| *** hwoarang has quit IRC | 00:03 | |
| *** hwoarang has joined #openstack-ansible | 00:06 | |
| *** markvoelker has joined #openstack-ansible | 00:09 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-ops master: Fix dashboards and possible port conflicts https://review.openstack.org/637082 | 00:15 |
|---|---|---|
| *** macza has joined #openstack-ansible | 00:27 | |
| *** macza has quit IRC | 00:32 | |
| *** gyee has quit IRC | 00:41 | |
| *** markvoelker has quit IRC | 00:42 | |
| prometheanfire | cloudnull: you running a 4.20 kernel? | 00:46 |
| prometheanfire | see if you can run a 'iptables-save -c' for me | 00:47 |
| prometheanfire | probably a gentoo regression, but maybe not | 00:47 |
| cloudnull | prometheanfire I am | 01:14 |
| * cloudnull needs to go get on my laptop | 01:14 | |
| prometheanfire | cloudnull: seems like a kconfig issue | 01:15 |
| *** ArchiFleKs has quit IRC | 01:35 | |
| *** zul has quit IRC | 01:37 | |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: Fix dashboards and possible port conflicts https://review.openstack.org/637082 | 01:38 |
| *** markvoelker has joined #openstack-ansible | 01:39 | |
| *** hwoarang has quit IRC | 01:43 | |
| *** hwoarang has joined #openstack-ansible | 01:44 | |
| *** ArchiFleKs has joined #openstack-ansible | 01:53 | |
| *** hwoarang has quit IRC | 01:54 | |
| *** hwoarang has joined #openstack-ansible | 01:55 | |
| *** markvoelker has quit IRC | 02:12 | |
| *** KeithMnemonic has quit IRC | 02:18 | |
| *** KeithMnemonic has joined #openstack-ansible | 02:19 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-ops master: Increment nginx check port https://review.openstack.org/637115 | 02:44 |
| *** partlycloudy has joined #openstack-ansible | 02:49 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_neutron master: Use the new services names for sfc https://review.openstack.org/622216 | 02:57 |
| openstackgerrit | James Denton proposed openstack/openstack-ansible-ops master: Correct formatting in README.rst https://review.openstack.org/637117 | 03:06 |
| *** markvoelker has joined #openstack-ansible | 03:09 | |
| partlycloudy | hello folks, i need a little help to follow through the OVS scenario documented at https://docs.openstack.org/openstack-ansible-os_neutron/latest/app-openvswitch.html. | 03:10 |
| jamesdenton | sure, whatcha got | 03:10 |
| partlycloudy | thanks james! i am confused about br-vlan and br-provider. is the br-provider in ovs the same as br-vlan in the linuxbridge version? | 03:11 |
| jamesdenton | ahh ok | 03:12 |
| jamesdenton | so, i agree that the docs need to be updated in that regard. It used to be that the architecture called for something like br-provider (ovs) -> br-vlan (lxb) -> eth1 | 03:13 |
| jamesdenton | BUT, now we can avoid br-vlan altogether | 03:14 |
| jamesdenton | so you can ignore that bit | 03:14 |
| jamesdenton | If you look midway down where it gives the example of br-provider and br-provider2. Those are OVS bridges that the playbooks would create on your behalf. and the 'network_interface' defined there would be an available interface on your system, like eth1 or bond1, that the playbooks would connect to the respective bridge on your behalf | 03:15 |
| *** priteau has joined #openstack-ansible | 03:16 | |
| partlycloudy | oh i see! Then, what is the relation between br-vxlan and br-tun? is it br-tun -> br-vxlan? | 03:16 |
| jamesdenton | well, not exactly. Neutron will create br-tun and connect it to br-int. Any 'tunnel' established between hosts will use the address configured on br-vxlan as the VTEP address | 03:17 |
| jamesdenton | so br-vxlan -> eth1.100 (dedicated vlan for overlay traffic) and the VTEP addr is configured on br-vxlan. When OVS passes vxlan traffic, it will traverse eth1.100 via br-vxlan (as thats where the IP is configured) | 03:18 |
| *** shyamb has joined #openstack-ansible | 03:19 | |
| jamesdenton | both br-vxlan and br-vlan bridges are important when neutron agents are deployed in containers, as the LXC containers connect to those bridges -> physical infrastructure. But as of Queens(?) the agents run on bare metal by default instead of LXC, so the bridges don't serve the same role | 03:20 |
| *** priteau has quit IRC | 03:20 | |
| partlycloudy | so for example, tunnel traffics would go from br-int -> br-tun -> br-vxlan (VTEP)? | 03:21 |
| jamesdenton | slowly trying to update the docs to reflect that. | 03:21 |
| jamesdenton | yes, but the connection from br-tun to br-vxlan is logical for our sake. OVS is using the IP configured on br-vxlan - it could just as easily be configured on eth1.100. BUT there is logic in the playbooks, i believe, that may expect it to be configured on br-vxlan so it's best to just leave it there | 03:23 |
| jamesdenton | This was recently added, and may help: https://docs.openstack.org/openstack-ansible/latest/user/network-arch/example.html | 03:24 |
| jamesdenton | doesn't call out OVS specifically, but gets close | 03:24 |
| partlycloudy | that was a great help! thank you james! | 03:26 |
| jamesdenton | Sure! Please let me know how the docs can be improved | 03:26 |
| partlycloudy | so with ovs, it is not necessarily to create br-vlan, if neutron agent is running on metal, right? | 03:27 |
| jamesdenton | that should be the case, yes | 03:27 |
| jamesdenton | What version are you deploying? | 03:28 |
| partlycloudy | i see. i was pulling my hairs off today… now i'm clear! :-) | 03:28 |
| *** shyam89 has joined #openstack-ansible | 03:28 | |
| jamesdenton | sorry bout that! | 03:28 |
| partlycloudy | tag 18.1.3 | 03:28 |
| jamesdenton | ok - just be aware that 'latest' docs usually reflect the master branch (v19 in this case) | 03:29 |
| jamesdenton | I cannot guarantee what's on that page will work w/o issue on v18 | 03:30 |
| jamesdenton | so the Rocky docs are here: https://docs.openstack.org/openstack-ansible-os_neutron/rocky/app-openvswitch.html | 03:30 |
| jamesdenton | but give the latest ones a go | 03:30 |
| jamesdenton | and if you do, be sure to add a 'container_interface: "<ethX>"' to the br-provider blocks that you create | 03:30 |
| jamesdenton | the eth interface you specify won't be used but will need to be there. eth100 is fine | 03:31 |
| *** shyamb has quit IRC | 03:31 | |
| partlycloudy | copy that! thank you so much! | 03:33 |
| jamesdenton | good luck! feel free to pop in - i'll be around tomorrow | 03:34 |
| partlycloudy | thank you James. you've been a great help! | 03:36 |
| *** markvoelker has quit IRC | 03:43 | |
| *** shyam89 has quit IRC | 03:47 | |
| *** dave-mccowan has quit IRC | 04:03 | |
| jamesdenton | cloudnull you around? | 04:11 |
| cjloader | https://review.openstack.org/#/c/636992/ | 04:27 |
| cjloader | can we get some eyes? cloudnull mnaser | 04:28 |
| cjloader | odyssey4me ^ | 04:28 |
| cjloader | GJ mgariepy for fixing the master bootstrap issue =) | 04:29 |
| cjloader | ty mnaser | 04:30 |
| cjloader | d34dh0r53 ^ | 04:38 |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: Increment nginx check port https://review.openstack.org/637115 | 04:39 |
| *** markvoelker has joined #openstack-ansible | 04:40 | |
| *** markvoelker has quit IRC | 05:12 | |
| *** pabelanger has quit IRC | 05:15 | |
| *** hwoarang has quit IRC | 05:21 | |
| *** hwoarang has joined #openstack-ansible | 05:26 | |
| *** shyam89 has joined #openstack-ansible | 05:30 | |
| *** priteau has joined #openstack-ansible | 05:39 | |
| *** priteau has quit IRC | 05:43 | |
| *** shyam89 has quit IRC | 06:06 | |
| *** shyam89 has joined #openstack-ansible | 06:07 | |
| *** markvoelker has joined #openstack-ansible | 06:09 | |
| *** markvoelker has quit IRC | 06:43 | |
| *** shyam89 has quit IRC | 06:49 | |
| *** shyam89 has joined #openstack-ansible | 06:55 | |
| *** kopecmartin|off is now known as kopecmartin | 07:15 | |
| *** shyam89 has quit IRC | 07:23 | |
| *** shyam89 has joined #openstack-ansible | 07:25 | |
| *** chkumar|pto is now known as chkumar|ruck | 07:29 | |
| *** markvoelker has joined #openstack-ansible | 07:40 | |
| fnpanic | hi | 07:44 |
| fnpanic | do i need to do something more then setting setup-ara true in bootstrap-ansible.sh to get it setup? | 07:47 |
| fnpanic | looks like ara-manage is not on the system | 07:47 |
| *** hamzaachi has joined #openstack-ansible | 07:51 | |
| *** shyam89 has quit IRC | 07:57 | |
| openstackgerrit | Chandan Kumar proposed openstack/openstack-ansible-os_nova master: Use venv_packages_to_symlink to symlink to import libvirt-python https://review.openstack.org/633474 | 08:01 |
| *** hamzaachi_ has joined #openstack-ansible | 08:04 | |
| *** hamzaachi has quit IRC | 08:05 | |
| *** KeithMnemonic has quit IRC | 08:08 | |
| *** KeithMnemonic has joined #openstack-ansible | 08:08 | |
| *** markvoelker has quit IRC | 08:12 | |
| *** rgogunskiy has joined #openstack-ansible | 08:16 | |
| *** tosky has joined #openstack-ansible | 08:38 | |
| *** shyam89 has joined #openstack-ansible | 08:44 | |
| *** aedc has joined #openstack-ansible | 08:59 | |
| *** DanyC has joined #openstack-ansible | 09:08 | |
| *** markvoelker has joined #openstack-ansible | 09:10 | |
| *** shardy has joined #openstack-ansible | 09:11 | |
| *** priteau has joined #openstack-ansible | 09:19 | |
| odyssey4me | fnpanic that will install it into the ansible runtime venv, not the system - so if it's there, it should be in /opt/ansible-runtime/bin | 09:25 |
| odyssey4me | you can activate the venv to use it more easily: source /opt/ansible-runtime/bin/activate; ara-manage ... | 09:25 |
| odyssey4me | jrosser so we have too many file watches in the repo container then? | 09:28 |
| odyssey4me | Can we overcome that, or do we need to look at removing pypiserver? | 09:29 |
| jrosser | odyssey4me: pypi-server seems to consume tons | 09:32 |
| *** DanyC has quit IRC | 09:32 | |
| jrosser | and i expect that as the number of upgrades you do builds up the contents of the repo server gets larger and larger (?) | 09:32 |
| *** DanyC has joined #openstack-ansible | 09:33 | |
| *** hamzaachi_ has quit IRC | 09:36 | |
| *** ArchiFleKs has quit IRC | 09:38 | |
| odyssey4me | jrosser probably, so now we have both lsync and pypiserver chewing up the watches :/ | 09:40 |
| jrosser | it seems surprisingly hard to find out what process has a watch on which files to understand what is happening | 09:40 |
| jrosser | but i expect this may explain a number of cases of failed lsync | 09:40 |
| odyssey4me | ok, I think most of the scaffolding is in place for us to back out of using pypiserver if we need to for rocky - that's what I'm planning anyway for master... however, that would be a bit of a distraction - can we work around it for now to buy me some time to figure it out? | 09:41 |
| odyssey4me | perhaps for now we can just do a known issue release note, explaining the issue and how to work around it | 09:41 |
| jrosser | just making the number of inotify large like my patch does should be totally aqequate for now | 09:41 |
| jrosser | *adequate | 09:41 |
| jrosser | it is a shame that the number of inotify is a system wide thing and exhasting them from one container will hose the whole system | 09:42 |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-os_gnocchi master: Convert systemd services to common role(s) https://review.openstack.org/637155 | 09:42 |
| *** markvoelker has quit IRC | 09:42 | |
| odyssey4me | mgariepy cjloader I've provided feedback in https://review.openstack.org/636992 - good attempt, but we can keep the host cleaner. | 09:43 |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-os_gnocchi master: Convert systemd services to common role(s) https://review.openstack.org/637155 | 09:43 |
| *** ArchiFleKs has joined #openstack-ansible | 09:48 | |
| jrosser | odyssey4me: another upgrade related thing i see in an environment that has gone from Q->R is that the repo build fails every time because there is still some remnant of molterniron hanging aroung which it tries to git clone, and fails | 09:50 |
| odyssey4me | jrosser hmm, that's odd - https://github.com/openstack/openstack-ansible/commit/e94c56ebe56559a2b2940cd2ef5378378905ac5e should be taking care of getting rid of the role, and it should no longer be in the py_pkgs lookup | 09:52 |
| jrosser | i'll circle back on that next week and get the actual error in a paste, i've not really dug into it yet | 09:52 |
| odyssey4me | jrosser could you type up a known issue for the issue with the repo container to accompany the patch you pushed up? | 09:55 |
| jrosser | ah right, yes and the ansible log should say if that role removal playbook got run | 09:55 |
| odyssey4me | a release note, I mean | 09:55 |
| jrosser | in a seperate patch, or to add to that? | 09:55 |
| odyssey4me | seperate patch is fine for master, given it's merged - but for the backports we should perhaps combine them | 09:55 |
| jrosser | ok sure i'll take a look later | 09:55 |
| fnpanic | odyssey4me: /opt/ansible-runtime/bin has no ara-manage | 10:13 |
| fnpanic | maybe i made a mistake | 10:13 |
| fnpanic | i did change this in the bootstrap script | 10:14 |
| fnpanic | export SETUP_ARA=${SETUP_ARA:-"True"} | 10:14 |
| *** hamzaachi has joined #openstack-ansible | 10:14 | |
| fnpanic | i guess this is wrong then | 10:14 |
| odyssey4me | fnpanic are you sure that it should - the ARA version being used may not have that command | 10:14 |
| fnpanic | oh | 10:14 |
| fnpanic | ok | 10:14 |
| *** shyam89 has quit IRC | 10:15 | |
| odyssey4me | fnpanic if you want to see whether it's installed, activate the venv and do 'pip freeze | grep ara' | 10:16 |
| CeeMac_ | morning | 10:20 |
| *** CeeMac_ is now known as CeeMac | 10:20 | |
| fnpanic | morning | 10:20 |
| fnpanic | nope | 10:21 |
| fnpanic | paramiko==2.4.1 | 10:21 |
| CeeMac | I ran into a small issue running the minor upgrade yesterday. Is there a specific requirement for which roles tp update in what order if I wanted to be more controlled about it? | 10:21 |
| *** DanyC has quit IRC | 10:23 | |
| *** DanyC has joined #openstack-ansible | 10:23 | |
| odyssey4me | fnpanic no ARA? if it's not there, then it seems there may be a bug in the implementation | 10:30 |
| *** priteau has quit IRC | 10:33 | |
| odyssey4me | fnpanic perhaps when antonym comes online he can share whether he's verified that it works | 10:33 |
| odyssey4me | it does appear to be working in our gate tests though | 10:33 |
| *** DanyC has quit IRC | 10:37 | |
| fnpanic | mhh ok so as i said it is most likely my mistake | 10:38 |
| *** markvoelker has joined #openstack-ansible | 10:39 | |
| *** DanyC has joined #openstack-ansible | 10:43 | |
| fnpanic | i wanted to setup octaivua | 10:46 |
| fnpanic | it added the provider network according to this doc | 10:46 |
| fnpanic | https://docs.openstack.org/openstack-ansible-os_octavia/rocky/configure-octavia.html#setup-a-neutron-network-for-use-by-octavia | 10:46 |
| fnpanic | it is named octavia like in the example | 10:46 |
| fnpanic | but afterwards it states in the docs: | 10:47 |
| fnpanic | There are a couple of variables which need to be adjusted if you don’t use lbaas for the provider network name and lbaas-mgmt for the neutron name. | 10:47 |
| fnpanic | so is just the example wrong? | 10:47 |
| fnpanic | also according to this doc octaiva is setup different | 10:47 |
| fnpanic | https://docs.openstack.org/openstack-ansible-os_neutron/rocky/configure-network-services.html#load-balancing-service-optional | 10:48 |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova master: Use venv_packages_to_symlink to symlink to import libvirt-python https://review.openstack.org/633474 | 10:56 |
| *** shyamb has joined #openstack-ansible | 11:06 | |
| *** DanyC has quit IRC | 11:07 | |
| *** shyamb has quit IRC | 11:12 | |
| *** markvoelker has quit IRC | 11:13 | |
| odyssey4me | fnpanic that would be for xgerman or johnsom to comment | 11:21 |
| *** shyamb has joined #openstack-ansible | 11:37 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_glance master: Remove glance_developer_mode and glance_venv_download https://review.openstack.org/620340 | 11:42 |
| phasespace | In an upgrade from pike to queens the hostname of the hypervisors are changed as to add '.openstack.local'. This breaks live migration for old instances. The node column of the instances table in the nova database will then include '.openstack.local' for new instances, but not old ones. When trying to live migrate old instances it fails because it apparently can't find the hypervisor named 'hostname' when it expects | 11:57 |
| phasespace | 'hostname.openstack.local'. | 11:57 |
| *** electrofelix has joined #openstack-ansible | 11:57 | |
| *** shyamb has quit IRC | 11:59 | |
| *** shyamb has joined #openstack-ansible | 11:59 | |
| *** hamzaachi has quit IRC | 11:59 | |
| *** hamzaachi has joined #openstack-ansible | 11:59 | |
| *** markvoelker has joined #openstack-ansible | 12:10 | |
| *** kmadac has joined #openstack-ansible | 12:12 | |
| odyssey4me | phasespace could you register a bug for that? that's not good | 12:27 |
| jamesdenton | mornin' | 12:29 |
| *** tosky has quit IRC | 12:30 | |
| chkumar|ruck | odyssey4me: is there a way to run this job locally http://logs.openstack.org/94/630694/12/check/openstack-ansible-deploy-aio_metal_heat-centos-7/062e8b9/ ? | 12:33 |
| phasespace | odyssey4me: will do | 12:33 |
| chkumar|ruck | odyssey4me: http://logs.openstack.org/94/630694/12/check/openstack-ansible-deploy-aio_metal_heat-centos-7/062e8b9/logs/openstack/aio1-utility/stestr_results.html and this job is creating tempest venv 18.0.0 any idea why? | 12:33 |
| odyssey4me | chkumar|ruck yes - clone openstack/openstack-ansible, then export SCENARIO=aio_metal_heat, then run through the basic AIO steps | 12:34 |
| *** nyloc has quit IRC | 12:34 | |
| odyssey4me | chkumar|ruck more details here: https://docs.openstack.org/openstack-ansible/latest/user/aio/quickstart.html | 12:34 |
| *** nyloc has joined #openstack-ansible | 12:35 | |
| *** shyamb has quit IRC | 12:35 | |
| *** shyamb has joined #openstack-ansible | 12:36 | |
| *** tosky has joined #openstack-ansible | 12:36 | |
| odyssey4me | chkumar|ruck the version there is probably because that was the last tag implemented on the master branch? the version now comes from https://github.com/openstack/openstack-ansible/commit/c44fbcbbe3b6d88e9393503e92ac670f9e811148 | 12:38 |
| phasespace | odyssey4me: https://bugs.launchpad.net/openstack-ansible/+bug/1816037 | 12:40 |
| openstack | Launchpad bug 1816037 in openstack-ansible "Pike->Queens upgrade adds ".openstack.local" to hypervisor hostname breaking live migration for old instances" [Undecided,New] | 12:40 |
| odyssey4me | phasespace thanks, hopefully someone will pick that up in the next triage session - or before | 12:42 |
| *** markvoelker has quit IRC | 12:42 | |
| odyssey4me | I'm surprised that hasn't come up in the work antonym has done - or maybe it has and it's not been resolved yet. | 12:42 |
| *** miloa has joined #openstack-ansible | 12:45 | |
| openstackgerrit | Marc Gariépy (mgariepy) proposed openstack/openstack-ansible master: Use ansible-runtime to run setup.py https://review.openstack.org/636992 | 12:50 |
| mgariepy | odyssey4me, ^^ | 12:50 |
| odyssey4me | mgariepy mind if I tweak the commit msg? | 12:52 |
| mgariepy | not at all :D | 12:52 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Use runtime venv to determine OSA version https://review.openstack.org/636992 | 12:55 |
| odyssey4me | mgariepy I think ^ will please jrosser more than the previous PS :) | 12:56 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Use runtime venv to determine OSA version https://review.openstack.org/636992 | 12:56 |
| odyssey4me | apologies for the last iteration - the extra whitespace was bugging me :p | 12:57 |
| odyssey4me | thanks mgariepy ! | 12:58 |
| *** shyamb has quit IRC | 13:03 | |
| mgariepy | thanks to you :D | 13:04 |
| mgariepy | whitespace are the worst. | 13:04 |
| mgariepy | but not as bad as tabs hha | 13:05 |
| openstackgerrit | Michael Vollman proposed openstack/openstack-ansible-os_neutron master: Avoid distro installing unused services https://review.openstack.org/633277 | 13:08 |
| *** shyamb has joined #openstack-ansible | 13:12 | |
| *** electrofelix has quit IRC | 13:14 | |
| *** flaviosr has quit IRC | 13:18 | |
| *** dave-mccowan has joined #openstack-ansible | 13:20 | |
| *** tosky has quit IRC | 13:25 | |
| *** tosky has joined #openstack-ansible | 13:26 | |
| *** flaviosr has joined #openstack-ansible | 13:27 | |
| *** markvoelker has joined #openstack-ansible | 13:39 | |
| *** phasespace has quit IRC | 13:44 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_neutron master: Set appropriate default for vxlan multicast ttl https://review.openstack.org/636676 | 13:56 |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-os_gnocchi master: Convert systemd services to common role(s) https://review.openstack.org/637155 | 13:58 |
| *** markvoelker has quit IRC | 14:12 | |
| *** kmadac has quit IRC | 14:18 | |
| openstackgerrit | Michael Vollman proposed openstack/openstack-ansible-os_neutron master: Avoid distro installing unused services https://review.openstack.org/633277 | 14:23 |
| *** shyamb has quit IRC | 14:28 | |
| *** shyamb has joined #openstack-ansible | 14:35 | |
| *** shyamb has quit IRC | 14:40 | |
| *** jorti has joined #openstack-ansible | 14:48 | |
| openstackgerrit | Florian Haas proposed openstack/openstack-ansible-os_cinder master: cinder.conf: add [nova] section, override interface defaults https://review.openstack.org/637198 | 14:56 |
| *** flaviosr has quit IRC | 15:00 | |
| *** sum12 has quit IRC | 15:00 | |
| *** flaviosr has joined #openstack-ansible | 15:01 | |
| *** itandops has joined #openstack-ansible | 15:04 | |
| *** jorti has quit IRC | 15:05 | |
| *** jorti has joined #openstack-ansible | 15:05 | |
| *** DanyC has joined #openstack-ansible | 15:07 | |
| *** markvoelker has joined #openstack-ansible | 15:10 | |
| *** sum12 has joined #openstack-ansible | 15:15 | |
| fnpanic | what is the plan for the heat ssl verify error | 15:15 |
| fnpanic | https://bugs.launchpad.net/openstack-ansible/+bug/1814909 | 15:15 |
| openstack | Launchpad bug 1814909 in openstack-ansible "Ssl verify error in heat -> keystone communication" [Undecided,New] | 15:15 |
| fnpanic | will the patch be merged ? | 15:15 |
| *** itandops has quit IRC | 15:15 | |
| jamesdenton | Is there an official patch for that yet? | 15:19 |
| openstackgerrit | Florian Haas proposed openstack/openstack-ansible-os_cinder master: Introduce cinder_scheduler_default_filters https://review.openstack.org/637201 | 15:19 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/ansible-role-python_venv_build master: Apply constraints when building the venv https://review.openstack.org/620331 | 15:23 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/ansible-role-python_venv_build master: Apply constraints when building the venv https://review.openstack.org/620331 | 15:30 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/ansible-role-python_venv_build master: Use the repo if available https://review.openstack.org/620339 | 15:33 |
| openstackgerrit | Matthew Thode proposed openstack/openstack-ansible-os_nova stable/rocky: Update deprecated option for pci passthrough https://review.openstack.org/636390 | 15:35 |
| openstackgerrit | Matthew Thode proposed openstack/openstack-ansible-os_nova stable/rocky: Update the pci config for nova. https://review.openstack.org/636391 | 15:35 |
| *** rgogunskiy has quit IRC | 15:36 | |
| *** chkumar|ruck is now known as chandankumar | 15:38 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_glance master: Update role for new source build process https://review.openstack.org/620340 | 15:39 |
| openstackgerrit | Matthew Thode proposed openstack/openstack-ansible-os_nova stable/queens: Update deprecated option for pci passthrough https://review.openstack.org/636396 | 15:40 |
| openstackgerrit | Matthew Thode proposed openstack/openstack-ansible-os_nova stable/queens: Update the pci config for nova. https://review.openstack.org/636397 | 15:40 |
| openstackgerrit | Matthew Thode proposed openstack/openstack-ansible-os_nova stable/pike: Update deprecated option for pci passthrough https://review.openstack.org/636398 | 15:42 |
| openstackgerrit | Matthew Thode proposed openstack/openstack-ansible-os_nova stable/pike: Update the pci config for nova. https://review.openstack.org/636399 | 15:42 |
| *** markvoelker has quit IRC | 15:42 | |
| openstackgerrit | Florian Haas proposed openstack/openstack-ansible-os_cinder master: Introduce cinder_scheduler_default_filters https://review.openstack.org/637201 | 15:43 |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-os_gnocchi master: Convert systemd services to common role(s) https://review.openstack.org/637155 | 15:48 |
| openstackgerrit | Matthew Thode proposed openstack/openstack-ansible-os_nova master: update pci_passthrough template variables https://review.openstack.org/637210 | 15:50 |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-os_gnocchi master: Convert systemd services to common role(s) https://review.openstack.org/637155 | 15:51 |
| *** gillesMo has joined #openstack-ansible | 15:56 | |
| *** francois has quit IRC | 16:17 | |
| antonym | odyssey4me: i think we worked around it internally in newton to just leave the hostname and drop the domainname but i think we have a bug internally for it for the queens upgrade... i'd guess the proper way would check for stand alone hostnames and then convert them to openstack.local if a custom domain wasn't set since greenfield assumes openstack.local is set or overridden | 16:26 |
| *** TxGirlGeek has joined #openstack-ansible | 16:28 | |
| *** macza has joined #openstack-ansible | 16:39 | |
| *** markvoelker has joined #openstack-ansible | 16:39 | |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-os_gnocchi master: Convert systemd services to common role(s) https://review.openstack.org/637155 | 16:40 |
| partlycloudy | jamesdenton: hello! are you there and have a moment? | 16:56 |
| *** gyee has joined #openstack-ansible | 17:01 | |
| *** phasespace has joined #openstack-ansible | 17:06 | |
| openstackgerrit | Merged openstack/openstack-ansible master: Use runtime venv to determine OSA version https://review.openstack.org/636992 | 17:07 |
| *** markvoelker has quit IRC | 17:13 | |
| jamesdenton | hi im here | 17:14 |
| partlycloudy | thank you james! still about OVS.. can i use a bridge instead of a nic or bond for the 'network_interface' for br-provider in openstack_user_variables.yml? | 17:18 |
| jamesdenton | what bridge did you have in mind? | 17:18 |
| jamesdenton | a linux bridge? | 17:18 |
| partlycloudy | i tried to use br-vlan (linuxbridge) that I created during provisioning. | 17:19 |
| guilhermesp | does someone know a good way to deal with this? http://paste.openstack.org/show/745173/ | 17:20 |
| guilhermesp | that's probably the reason for that bug https://bugs.launchpad.net/openstack-ansible/+bug/1815902 | 17:20 |
| openstack | Launchpad bug 1815902 in openstack-ansible "Error galera_server : Download extra packages" [Undecided,New] | 17:20 |
| jamesdenton | you should be able to use a bridge, however, i don't know that network_interface works in rocky. You can always try 'ovs-vsctl add-if br-provider br-vlan' | 17:20 |
| *** gillesMo has quit IRC | 17:21 | |
| *** TxGirlGeek has quit IRC | 17:21 | |
| partlycloudy | i tried to use br-vlan, but according to 'ovs-vsctl show', the br-provider only had two ports: br-provider and phy-br-provider | 17:21 |
| jamesdenton | try the ovs-vsctl command ^^^ | 17:22 |
| partlycloudy | ok. let me try it | 17:22 |
| evrardjp | odyssey4me: mnaser how is master nowadays? should I try a alpha release? We are not tied to anything anymore, and I like to either do that, or do a pbr bump | 17:22 |
| evrardjp | I would prefer releasing an alpha | 17:23 |
| evrardjp | mnaser: we should also discuss the releasing of ansible-config_template | 17:23 |
| partlycloudy | jamesdenton: ok. i used 'ovs-vsctl add-port br-provider br-vlan' and it worked. | 17:24 |
| jamesdenton | cool. | 17:25 |
| partlycloudy | doing it on all controller(network)nodes now.. see if i can get packets out after. | 17:26 |
| partlycloudy | jamesdenton: hmm… i added br-vlan to br-provider and restarted neutron on all controller(network) hosts. still i can't reach vms using floating ip. actually i can't even reach the router. | 17:33 |
| jamesdenton | can you paste the output of 'ovs-vsctl show' into pastebin? | 17:33 |
| partlycloudy | jamesdenton: https://pasted.tech/pastes/5ea040cfd60a767955d23be833697405097b9ce2 | 17:35 |
| partlycloudy | i found pasted.tech is nicer :-) | 17:35 |
| jamesdenton | :D | 17:36 |
| jamesdenton | and just to confirm, br-vlan is setup and UP with some physical interface? | 17:36 |
| *** miloa has quit IRC | 17:38 | |
| partlycloudy | jamesdenton: i see the non-ha router gateway interface is on infra1. so here is the 'ovs-vsctl show' from infra1: https://pasted.tech/pastes/dc125bb1a93f316818754ffd5974df43c9e3c7d0 | 17:39 |
| partlycloudy | yes, br-vlan is up and has eno2 enslaved to it. | 17:40 |
| jamesdenton | if you pop into the qrouter namespace on infra1, can it ping its gateway? | 17:40 |
| partlycloudy | let me try. ip netns exec ….. right? | 17:40 |
| jamesdenton | yep | 17:40 |
| partlycloudy | yes, i can. | 17:42 |
| jamesdenton | ok, so that tells me that br-vlan is indeed working for you | 17:42 |
| jamesdenton | at least on that infra | 17:42 |
| jamesdenton | and your tenant networks are vxlan? or vlan? | 17:42 |
| partlycloudy | using vxlan for overlay | 17:43 |
| *** TxGirlGeek has joined #openstack-ansible | 17:43 | |
| jamesdenton | ok - can you confirm in a vm console log if dhcp worked? | 17:45 |
| jamesdenton | also, did you update sec group rules to allow the connectivity? | 17:45 |
| partlycloudy | yes, i've added Allow all ICMP to sec group. | 17:46 |
| jamesdenton | can you confirm connectivity between infra1 and some compute via the br-vxlan IP? Just to make sure that works | 17:47 |
| jamesdenton | 172.29.240.11 to 172.29.240.12, for example | 17:48 |
| partlycloudy | from infra1 (172.29.240.11) i can reach all br-vxlan end points on other infras and computes | 17:50 |
| jamesdenton | kk | 17:50 |
| partlycloudy | let me try to recreate all networks, routers and vms. | 17:51 |
| jamesdenton | k | 17:52 |
| odyssey4me | evrardjp I dunno if you noticed, but https://review.openstack.org/636992 was needed to make the versioning work outside the gate. | 17:54 |
| odyssey4me | guilhermesp I think https://bugs.launchpad.net/openstack-ansible/+bug/1815902 was fixed - added a comment to the bug. | 17:56 |
| openstack | Launchpad bug 1815902 in openstack-ansible "Error galera_server : Download extra packages" [Undecided,New] | 17:56 |
| odyssey4me | evrardjp releasing an alpha/beta doesn't matter much to me - I guess it depends on precedence in openstack as a whole | 17:56 |
| odyssey4me | evrardjp but yeah, I think master is relatively releasable right now - although it'd need a role pin | 17:57 |
| partlycloudy | jamesdenton: after rebuild, both interfaces on the router remain down. | 18:02 |
| partlycloudy | from horizon interface status page it says "VIF Type binding failed" | 18:02 |
| jamesdenton | hmm, well there will be an error in the openvswitch-agent log, hopefully, that coincides with that | 18:03 |
| guilhermesp | odyssey4me: yeah, was fixed but raised another error. The fact that when it tries to get the url item ansible says it is undefined https://github.com/openstack/openstack-ansible-galera_server/blob/5e78067f409a438411c5c563811ac6edba358229/tasks/galera_install_download_extra_packages.yml#L25 | 18:03 |
| jamesdenton | check the log and see if there's anything interesting ther | 18:03 |
| guilhermesp | I was debugging locally and I saw that the values are lists and then ansible considers url as undefined variable | 18:04 |
| guilhermesp | that's a part of my debugging http://paste.openstack.org/show/745173/ | 18:04 |
| guilhermesp | item.value is a list, so I can't access the 'url' attribute directly. because we're looping over a dict at the higher level | 18:05 |
| jamesdenton | partlycloudy i need to grab a bite, but will be back in a few | 18:06 |
| partlycloudy | jamesdenton: yeah of course! take ur time. thanks so much. | 18:07 |
| odyssey4me | guilhermesp that's weird, because https://github.com/openstack/openstack-ansible-galera_server/blob/5e78067f409a438411c5c563811ac6edba358229/tasks/galera_install.yml#L34 | 18:09 |
| *** markvoelker has joined #openstack-ansible | 18:10 | |
| guilhermesp | yeah and then the task gets the whole dictionary https://github.com/openstack/openstack-ansible-galera_server/blob/5e78067f409a438411c5c563811ac6edba358229/tasks/galera_install_download_extra_packages.yml#L23 | 18:10 |
| guilhermesp | then item in this task is a list | 18:10 |
| guilhermesp | and boom, breaks with undefined variable | 18:11 |
| odyssey4me | guilhermesp Those tasks are meant to only execute for arm/PPC, so I think it's best to focus on what's going wrong there first | 18:12 |
| odyssey4me | actually, based on https://github.com/openstack/openstack-ansible-galera_server/blob/5e78067f409a438411c5c563811ac6edba358229/tasks/galera_install.yml#L34 - it should only execute on PPC hosts | 18:12 |
| *** hamzaachi has quit IRC | 18:15 | |
| *** hamzaachi has joined #openstack-ansible | 18:15 | |
| guilhermesp | yes odyssey4me I agree. If we look at this aspect, cool, it works. The thins is just the logic of "Download extra packages". I think the structure to iterate over the values needs to be refactored so we can get all the urls and checksums we need to install the PPC packages | 18:17 |
| guilhermesp | guys at #ansible are making some suggestions here. To create a list of those values to iterate using with_items | 18:18 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Include package pins for glance playbook https://review.openstack.org/637240 | 18:18 |
| prometheanfire | message-id CAJCXu8ep1ER3AH0S=ukxyqpi0=kkDPWstr_voLQ-Gn=v-z3LCQ@mail.gmail.com had a good comment about security updates for constraints on stable branches | 18:18 |
| prometheanfire | I'd like to see OSA's perspective | 18:18 |
| guilhermesp | as it runs only for PPC we don't need to worry about the keys of the dict | 18:18 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_glance master: Update role for new source build process https://review.openstack.org/620340 | 18:18 |
| odyssey4me | prometheanfire um, I think a link would be more useful than a mail ID | 18:19 |
| odyssey4me | guilhermesp maybe we just need to backstop it with a conditional of whether the item exists? although I'm curious why it even executes the task in the first place | 18:20 |
| *** DanyC has quit IRC | 18:20 | |
| prometheanfire | isn't online yet | 18:20 |
| prometheanfire | refresh :D http://lists.openstack.org/pipermail/openstack-discuss/2019-February/002802.html | 18:21 |
| prometheanfire | and fungi's response http://lists.openstack.org/pipermail/openstack-discuss/2019-February/002803.html | 18:21 |
| odyssey4me | logan- was http://logs.openstack.org/31/632131/2/check/openstack-ansible-functional-ubuntu-xenial/9addd27/job-output.txt.gz#_2019-01-22_14_37_40_350931 ever sorted out? | 18:21 |
| odyssey4me | prometheanfire if I read that right, requests had a CVE - from our standpoint, as long as upper-constraints updates the version to include the fix, we inherit it | 18:23 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/monitorstack master: Update monitorstack to use updated openstacksdk and es https://review.openstack.org/637242 | 18:23 |
| prometheanfire | odyssey4me: right, but what's being argued is that reqs should not be expected to update stable branches for security things | 18:24 |
| odyssey4me | prometheanfire well, I think that packagers and us all rely on that mechanism - however, if that's the stance we can work around it because we have our own override mechanisms... but those are meant to only be for temporary purposes | 18:25 |
| odyssey4me | I would hope that even for devstack tests the desire would be to test with the same thing that everyone is using | 18:26 |
| prometheanfire | ya, I kinda think we should backport stuff as we fix it but not be trusted to constantly watch for CVEs | 18:27 |
| openstackgerrit | Michael Vollman proposed openstack/openstack-ansible-os_manila master: Basic working os_manila role https://review.openstack.org/611930 | 18:27 |
| odyssey4me | prometheanfire I think 'best-effort' from a requirements standpoint is totally suitable | 18:27 |
| prometheanfire | fungi: ^ that sound good? | 18:28 |
| odyssey4me | vollman wrt the distro test - perhaps it'd be best to just get the source builds right, get that merged, then follow on with distro builds? | 18:28 |
| odyssey4me | vollman I'm concerned that the patch is growing to be even more difficult to review than it already is | 18:29 |
| prometheanfire | odyssey4me: also, update the reviews to add cherry-picked-from | 18:29 |
| vollman | odyssey4me: I'm happy to make the distro gate non-voting. source gate is working consistently | 18:29 |
| odyssey4me | vollman many of the roles don't even have the distro builds setup, so there's no requirement to make it work... although thanks to you for improving those builds! | 18:30 |
| guilhermesp | odyssey4me: I came up with a solution to access the values transforming the dicts values in a list and then iterate using with_items | 18:31 |
| guilhermesp | I can submit a PR so you can see how it goes | 18:31 |
| odyssey4me | guilhermesp okie dokey, thanks! | 18:31 |
| vollman | odyssey4me: np. thanks for all the reviews. I'll update the os_manila tests | 18:34 |
| odyssey4me | prometheanfire I'll respond to the email. | 18:36 |
| prometheanfire | odyssey4me: I just did as well | 18:37 |
| jamesdenton | odyssey4me https://github.com/logan2211/ansible-etcd/commit/fa1c447b6a979a614fc024725b5ecad215261c4a | 18:39 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/monitorstack master: Update monitorstack to use updated openstacksdk and es https://review.openstack.org/637242 | 18:40 |
| jamesdenton | partlycloudy any luck in the logs? | 18:41 |
| partlycloudy | thanks for coming back! the downed router interfaces have been solved. somehow the ovs daemon died on all 3 infra nodes. | 18:43 |
| *** markvoelker has quit IRC | 18:43 | |
| partlycloudy | now i can ping the router external gateway from internet. | 18:44 |
| jamesdenton | ok, so you're back where you started | 18:44 |
| partlycloudy | as you mentioned momentarily, dhcp may be the cause of the problem. | 18:46 |
| *** kopecmartin is now known as kopecmartin|off | 18:46 | |
| jamesdenton | well dhcp is a symptom usually | 18:47 |
| jamesdenton | or lack of dhcp, that is | 18:47 |
| partlycloudy | i checked the vm log and it contains 'checking http://169.254.169.254/2009-04-04/instance-id … failed 1/20: up 210.38. request failed' all the way to 20/20 | 18:47 |
| *** kmadac has joined #openstack-ansible | 18:47 | |
| jamesdenton | can you paste the entire log? | 18:47 |
| openstackgerrit | Michael Vollman proposed openstack/openstack-ansible-os_manila master: Basic working os_manila role https://review.openstack.org/611930 | 18:48 |
| partlycloudy | sure. and btw i may have some difficult time to explain to cloudnull as to why pasted.tech is down…. | 18:48 |
| jamesdenton | you broke it? | 18:48 |
| partlycloudy | i never abused it… only made a few pastes… | 18:48 |
| jamesdenton | lol | 18:49 |
| jamesdenton | he'll make it better. stronger. faster | 18:49 |
| partlycloudy | no doubt about it! | 18:50 |
| cloudnull | moved to a new clouds sever. one sec | 18:50 |
| partlycloudy | here is the full boot up log: http://paste.openstack.org/show/745180/ | 18:50 |
| partlycloudy | cloudnull: kudo! | 18:51 |
| *** kmadac has quit IRC | 18:51 | |
| openstackgerrit | Michael Vollman proposed openstack/openstack-ansible-os_manila master: Basic working os_manila role https://review.openstack.org/611930 | 18:51 |
| cloudnull | its back ! | 18:52 |
| cloudnull | sorry. seems my VMs were down, they've now been rebuilt | 18:52 |
| partlycloudy | high 5! | 18:52 |
| jamesdenton | partlycloudy Sending discover... repeats 3 times then moves on. so dhcp is not working. it is most likely related to vxlan | 18:53 |
| jamesdenton | do you have l2pop enabled? | 18:53 |
| jrosser | mutlicast...... | 18:53 |
| jrosser | depending on your switch, not having a querier running would do that | 18:54 |
| jrosser | nxos for sure | 18:54 |
| jamesdenton | and juniper, too | 18:54 |
| jamesdenton | enabled l2pop, if it isn't enabled, may be fastest path to success | 18:54 |
| partlycloudy | jamesdenton: l2_population=false | 18:54 |
| *** DanyC has joined #openstack-ansible | 18:55 | |
| jamesdenton | ok, so if you can, delete the routers and networks, then add an override for 'neutron_l2_population: true' and rerun the neutron playbooks. | 18:55 |
| partlycloudy | will do. i pretty much followed the sample configurations. does that mean there should be a line added into the user_variables.yml for OVS Scenario? | 18:57 |
| odyssey4me | ok folks, I'm out for the w/end - have a fantastic one! | 19:06 |
| jamesdenton | not necessarily, but the docs need to be updated to reflect multicast/non-multicast scenarios | 19:06 |
| jamesdenton | and caveats to each | 19:06 |
| *** TxGirlGeek has quit IRC | 19:07 | |
| *** TxGirlGe_ has joined #openstack-ansible | 19:07 | |
| openstackgerrit | Guilherme Steinmuller Pimentel proposed openstack/openstack-ansible-galera_server master: Iterate over list of values of PPC packages dict https://review.openstack.org/637275 | 19:10 |
| guilhermesp | odyssey4me: ^^ | 19:10 |
| *** shananigans has joined #openstack-ansible | 19:13 | |
| shananigans | I'm unfamiliar with the backport process. Do we need to do anything to get this guy back ported to rocky and queens? https://review.openstack.org/#/c/636071/ | 19:15 |
| *** colin- has left #openstack-ansible | 19:16 | |
| partlycloudy | jamesdenton: it's working now 8-) | 19:17 |
| jamesdenton | woot! | 19:17 |
| openstackgerrit | Shannon Mitchell proposed openstack/openstack-ansible-os_cinder stable/rocky: Add missing CLI_OPTIONS when setting up qos volume types https://review.openstack.org/637279 | 19:19 |
| partlycloudy | so, what caused this issue in the first place? something related to my switch? (i saw jrosser mentioned it, but i didn't get it) | 19:19 |
| openstackgerrit | Shannon Mitchell proposed openstack/openstack-ansible-os_cinder stable/queens: Add missing CLI_OPTIONS when setting up qos volume types https://review.openstack.org/637281 | 19:19 |
| fungi | prometheanfire: odyssey4me: the counterargument we get from distros packaging out of our stable branches is that things shouldn't be changing unless they directly relate to fixes in the software. this particular requests update may not require patches to stable branches depending on the library, but other similar situations the the future could and we need to be able to continue supporting the | 19:27 |
| fungi | distros who are backporting fixes to our dependencies and not break them too. also as noted our security focus is on fixing vulnerabilities within our source code so we shouldn't give downstream consumers the false impression we track vulnerabilities in our dependencies | 19:27 |
| prometheanfire | that false impression is the main issue I have with taking it | 19:28 |
| fungi | as for devstack testing "the same thing that everyone is using" sticking with the frozen version of requests is likely closer to testing against an lts distro's contemporary python-requests package with a backported fix than testing against the latest upstream release of requests is | 19:28 |
| jamesdenton | partlycloudy Depending on the switch, you have to configure IGMP snooping for multicast to work, or have a multicast router setup. L2 population relies on the neutron agent pre-populating the forwarding tables on all hosts. lot more overhead and places where things can go wrong | 19:31 |
| *** gyee has quit IRC | 19:32 | |
| openstackgerrit | Florian Haas proposed openstack/openstack-ansible-os_cinder master: Introduce cinder_scheduler_default_filters https://review.openstack.org/637201 | 19:34 |
| partlycloudy | jamesdenton: so, by default, l2population should be enabled and dhcp request from vms should be diverted directly to the dhcp agent, right? | 19:37 |
| partlycloudy | i need to go back and re-read your book now. :-) | 19:38 |
| jamesdenton | well, l2pop should be enabled/disabled based on the capabilities of a given environment (and maybe documented why you would choose one or the other). When enabled, the compute node knows exactly which host to forward traffic to, and vice-versa. When disabled, the hosts rely on multicast to populate their fdb | 19:39 |
| jamesdenton | hah, well i don't know how well i documented it there, either | 19:39 |
| *** markvoelker has joined #openstack-ansible | 19:41 | |
| *** shardy has quit IRC | 19:43 | |
| partlycloudy | i see. thanks so very much for helping and explaining everything. i am really appreciated! | 19:44 |
| *** gyee has joined #openstack-ansible | 19:46 | |
| openstackgerrit | Florian Haas proposed openstack/openstack-ansible-os_cinder master: Introduce cinder_scheduler_default_filters https://review.openstack.org/637201 | 19:58 |
| -openstackstatus- NOTICE: The StoryBoard service on storyboard.openstack.org is offline momentarily for maintenance: http://lists.openstack.org/pipermail/openstack-discuss/2019-February/002666.html | 20:02 | |
| *** TxGirlGe_ has quit IRC | 20:11 | |
| *** markvoelker has quit IRC | 20:13 | |
| *** TxGirlGeek has joined #openstack-ansible | 20:13 | |
| *** ArchiFleKs has quit IRC | 20:44 | |
| *** TxGirlGeek has quit IRC | 20:45 | |
| *** shananigans has quit IRC | 20:45 | |
| *** TxGirlGeek has joined #openstack-ansible | 20:45 | |
| *** shananigans has joined #openstack-ansible | 20:47 | |
| *** ArchiFleKs has joined #openstack-ansible | 21:06 | |
| *** markvoelker has joined #openstack-ansible | 21:10 | |
| openstackgerrit | Michael Vollman proposed openstack/openstack-ansible-os_manila master: Basic working os_manila role https://review.openstack.org/611930 | 21:31 |
| *** markvoelker has quit IRC | 21:43 | |
| *** jra has joined #openstack-ansible | 21:46 | |
| jra | so we're running a queens openstack deploy with OSA 17.0.8, all of which was originally installed on centos 7.5. Now 7.6 is out, and it's causing significant problems: there's no co 7.5 LXC base image available anymore, so I basically can't install a new infrastructure node unless I'm ready to move at least some of the deployment up to 7.6, which means rebuilding the repo | 21:48 |
| jra | How have other folks dealt with this? | 21:48 |
| *** markvoelker has joined #openstack-ansible | 22:40 | |
| *** shananigans has quit IRC | 22:41 | |
| *** aedc has quit IRC | 22:45 | |
| *** dave-mccowan has quit IRC | 22:45 | |
| *** marst has joined #openstack-ansible | 22:47 | |
| *** marst has quit IRC | 23:05 | |
| *** markvoelker has quit IRC | 23:13 | |
| openstackgerrit | Michael Vollman proposed openstack/openstack-ansible-os_manila master: Basic working os_manila role https://review.openstack.org/611930 | 23:22 |
| *** gyee has quit IRC | 23:50 | |
| *** hwoarang has quit IRC | 23:50 | |
| *** gyee has joined #openstack-ansible | 23:54 | |
| *** hwoarang has joined #openstack-ansible | 23:54 | |
| cloudnull | jra you have two options, 1 rerun the lxc-hosts playbook, 2 copy the lxc image cache to the new version name | 23:56 |
| *** DanyC has quit IRC | 23:56 | |
| cloudnull | the cache should in under /var/cache/lxc | 23:56 |
| *** DanyC has joined #openstack-ansible | 23:56 | |
| cloudnull | I guess the distro version according to ansible also includes the minor release on centos ? | 23:57 |
| *** hamzaachi has quit IRC | 23:57 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!