylinux1_x86_64.whl is not a supported wheel on this platform. | 22:14 |
| errr | that seems odd. | 22:14 |
| cloudnull | stelouie: the hpe playbook are setup using an ansible anti pattern which works well for their product however we're aiming to be inline with Ansible best practices: http://docs.ansible.com/ansible/playbooks_best_practices.html | 22:15 |
| michaelgugino | stelouie: we have a large code base. Using another set of plays would be more to maintain. | 22:15 |
| cloudnull | so you'll see some structural differences between what we're doing if youre used to the hpe playbooks | 22:15 |
| cloudnull | errr: hum... | 22:15 |
| errr | cloudnull: I just have a cloud server with an AIO | 22:16 |
| *** thorst_ has quit IRC | 22:16 | |
| cloudnull | 14.04 ? | 22:16 |
| errr | yeah | 22:16 |
| *** thorst_ has joined #openstack-ansible | 22:17 | |
| cloudnull | well thats odd. | 22:17 |
| * cloudnull looking more | 22:17 | |
| michaelgugino | stelouie: it took me a bit to get used to the layout of this project and how things work. Now that I'm accustomed to it, I feel this is probably the best way to develop roles and plays with ansible. This project serves as a reference point for all other ansible plays I write. | 22:17 |
| stelouie | michaelgugino: I think thats what I'm still trying to grasp, at a glance the hpe ansible stuff is easier to ingest quickly. | 22:18 |
| cloudnull | errr: the rpc-repo package is working | 22:18 |
| cloudnull | pip install http://rpc-repo.rackspace.com/os-releases/11.2.15/msgpack_python-0.4.7-cp27-cp27mu-manylinux1_x86_64.whl | 22:19 |
| *** thorst_ has quit IRC | 22:19 | |
| cloudnull | and it looks the same based on thename ? | 22:19 |
| cloudnull | x86_64 system ? | 22:19 |
| errr | cloudnull: yes and they both have the same md5sum | 22:20 |
| *** Ger-cher_ has joined #openstack-ansible | 22:20 | |
| errr | cloudnull: and if I try to install the one you linked same error about platform | 22:21 |
| *** spotz is now known as spotz_zzz | 22:21 | |
| stelouie | michaelgugino: for an upgrade, would you just run the given role with the tag: <service>-upgrade, I'm not seeing the obvious upgrade flow | 22:22 |
| *** sdake has joined #openstack-ansible | 22:22 | |
| cloudnull | errr: what kernel ? | 22:22 |
| errr | 3.13.0-85-generic | 22:22 |
| errr | thats what I get in the container from a uname -r | 22:22 |
| *** Ger-chervyak has quit IRC | 22:23 | |
| stevelle | stelouie: take a look at the mitaka stable branch for an example of how we deal with upgrades: https://github.com/openstack/openstack-ansible/blob/stable/mitaka/scripts/run-upgrade.sh | 22:23 |
| cloudnull | errr: that all looks fine | 22:25 |
| michaelgugino | stelouie: there is an upgrade script | 22:25 |
| cloudnull | the only think i could think of is that the base os is 32 bit ? | 22:25 |
| michaelgugino | ^^^ | 22:25 |
| *** jayc has quit IRC | 22:25 | |
| stevelle | the tags are a convenience provided for operators to make changes to an existing deploy, but there are few if any hard dependencies on those tags | 22:25 |
| *** Ger-cher_ has quit IRC | 22:26 | |
| errr | cloudnull: Linux mike-rpc-aio 3.13.0-85-generic #129-Ubuntu SMP Thu Mar 17 20:50:15 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux | 22:26 |
| errr | cloudnull: its one of our standard 15g cloud servers. I think its in ord | 22:26 |
| cloudnull | that is odd. | 22:27 |
| cloudnull | errr: http://cdn.pasteraw.com/3fp53ozud2x9g8qqb9ivcbr2n4n1qnv http://cdn.pasteraw.com/3fp53ozud2x9g8qqb9ivcbr2n4n1qnv | 22:29 |
| cloudnull | bah... | 22:29 |
| cloudnull | http://cdn.pasteraw.com/3fp53ozud2x9g8qqb9ivcbr2n4n1qnv | 22:30 |
| cloudnull | so thats my kernel, OS , and installing the package with ``pip install http://rpc-repo.rackspace.com/os-releases/11.2.15/msgpack_python-0.4.7-cp27-cp27mu-manylinux1_x86_64.whl`` | 22:30 |
| *** thorst_ has joined #openstack-ansible | 22:30 | |
| cloudnull | that was on the host | 22:31 |
| cloudnull | this is in the container | 22:31 |
| cloudnull | http://cdn.pasteraw.com/4ys8xrq71pqfxx5vaqno3absjt8tdlg | 22:31 |
| *** karimb has quit IRC | 22:31 | |
| stevelle | stelouie: upgrades may be a tangent so not to dwell on it but if it is genuinely a topic for you, while the mitaka upgrade script is essentially a stub the expectation as we complete the work for that is to add playbooks of tasks necessary to do the upgrade but it doesn't actually use tags. The kilo branch has stuff in scripts/upgrade-utilities/playbooks for this | 22:31 |
| *** thorst_ has quit IRC | 22:31 | |
| *** thorst_ has joined #openstack-ansible | 22:31 | |
| stelouie | stevelle: thanks I was trying to follow where that update script calls any playbooks, was only seeing scripts-library.sh and upgrading pip modules | 22:32 |
| stevelle | the flow is to update the roles and playbooks, then run the upgrade plays, which will also coordinate the execution of the various install playbooks | 22:32 |
| *** markvoelker_ has quit IRC | 22:32 | |
| stevelle | no tags are used in this model | 22:32 |
| stelouie | stevelle: I'm so curious about upgrades because it is something I am interested in for kubernetes/contrib/ansible and wanted to use openstack-ansible as a model since the project is large and follows best practices | 22:33 |
| errr | cloudnull: Ive noticed that wheel is already installed on cielometer container so no idea why its not working on this one | 22:33 |
| stevelle | the really tl;dr is 1) update the roles and playbooks, 2) do some things to upgrade the roles (changes to variables, introduction of new required variables, etc), 3) run the new version of the playbook | 22:34 |
| cloudnull | can you login to another container and install it there? | 22:34 |
| errr | cloudnull: yeah I did, and it worked. Im so confused why its failing in this one | 22:34 |
| cloudnull | was that container built by hand ? | 22:34 |
| cloudnull | maybe its not using amd64 as its base os? | 22:34 |
| errr | cloudnull: nope the playbooks built it | 22:34 |
| cloudnull | ok nvmd | 22:35 |
| errr | its using the same setting as the heat container | 22:35 |
| errr | I basically copied everything from heat to get this role going | 22:35 |
| stevelle | stelouie: the important part of an upgrade it laying down new bits and kicking the services. All the rest of the things to make that work become tasks in upgrade-specific playbooks. | 22:35 |
| cloudnull | destroy it and make it again ? | 22:35 |
| stevelle | because the bits and kicking services is part of install. | 22:36 |
| errr | yeah Ill do that | 22:36 |
| *** Ger-chervyak has joined #openstack-ansible | 22:36 | |
| errr | thanks for the help | 22:36 |
| cloudnull | sorry :\ | 22:36 |
| stelouie | is the accepted method to create upgrade playbooks that call out specific tasks in a role, or to use tags to pick and chose tasks | 22:36 |
| stevelle | stelouie: we create upgrade playbooks | 22:36 |
| *** Ger-cher_ has joined #openstack-ansible | 22:37 | |
| *** jayc has joined #openstack-ansible | 22:37 | |
| stevelle | stelouie: if the install playbooks are well formed, you don't need to worry about picking and choosing tasks from them, just run that after you do the upgrade-specific steps | 22:38 |
| stelouie | ^agreed, this is a part of the refactor I'm taking on with kubernetes/contrib/ansible. Right now the tasks are not granular enough that you could recycle, and I'm trying to take the structure here and apply it | 22:40 |
| *** Ger-chervyak has quit IRC | 22:41 | |
| *** Brew has quit IRC | 22:41 | |
| stelouie | I believe kubernetes/contrib will have to have an upgrade playbook in each role, that would wrap some of the tasks from main.yml | 22:42 |
| stevelle | stelouie: as we move from version to version, we don't carry our upgrades forward (juno to kilo playbooks are not in the mitaka branch, only kilo to mitaka will be). That was an intentional decision we set down to require one step at a time upgrades, because OpenStack. YMMV I suspect. | 22:44 |
| stevelle | stelouie: dig through our kilo branch upgrades a bit. We broke things out into very small pieces precisely because upgrades are so hard and a failure part way may need to be dealt with. That may translate for kube directly. | 22:45 |
| stelouie | stevelle: where can I find that work | 22:46 |
| stelouie | stevelle: just here, https://github.com/openstack/openstack-ansible/tree/kilo/scripts/upgrade-utilities | 22:46 |
| stevelle | stelouie: trace from https://github.com/openstack/openstack-ansible/blob/kilo/scripts/run-upgrade.sh | 22:46 |
| stevelle | upgrade utilities, what you linked, is the right spot for the deets | 22:47 |
| stelouie | this would be everything since the breakout of repos just happened for mitaka right? | 22:48 |
| stevelle | stelouie: exactly. | 22:49 |
| stelouie | stevelle: I'll dig around and probably be back when I have questions. thanks for all the help | 22:49 |
| stevelle | glhf | 22:49 |
| *** weezS has joined #openstack-ansible | 22:51 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-galera_client: Update Multi-distro support for CentOS7 https://review.openstack.org/314717 | 22:54 |
| *** stelouie has quit IRC | 23:02 | |
| *** Ger-cher_ has quit IRC | 23:03 | |
| *** weezS has quit IRC | 23:08 | |
| *** Mudpuppy has joined #openstack-ansible | 23:12 | |
| *** jamielennox is now known as jamielennox|away | 23:13 | |
| *** Ger-chervyak has joined #openstack-ansible | 23:16 | |
| Mudpuppy | Does anyone know offhand if python virtualenvs is for ansible 2.0+? | 23:17 |
| Mudpuppy | ie supported | 23:17 |
| *** saneax_AFK is now known as saneax | 23:20 | |
| *** jamielennox|away is now known as jamielennox | 23:21 | |
| stevelle | Mudpuppy: not sure abt question | 23:23 |
| Mudpuppy | rephrase: What version of ansible supports using python virtualenv to install pip packages. Is that a new 2.0 thing? or 1.8? | 23:24 |
| *** sdake has quit IRC | 23:30 | |
| stevelle | Mudpuppy: https://docs.ansible.com/ansible/pip_module.html claims 2.0 for first support | 23:30 |
| Mudpuppy | ty | 23:31 |
| stevelle | Mudpuppy: executable can be used to simulate if you create the venv separately | 23:31 |
| stevelle | p sure that is how we do it | 23:31 |
| Mudpuppy | I wanted to push the lab tools using their own environment, as I hit some snags with pip and cryptography library | 23:32 |
| Mudpuppy | being wonky | 23:32 |
| stevelle | Mudpuppy: I think there are lies involved in those docs | 23:32 |
| stevelle | https://github.com/openstack/openstack-ansible-os_ceilometer/blob/master/tasks/ceilometer_install.yml#L182 | 23:33 |
| stevelle | looks like there is virtualenv available in somewhere in the 1.5 - 1.9 gen | 23:33 |
| *** sacharya has quit IRC | 23:34 | |
| *** sacharya has joined #openstack-ansible | 23:35 | |
| Mudpuppy | ok. I'll play around with it | 23:37 |
| *** sdake has joined #openstack-ansible | 23:38 | |
| cloudnull | Mudpuppy: virtualenv was available in ansible 1.9 https://github.com/ansible/ansible-modules-core/blob/stable-1.9/packaging/language/pip.py#L49-L72 | 23:40 |
| Mudpuppy | cool | 23:41 |
| Mudpuppy | So confusing with all these docs | 23:41 |
| Mudpuppy | ty | 23:41 |
| cloudnull | loosk lke the same bits were in 1.8 too | 23:41 |
| cloudnull | https://github.com/ansible/ansible-modules-core/blob/release1.8.0/packaging/language/pip.py#L49-L72 | 23:41 |
| cloudnull | it would seem the docs are pointing folks to 2.x | 23:42 |
| stevelle | I seem to recall doing it with 1.7 but could be wrong | 23:42 |
| *** chandanc has joined #openstack-ansible | 23:42 | |
| *** chandanc_ has joined #openstack-ansible | 23:43 | |
| *** woodard_ has quit IRC | 23:49 | |
| *** woodard has joined #openstack-ansible | 23:49 | |
| *** quantumshadow1 has joined #openstack-ansible | 23:52 | |
| *** quantumshadow has quit IRC | 23:53 | |
| *** thorst_ has quit IRC | 23:57 | |
| *** thorst_ has joined #openstack-ansible | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!