*** k_mouza has joined #openstack-kolla | 00:28 | |
*** k_mouza has quit IRC | 00:33 | |
*** kplant has quit IRC | 01:11 | |
*** altlogbot_0 has quit IRC | 01:28 | |
*** altlogbot_0 has joined #openstack-kolla | 01:29 | |
*** brett-soric has joined #openstack-kolla | 02:08 | |
*** BjoernT has joined #openstack-kolla | 02:33 | |
*** brett-soric has left #openstack-kolla | 02:54 | |
*** BjoernT_ has joined #openstack-kolla | 03:08 | |
*** BjoernT has quit IRC | 03:09 | |
*** skramaja has joined #openstack-kolla | 03:54 | |
*** factor has quit IRC | 04:13 | |
*** factor has joined #openstack-kolla | 04:14 | |
*** Sravan has joined #openstack-kolla | 04:19 | |
*** Sravan has quit IRC | 04:28 | |
*** Sravan has joined #openstack-kolla | 04:32 | |
*** Sravan has quit IRC | 04:37 | |
*** whoami-rajat has joined #openstack-kolla | 05:30 | |
*** BjoernT_ has quit IRC | 05:32 | |
*** Luzi has joined #openstack-kolla | 05:32 | |
*** Tengu has quit IRC | 05:38 | |
*** Tengu has joined #openstack-kolla | 05:43 | |
*** baha has quit IRC | 05:45 | |
*** shyamb has joined #openstack-kolla | 05:50 | |
yoctozepto | morning | 05:57 |
---|---|---|
*** pcaruana has joined #openstack-kolla | 06:12 | |
*** jbadiapa has joined #openstack-kolla | 06:19 | |
*** cah_link has joined #openstack-kolla | 06:21 | |
*** luksky11 has joined #openstack-kolla | 06:34 | |
*** shyamb has quit IRC | 06:39 | |
*** shyamb has joined #openstack-kolla | 06:46 | |
*** unicell has joined #openstack-kolla | 07:09 | |
*** ivve has joined #openstack-kolla | 07:13 | |
*** unicell has quit IRC | 07:25 | |
*** JangwonLee has joined #openstack-kolla | 07:32 | |
*** shyamb has quit IRC | 07:51 | |
mgoddard | morning | 07:56 |
*** Wasaac has joined #openstack-kolla | 07:59 | |
mgoddard | goldyfruit: is masakari working now? | 08:00 |
mnasiadka | morning | 08:00 |
mgoddard | morning mnasiadka | 08:01 |
mnasiadka | o/ mgoddard | 08:01 |
mgoddard | goldyfruit: I found this, may or may not be useful to you: https://review.opendev.org/#/c/561940/ | 08:02 |
*** rpittau|afk is now known as rpittau | 08:03 | |
*** Wasaac has quit IRC | 08:04 | |
*** Wasaac has joined #openstack-kolla | 08:13 | |
*** Wasaac has quit IRC | 08:14 | |
*** Wasaac has joined #openstack-kolla | 08:15 | |
yoctozepto | mgoddard, mnasiadka: any of you observed the ansible/jinja2 behavior that templating fails wherever there is an undefined variable used even if its only place is guarded by an 'if' which is false? | 08:15 |
*** cgrosjean has joined #openstack-kolla | 08:17 | |
mgoddard | yoctozepto: I don't think so, I thought it was lazy | 08:22 |
mgoddard | is this in the template module? | 08:22 |
yoctozepto | mgoddard: it was in our custom merge_configs, I replaced it with template for a moment but the result was the same | 08:23 |
mgoddard | hmm | 08:24 |
yoctozepto | "msg": "'ansible.vars.hostvars.HostVarsVars object' has no attribute u'ansible_nonexistent'" | 08:25 |
yoctozepto | 'if' works correctly | 08:25 |
yoctozepto | as in, after the offending variable is removed, the part is not in the resulting file | 08:25 |
yoctozepto | this is re: https://bugs.launchpad.net/kolla-ansible/+bug/1835713 | 08:27 |
openstack | Launchpad bug 1835713 in kolla-ansible train "migration_interface must exist for non-compute controllers" [Medium,In progress] - Assigned to Radosław Piliszek (yoctozepto) | 08:27 |
yoctozepto | the 'if' was missing but adding it by itself does not improve the situation | 08:27 |
yoctozepto | I ended up with coding relative 'include' support for our merge modules and using it there, will submit a change in a moment | 08:28 |
yoctozepto | so as long as you include it does not try to evaluate it for the failing cases | 08:30 |
mgoddard | yoctozepto: how about using hostvars[host].get('migration_interface')? | 08:31 |
yoctozepto | migration_interface_address: "{{ hostvars[inventory_hostname]['ansible_' + migration_interface]['ipv4']['address'] }}" | 08:41 |
yoctozepto | ^ this is the offending variable | 08:42 |
yoctozepto | not tried using get in this context but that would be cumbersome I believe | 08:42 |
yoctozepto | I was surprised jinja2 tries to evaluate a variable under false 'if' | 08:43 |
*** hamzaachi has joined #openstack-kolla | 08:46 | |
mgoddard | yoctozepto: we have a pattern of using migration_interface_address elsewhere, would that help? | 08:47 |
mgoddard | oops, I see it is used :) | 08:48 |
mgoddard | how about using the default filter? | 08:48 |
*** shyamb has joined #openstack-kolla | 08:49 | |
*** skramaja_ has joined #openstack-kolla | 09:04 | |
*** skramaja has quit IRC | 09:05 | |
*** priteau has joined #openstack-kolla | 09:08 | |
mgoddard | this one needs a review: https://review.opendev.org/#/c/669321/1 | 09:16 |
openstackgerrit | Merged openstack/kolla-ansible stable/stein: Specify endpoint when creating monasca user https://review.opendev.org/669327 | 09:21 |
openstackgerrit | Merged openstack/kolla-ansible master: Bump minimum Ansible version to 2.5 https://review.opendev.org/668387 | 09:21 |
openstackgerrit | Merged openstack/kolla-ansible master: Fixes for MariaDB bootstrap and recovery https://review.opendev.org/667904 | 09:21 |
openstackgerrit | Merged openstack/kolla-ansible stable/stein: Add upgrade-bifrost command https://review.opendev.org/667630 | 09:21 |
openstackgerrit | Merged openstack/kolla-ansible master: CI: Pull images before upgrade https://review.opendev.org/666881 | 09:21 |
openstackgerrit | Bharat Kunwar proposed x/kayobe-config master: Make source kayobe-env zsh compatible https://review.opendev.org/669364 | 09:22 |
openstackgerrit | Merged openstack/kolla-ansible master: CI: Test ironic also when nova role is modified https://review.opendev.org/669136 | 09:23 |
yoctozepto | <mgoddard> how about using the default filter? | 09:24 |
yoctozepto | wonder if that would not hide a real issue later | 09:24 |
yoctozepto | sorry for the delay, had some other tasks to do | 09:25 |
openstackgerrit | Mark Goddard proposed openstack/kolla-ansible stable/stein: Bump minimum Ansible version to 2.5 https://review.opendev.org/669627 | 09:31 |
mgoddard | yoctozepto: could do | 09:32 |
mgoddard | and this one please: https://review.opendev.org/#/c/669627/1 | 09:32 |
*** ohwhyosa has joined #openstack-kolla | 09:33 | |
openstackgerrit | Merged openstack/kolla-ansible stable/stein: CI: Keep stderr in ansible logs https://review.opendev.org/669086 | 09:34 |
yoctozepto | <mgoddard> yoctozepto: could do | 09:34 |
yoctozepto | hence proposing alternative solution | 09:34 |
yoctozepto | what a hectic day, can't sit down | 09:35 |
*** shyamb has quit IRC | 09:35 | |
*** shyamb has joined #openstack-kolla | 09:44 | |
openstackgerrit | Merged openstack/kolla-ansible master: CI: add periodic-stable-jobs Zuul project template https://review.opendev.org/667956 | 09:46 |
openstackgerrit | Merged openstack/kolla-ansible stable/stein: CI: Use template-overrides.j2 from kolla https://review.opendev.org/668983 | 09:47 |
openstackgerrit | Merged openstack/kolla stable/rocky: CI: Move to periodic-stable https://review.opendev.org/667962 | 09:47 |
mnasiadka | uhh, ok, time go back to ansible bump up playground I see | 09:57 |
mnasiadka | ;) | 09:57 |
openstackgerrit | Michal Nasiadka proposed openstack/kolla-ansible master: DNM: Troubleshoot ceph-nfs on ubuntu https://review.opendev.org/669315 | 10:08 |
*** luksky11 has quit IRC | 10:12 | |
openstackgerrit | Merged openstack/kolla-ansible stable/rocky: Specify endpoint when creating monasca user https://review.opendev.org/669328 | 10:17 |
ohwhyosa | morning everyone o/ | 10:21 |
*** vmixor has joined #openstack-kolla | 10:23 | |
openstackgerrit | Radosław Piliszek proposed openstack/kolla-ansible stable/stein: CI: add periodic-stable-jobs Zuul project template https://review.opendev.org/669630 | 10:24 |
vmixor | Hello, Iam | 10:26 |
vmixor | Hello, I am new to OpenStack Kolla developement and I have a few questions about kolla-ansible | 10:29 |
mgoddard | Hello vmixor, welcome | 10:29 |
vmixor | Hello, mgoddard | 10:30 |
vmixor | My actual task is to implement changes tu run OpenStack Kolla Stein on Debian Buster and push my changes to upstream if possible | 10:31 |
mgoddard | vmixor: do you work with hrw? | 10:34 |
vmixor | mgoddard: no | 10:35 |
mgoddard | vmixor: just wondering, he works with Debian | 10:36 |
mgoddard | vmixor: what questions do you have? | 10:36 |
brtknr | mgoddard: any idea why my `kayobe overcloud service reconfigure -kt monasca` fails with: TASK [public-openrc : Read the admin OpenStack API environment file] **************************************************************************************************************** | 10:38 |
brtknr | fatal: [localhost]: FAILED! => {"changed": false, "msg": "file not found: /opt/alaska/rundeck/alaska-prod/kayobe/src/kayobe-config/etc/kolla/admin-openrc.sh"} | 10:38 |
mgoddard | brtknr: yes, you need https://review.opendev.org/667451 | 10:39 |
openstackgerrit | Radosław Piliszek proposed openstack/kolla-ansible master: Do not require valid migration_interface for controllers https://review.opendev.org/669631 | 10:39 |
brtknr | mgoddard: I'm on rocky | 10:39 |
mgoddard | brtknr: cherry pick it | 10:40 |
openstackgerrit | Mark Goddard proposed openstack/kolla-ansible stable/stein: CI: Pull images before upgrade https://review.opendev.org/666953 | 10:40 |
openstackgerrit | Radosław Piliszek proposed openstack/kolla-ansible master: Do not require valid migration_interface for controllers https://review.opendev.org/669631 | 10:42 |
*** priteau has quit IRC | 10:42 | |
goldyfruit | mgoddard, masakari works for instances for now, I'm working on the hacluster role | 10:42 |
goldyfruit | So far I'm able to have corosync/pacemaker/pcs&crmsh into different containers | 10:43 |
goldyfruit | I working on the pacemaker-remote container now | 10:43 |
mgoddard | goldyfruit: did you see Sam P's comment on ML about running pacemaker on the host? | 10:45 |
mgoddard | goldyfruit: should we merge with just instance support? | 10:46 |
openstackgerrit | Merged x/kayobe-config master: Make source kayobe-env zsh compatible https://review.opendev.org/669364 | 10:46 |
goldyfruit | mgoddard, Yes I saw Sam comment but I think itś doable | 10:47 |
goldyfruit | Just want need some work | 10:48 |
mgoddard | goldyfruit: ok | 10:48 |
goldyfruit | mgoddard, yes so far we could just merge with instance support | 10:48 |
mgoddard | goldyfruit: is the patch with CI changes passing? | 10:48 |
goldyfruit | I will do an another review to support hostmonitor (and maybe process monitor) | 10:48 |
mgoddard | and should we add a new job for it? | 10:49 |
goldyfruit | this one https://review.opendev.org/#/c/616050 ? | 10:49 |
vmixor | mgoddard: Is change like this: http://paste.openstack.org/show/754117/ ok for community? | 10:50 |
brtknr | mgoddard: tyvm | 10:50 |
mgoddard | goldyfruit: yeah | 10:51 |
mgoddard | brtknr: yvw | 10:51 |
goldyfruit | mgoddard, I got +1 from Zull since I rebased/updated but I'm sure if this patch is very useful | 10:53 |
mgoddard | goldyfruit: yeah, just commenting - it doesn't do anything right now :) | 10:53 |
goldyfruit | To validate Masakari, we need to spawn an instance with HA_enable property and then kill the KVM process and check if Nova starts the VM back | 10:54 |
goldyfruit | mgoddard +! | 10:54 |
goldyfruit | +1 | 10:54 |
mgoddard | goldyfruit: that should be possible. We have test scripts in tests/test-*.sh | 10:54 |
mgoddard | vmixor: which branch are you on? check master - we already have it | 10:56 |
yoctozepto | mgoddard: https://review.opendev.org/669631 <- or would you prefer that split? | 10:56 |
mgoddard | vmixor: also stable/stein has it | 10:56 |
goldyfruit | mgoddard, I'll have a look | 10:57 |
mgoddard | yoctozepto: the split looks like it would work. I can see this getting out of hand though if we have many branches in config files and variables that could be undefined | 11:01 |
mgoddard | yoctozepto: is this something that has changed in a recent version of ansible or jinja? | 11:03 |
yoctozepto | mgoddard: checked with ansible 2.5 and 2.8 | 11:03 |
mgoddard | hmm | 11:03 |
yoctozepto | did not check jinja | 11:03 |
yoctozepto | in both cases it was the latest - 2.10.1 | 11:04 |
yoctozepto | seems they had a delay in releases | 11:04 |
yoctozepto | and it has a security fix for 2.10 released 2 years earlier | 11:04 |
yoctozepto | https://github.com/pallets/jinja/commit/a2a6c930bcca591a25d2b316fcfd2d6793897b26 | 11:06 |
yoctozepto | doubt it could be breaking | 11:06 |
*** hjensas is now known as hjensas|lunch | 11:06 | |
yoctozepto | it got me too, I thought the conditional was wrong but it was fine | 11:07 |
*** priteau has joined #openstack-kolla | 11:08 | |
*** luksky11 has joined #openstack-kolla | 11:11 | |
mgoddard | yoctozepto: https://github.com/markgoddard/ansible-experiments/tree/master/04-templates | 11:12 |
mgoddard | yoctozepto: simple case seems to work | 11:12 |
*** priteau has quit IRC | 11:16 | |
*** kplant has joined #openstack-kolla | 11:17 | |
openstackgerrit | Mark Goddard proposed openstack/kolla-ansible master: Test MariaDB operations in CI https://review.opendev.org/655663 | 11:18 |
yoctozepto | mgoddard: yeah, I knew it worked | 11:24 |
yoctozepto | now why not here | 11:24 |
mgoddard | hmm | 11:24 |
yoctozepto | I'll experiment | 11:24 |
mgoddard | would be nice to see a minimal example that doesn't work | 11:25 |
mgoddard | ansible would be simple, if it weren't for all the little oddities | 11:25 |
*** Wasaac has quit IRC | 11:27 | |
mgoddard | yoctozepto: isn't the problem just that we did not have {% if service_name == 'nova-compute' %} | 11:30 |
yoctozepto | mgoddard: nope, tested thrice ;p | 11:31 |
mgoddard | :( | 11:32 |
yoctozepto | trying for the smallest failure sample | 11:32 |
yoctozepto | ansible quirk hunters 8-) | 11:32 |
mnasiadka | so... the ubuntu ceph-nfs problem is some RPC problem... digging... | 11:34 |
yoctozepto | mnasiadka: missing rpcbind or dbus stuff? | 11:35 |
mnasiadka | yoctozepto: might be - http://logs.openstack.org/15/669315/2/check/kolla-ansible-ubuntu-source-ceph/3952232/primary/logs/kolla/ceph/ceph-nfs.txt.gz | 11:35 |
*** happyhemant has joined #openstack-kolla | 11:37 | |
mnasiadka | I'll try to force nfsv4, which shouldn't need rpcbind | 11:38 |
ohwhyosa | https://bugs.launchpad.net/kolla-ansible/+bug/1835735 --> When you have time (i see you are kinda busy) could someone take a look at this? | 11:39 |
openstack | Launchpad bug 1835735 in kolla-ansible "initrunonce fails with opendaylight" [Undecided,New] | 11:39 |
yoctozepto | ohwhyosa: yikes, horrifying | 11:42 |
yoctozepto | need someone knowledgeable to take a look | 11:42 |
ohwhyosa | yoctozepto: horrifying as a bug report (meaning, I should work on my bug reporting skills) or as a bug itself? | 11:42 |
ohwhyosa | or both xP | 11:43 |
openstackgerrit | Raimund Hook (stingrayza) proposed openstack/kolla-ansible stable/stein: fixed spelling of 'Bootstrapping' in tools/kolla-ansible https://review.opendev.org/669643 | 11:44 |
stingrayza | :) hoping to get :point_up: into stein release, but not major if it doesn't get there :) | 11:45 |
yoctozepto | ohwhyosa: horrifying error to get as it makes little sense to me atm | 11:47 |
*** rgogunskiy has joined #openstack-kolla | 11:47 | |
yoctozepto | stingrayza: hopefully we arrive at the point where all our changes will be like yours now | 11:48 |
ohwhyosa | yoctozepto: Yeah, I was a bit like... why doesn't the database exist? | 11:49 |
stingrayza | :D | 11:50 |
*** henriqueof has joined #openstack-kolla | 11:59 | |
*** sean-k-mooney has quit IRC | 12:03 | |
mgoddard | ohwhyosa: init-runonce is really just for testing. I don't think it's guaranteed to work in every environment | 12:04 |
mgoddard | ohwhyosa: looks like the neutron ODL database tables have not been configured. Normally this happens during initial deploy. There is a bug where it is not done on every deploy though, that could be the issue | 12:06 |
*** shyamb has quit IRC | 12:14 | |
*** shyamb has joined #openstack-kolla | 12:14 | |
*** sean-k-mooney has joined #openstack-kolla | 12:16 | |
ohwhyosa | mgoddard: Thanks! | 12:18 |
ohwhyosa | So it does work on some deploys on the same systems, or I have it guaranteed that it doesn't work on the same system? | 12:19 |
yoctozepto | https://github.com/markgoddard/ansible-experiments/pull/1 | 12:20 |
yoctozepto | mgoddard: ^ certainly you did not expect your repo to receive a pull request ;p | 12:20 |
yoctozepto | this loss of bug needed fixing asap | 12:21 |
*** goldyfruit has quit IRC | 12:22 | |
*** skramaja_ has quit IRC | 12:24 | |
mgoddard | yoctozepto: nice :) | 12:34 |
mgoddard | yoctozepto: so simple fix would be to use migration_interface rather than migration_interface_address? | 12:35 |
mgoddard | yoctozepto: or allow for undefined migration_interface and add a precheck to ensure it is defined where necessary | 12:40 |
*** Wasaac has joined #openstack-kolla | 12:40 | |
mgoddard | just thinking of things we could apply to other *_interface variables easily | 12:41 |
*** edleafe has joined #openstack-kolla | 12:42 | |
openstackgerrit | Mariusz Karpiarz proposed openstack/kolla-ansible master: Handle more return codes from nova-status upgrade check https://review.opendev.org/668177 | 12:43 |
ohwhyosa | Hmmm I happily destroyed and tried again with normal openvswitch (which AIO worked amazingly) but it does not create the bridges so creating an instance in nova fails miserably | 12:47 |
ohwhyosa | And then ovs-vsctl just hangs there | 12:47 |
ohwhyosa | Should my eno2 (the one I assign on the globals file to neutron external) be down or up during deployment? is it okay if it has an ipv6 but no ipv4? | 12:48 |
yoctozepto | <mgoddard> yoctozepto: so simple fix would be to use migration_interface rather than migration_interface_address? | 12:48 |
yoctozepto | ugly! | 12:48 |
yoctozepto | <mgoddard> yoctozepto: or allow for undefined migration_interface and add a precheck to ensure it is defined where necessary | 12:48 |
yoctozepto | the problem is they are all "defined" | 12:48 |
yoctozepto | only the indexing is really a failure | 12:48 |
mgoddard | yeah, but that failure makes it undefined | 12:49 |
yoctozepto | also, using |default does not help | 12:49 |
yoctozepto | maybe let's go with what I crafted so carefully and in the meantime report a bug to ansible? | 12:50 |
yoctozepto | seems odd enough to be a bug | 12:50 |
mgoddard | yoctozepto: I'm not against your approach as a one off, just concerned we'll hit the same elsewhere and we'll end up splitting lots of config files which will get messy | 12:50 |
yoctozepto | I share your concerns, hence upstreaming the problem to ansible might be worth it | 12:51 |
mgoddard | ohwhyosa: I think your neutron external interface should be up | 12:51 |
yoctozepto | anyways, nova.conf grew pretty much so it deserves splitting ;p | 12:51 |
mgoddard | yoctozepto: +1, would be nice to get a fix in ansible. Could take time though | 12:52 |
yoctozepto | mgoddard: then we might find a solution till then, no more similar issues so far so could be only this one since nova shares so much config | 12:53 |
yoctozepto | we usually bail out at the ansible level | 12:54 |
*** hjensas|lunch is now known as hjensas | 12:54 | |
ohwhyosa | mgoddard: good, it is up! I did a kolla-reconfigure after changing type=vxlan for type=flat, it managed to create and bring up the br-ext... but now it complains that br-int does not exist when creating an instance, though ovs-vsctl does show a br-int configured | 12:55 |
mgoddard | yoctozepto: I think tunnel_interface_address will have the same issue | 12:55 |
*** KeithMnemonic has joined #openstack-kolla | 12:57 | |
mgoddard | ohwhyosa: is this still dpdk? | 12:57 |
ohwhyosa | Nope, i tried to reduce complexity | 12:57 |
ohwhyosa | Pastebinning my globals.yml in a second | 12:58 |
ohwhyosa | mgoddard: https://pastebin.com/C5mqq2w3 | 13:00 |
ohwhyosa | I believe that is all, apart from the change i did to neutron plugin in /etc/kolla/neutron-openvswitch-agent/ml2_xxx.ini | 13:01 |
ohwhyosa | https://pastebin.com/XYbWf68H this is ovs-vsctl show | 13:03 |
mgoddard | ohwhyosa: looks like neutron OVS agent can't reach OVS for some reason. did you reconfigure neutron too? | 13:04 |
ohwhyosa | Hmmm nope, i think | 13:04 |
ohwhyosa | By the way, there is no way for the config to be all in /etc/kolla/{service} on deployment host, right? instead of it being on each node | 13:05 |
ohwhyosa | And also, for adding central_logging, kolla-ansible deploy or kolla-ansible reconfigure? silly question, i guess | 13:07 |
mgoddard | ohwhyosa: you only ever need to change config on localhost, kolla-ansible generates config on nodes | 13:07 |
mgoddard | deploy for adding a new service | 13:08 |
ohwhyosa | mgoddard: But which localhost? I have 1 deployment, 1 controller, 1 storage and 1 network-compute | 13:08 |
ohwhyosa | deploy doesn't have any service config (apart from globals) | 13:09 |
ohwhyosa | they are on each node (as in, kolla-controller has some in /etc/kolla/{service} , kolla-storage some others | 13:09 |
*** goldyfruit has joined #openstack-kolla | 13:10 | |
mgoddard | ohwhyosa: deployment | 13:10 |
ohwhyosa | that is the expected behaviour right? | 13:10 |
mgoddard | that's expected | 13:11 |
*** Wasaac has quit IRC | 13:11 | |
ohwhyosa | Oh, good, okay, I thought the config files would also be centralized on deployment for ease of access | 13:12 |
*** Wasaac has joined #openstack-kolla | 13:12 | |
ohwhyosa | Which doesn't mean it is not easy, still lovng kolla, in case it sounded like that | 13:13 |
*** Wasaac has quit IRC | 13:14 | |
openstackgerrit | Mariusz Karpiarz proposed openstack/kolla-ansible master: Handle more return codes from nova-status upgrade check https://review.opendev.org/668177 | 13:14 |
*** Wasaac has joined #openstack-kolla | 13:14 | |
*** henriqueof has quit IRC | 13:15 | |
*** gkadam has joined #openstack-kolla | 13:24 | |
*** gkadam has quit IRC | 13:26 | |
*** Wasaac has quit IRC | 13:29 | |
*** Wasaac has joined #openstack-kolla | 13:29 | |
mgoddard | ohwhyosa: when you say ease of access, do you mean for reading not writing? | 13:36 |
*** shyamb has quit IRC | 13:46 | |
ohwhyosa | Both | 13:47 |
*** Luzi has quit IRC | 13:48 | |
ohwhyosa | Meaning, if I have it all on my deployment node on /etc/kolla/{service}/{config} then it would be easy to just write and check config and then kolla-ansible reconfigure from the same host (the deployment one) | 13:48 |
ohwhyosa | Now i have to go to, say, neutron host, change the config there, then on the deployment host kolla-ansible -i multinode reconfigure | 13:50 |
*** BjoernT has joined #openstack-kolla | 13:52 | |
mgoddard | ohwhyosa: you shouldn't ever modify files under /etc/kolla/{service} | 13:57 |
mgoddard | ohwhyosa: changes will be overwritten | 13:57 |
mgoddard | if you need to modify config files, follow this: https://docs.openstack.org/kolla-ansible/latest/admin/advanced-configuration.html#openstack-service-configuration-in-kolla | 13:58 |
mgoddard | only needs changing on the deployment host | 13:58 |
ohwhyosa | mgoddard: Thanks, easier still then | 14:01 |
ohwhyosa | I read that but i got it all backwards | 14:01 |
*** priteau has joined #openstack-kolla | 14:02 | |
*** baha has joined #openstack-kolla | 14:02 | |
*** rgogunskiy has quit IRC | 14:04 | |
*** FlorianFa has quit IRC | 14:05 | |
ohwhyosa | Okay, some progress | 14:08 |
ohwhyosa | it seems ovs-vsctl now created the br-int and some others properly | 14:08 |
ohwhyosa | after restarting some containers in good order | 14:08 |
ohwhyosa | and there is one netns now | 14:08 |
ohwhyosa | but still neutron dhcp complains there is no br-int | 14:09 |
ohwhyosa | https://pastebin.com/L5n4MCRY | 14:09 |
ohwhyosa | I put there both the log and the ovs-vsctl | 14:09 |
*** dmellado has quit IRC | 14:09 | |
ohwhyosa | The former first, the latter at the bottom | 14:09 |
*** dmellado has joined #openstack-kolla | 14:11 | |
ohwhyosa | Well, now i'm getting another seemingly nasty error | 14:15 |
openstackgerrit | Pierre Riteau proposed x/kayobe master: Move paramiko cap to test-requirements.txt https://review.opendev.org/665093 | 14:15 |
ohwhyosa | But progress, that's something at least | 14:15 |
*** dpawlik has quit IRC | 14:24 | |
*** dmellado has quit IRC | 14:30 | |
*** k_mouza has joined #openstack-kolla | 14:31 | |
*** dmellado has joined #openstack-kolla | 14:32 | |
yoctozepto | <mgoddard> yoctozepto: I think tunnel_interface_address will have the same issue | 14:44 |
yoctozepto | mhm | 14:44 |
yoctozepto | could be | 14:44 |
yoctozepto | nope: | 14:45 |
yoctozepto | services_need_ml2_conf_ini: | 14:45 |
yoctozepto | - "neutron-linuxbridge-agent" | 14:45 |
yoctozepto | - "neutron-openvswitch-agent" | 14:45 |
yoctozepto | - "neutron-infoblox-ipam-agent" | 14:45 |
yoctozepto | - "neutron-server" | 14:45 |
yoctozepto | and they all need tunnel_interface anyway | 14:45 |
yoctozepto | unless I'm missing something? | 14:46 |
yoctozepto | though precheck for tunnel interface is missing | 14:46 |
*** k_mouza has quit IRC | 14:47 | |
*** gouthamr has quit IRC | 14:47 | |
yoctozepto | it was not missing for migration though | 14:47 |
yoctozepto | ;p | 14:47 |
mnasiadka | so, in ganesha 2.6.0 (supplied by Ubuntu with Ceph Mimic) - forget about not running rpcbind or similar | 14:49 |
*** dmellado has quit IRC | 14:49 | |
*** cah_link has quit IRC | 14:49 | |
yoctozepto | mnasiadka: hmm? why so? | 14:49 |
mnasiadka | yoctozepto: because the code to support nfsv4 only without checking for rpcbind is in 2.7.x | 14:50 |
mgoddard | yoctozepto: just finishing off the last every kayobe meeting :) | 14:50 |
yoctozepto | mnasiadka: :-( | 14:50 |
*** cah_link has joined #openstack-kolla | 14:50 | |
yoctozepto | mgoddard: cool, I'm on the move so will be back in a bit again | 14:51 |
yoctozepto | I made a todo note to myself to report the bug to ansible | 14:52 |
yoctozepto | maybe they could take care of it | 14:52 |
openstackgerrit | Michal Nasiadka proposed openstack/kolla-ansible master: DNM: Troubleshoot ceph-nfs on ubuntu https://review.opendev.org/669315 | 14:53 |
*** dmellado has joined #openstack-kolla | 14:53 | |
mnasiadka | yoctozepto: usually it means they will take care of it in 2.11 :) | 14:53 |
mnasiadka | (if you report the bug today) | 14:53 |
*** gouthamr has joined #openstack-kolla | 14:56 | |
*** jovial[m] has joined #openstack-kolla | 14:57 | |
openstackgerrit | Merged openstack/kolla-ansible stable/stein: Bump minimum Ansible version to 2.5 https://review.opendev.org/669627 | 14:59 |
openstackgerrit | Merged openstack/kolla-ansible stable/stein: Wait for all compute services before cell discovery https://review.opendev.org/669321 | 14:59 |
openstackgerrit | Mark Goddard proposed x/kayobe master: Use new kolla-ansible upgrade-bifrost command https://review.opendev.org/667632 | 15:03 |
openstackgerrit | Mark Goddard proposed x/kayobe master: Revert "Workaround: stop bifrost services before seed upgrade" https://review.opendev.org/668888 | 15:04 |
mgoddard | yoctozepto: strictly I don't think neutron-server needs data plane access | 15:07 |
mgoddard | not sure about infoblox | 15:08 |
openstackgerrit | Merged openstack/kolla-ansible stable/stein: fixed spelling of 'Bootstrapping' in tools/kolla-ansible https://review.opendev.org/669643 | 15:10 |
*** BjoernT_ has joined #openstack-kolla | 15:11 | |
*** k_mouza has joined #openstack-kolla | 15:14 | |
*** BjoernT has quit IRC | 15:14 | |
*** ivve has quit IRC | 15:16 | |
*** Wasaac has quit IRC | 15:18 | |
*** Wasaac_ has joined #openstack-kolla | 15:18 | |
openstackgerrit | Mark Goddard proposed openstack/kolla-ansible stable/rocky: Wait for all compute services before cell discovery https://review.opendev.org/669698 | 15:19 |
openstackgerrit | Mark Goddard proposed openstack/kolla-ansible stable/queens: Wait for all compute services before cell discovery https://review.opendev.org/669700 | 15:20 |
*** Wasaac_ has quit IRC | 15:24 | |
*** Wasaac has joined #openstack-kolla | 15:25 | |
*** Wasaac has quit IRC | 15:26 | |
*** Wasaac has joined #openstack-kolla | 15:27 | |
openstackgerrit | Mark Goddard proposed openstack/kolla-ansible stable/rocky: Fixes for MariaDB bootstrap and recovery https://review.opendev.org/669701 | 15:28 |
*** Wasaac has quit IRC | 15:29 | |
*** Wasaac has joined #openstack-kolla | 15:30 | |
*** henriqueof has joined #openstack-kolla | 15:30 | |
*** Wasaac has quit IRC | 15:32 | |
openstackgerrit | Mark Goddard proposed openstack/kolla-ansible stable/queens: Fixes for MariaDB bootstrap and recovery https://review.opendev.org/669703 | 15:32 |
*** Wasaac has joined #openstack-kolla | 15:32 | |
*** Wasaac has quit IRC | 15:34 | |
*** Wasaac has joined #openstack-kolla | 15:35 | |
mgoddard | Proposed patch to release kolla and kolla-ansible 8.0.0: 8a09547e270e716495183600b5d1679835779247 | 15:37 |
mgoddard | https://review.opendev.org/669707 | 15:37 |
yoctozepto | mgoddard: +1 | 15:40 |
yoctozepto | <mgoddard> yoctozepto: strictly I don't think neutron-server needs data plane access | 15:41 |
yoctozepto | right-yo, it's neutron-api if it followed nova naming | 15:41 |
mgoddard | yeah | 15:41 |
yoctozepto | ok, checked the inventory, it would be separate and fail if I ran network nodes separate from control nodes | 15:43 |
ohwhyosa | hmmm is it still current, with kibana, adding the log-* index? | 15:43 |
ohwhyosa | It tells me | 15:43 |
yoctozepto | {% if inventory_hostname in groups["network"] or inventory_hostname in groups["compute"] %} | 15:43 |
yoctozepto | local_ip = {{ tunnel_interface_address }} | 15:43 |
yoctozepto | {% endif %} | 15:43 |
ohwhyosa | Unable to fetch mapping. Do you have indices matching the pattern? | 15:43 |
yoctozepto | we guard it but know we know it will fail anyways | 15:43 |
yoctozepto | now* we know | 15:44 |
yoctozepto | infoblox should have tunnel interface by the way it looks though | 15:45 |
yoctozepto | so it's neutron-server to protect here | 15:45 |
yoctozepto | and also add a precheck | 15:45 |
yoctozepto | because there is none | 15:45 |
yoctozepto | ;p | 15:45 |
vmixor | mgoddard: my mistake, I was using my branch created from master 2 or 3 weeks ago | 15:57 |
ohwhyosa | okay, it is now flog, i have it now | 15:57 |
mgoddard | vmixor: no problem | 16:02 |
*** hamzaachi has quit IRC | 16:02 | |
openstackgerrit | Gaëtan Trellu proposed openstack/kolla-ansible master: Testing Masakari role in gate https://review.opendev.org/616050 | 16:03 |
*** k_mouza has quit IRC | 16:07 | |
*** whoami-rajat has quit IRC | 16:10 | |
*** BjoernT has joined #openstack-kolla | 16:11 | |
*** hamzaachi has joined #openstack-kolla | 16:14 | |
openstackgerrit | Merged x/kayobe master: Support customising Prometheus config https://review.opendev.org/655669 | 16:14 |
*** whoami-rajat has joined #openstack-kolla | 16:15 | |
*** BjoernT_ has quit IRC | 16:15 | |
openstackgerrit | Merged x/kayobe master: Support customising refactored HAProxy config https://review.opendev.org/668547 | 16:17 |
yoctozepto | a few more deliverables and we become a lively channel | 16:18 |
*** k_mouza has joined #openstack-kolla | 16:19 | |
openstackgerrit | Merged x/kayobe master: Fix custom ironic python agent environmental variables https://review.opendev.org/667422 | 16:24 |
*** Wasaac has quit IRC | 16:26 | |
*** Sravan has joined #openstack-kolla | 16:27 | |
*** Sravan has quit IRC | 16:32 | |
openstackgerrit | Mark Goddard proposed x/kayobe stable/rocky: Fix custom ironic python agent environmental variables https://review.opendev.org/669711 | 16:32 |
openstackgerrit | Mark Goddard proposed x/kayobe stable/queens: Fix custom ironic python agent environmental variables https://review.opendev.org/669712 | 16:32 |
openstackgerrit | Mark Goddard proposed x/kayobe stable/rocky: Add retries to ansible galaxy install for all envs https://review.opendev.org/669713 | 16:33 |
openstackgerrit | Mark Goddard proposed x/kayobe stable/rocky: Add config tags to kolla-openstack play https://review.opendev.org/669714 | 16:33 |
openstackgerrit | Mark Goddard proposed x/kayobe stable/queens: Add config tags to kolla-openstack play https://review.opendev.org/669715 | 16:33 |
*** igordc has joined #openstack-kolla | 16:37 | |
*** Sravan has joined #openstack-kolla | 16:39 | |
vmixor | mgoddard: Is there mechanism to override default Debian sources.list and/or to add backports? Buster release contains packages for Rocky by default. | 16:40 |
mgoddard | vmixor: stein containers should have stein packages | 16:41 |
mgoddard | vmixor: although hrw knows most about debian but he is out this week | 16:43 |
mgoddard | vmixor: you can override the block base_ubuntu_package_sources_list if you want to provide your own sources.list though | 16:44 |
mgoddard | vmixor: https://docs.openstack.org/kolla/latest/admin/index.html#building-container-images | 16:44 |
*** k_mouza has quit IRC | 16:46 | |
*** ivve has joined #openstack-kolla | 16:48 | |
*** rpittau is now known as rpittau|afk | 16:50 | |
*** cah_link has quit IRC | 16:50 | |
*** luksky11 has quit IRC | 16:51 | |
openstackgerrit | Pierre Riteau proposed x/kayobe master: Add support for overriding OpenStack configuration globally https://review.opendev.org/665721 | 16:51 |
*** cah_link has joined #openstack-kolla | 16:52 | |
openstackgerrit | Merged openstack/kolla master: Add Python 3 Train unit tests https://review.opendev.org/669375 | 17:00 |
*** cah_link has quit IRC | 17:02 | |
*** Sravan has quit IRC | 17:05 | |
*** k_mouza has joined #openstack-kolla | 17:06 | |
*** k_mouza has quit IRC | 17:06 | |
*** dasp has quit IRC | 17:06 | |
vmixor | mgodard: thank you | 17:07 |
*** igordc has quit IRC | 17:10 | |
*** henriqueof has quit IRC | 17:12 | |
*** Sravan has joined #openstack-kolla | 17:17 | |
*** priteau has quit IRC | 17:21 | |
openstackgerrit | Merged openstack/kolla-ansible master: Add Python 3 Train unit tests https://review.opendev.org/669376 | 17:24 |
*** jonaspaulo has joined #openstack-kolla | 17:25 | |
*** hamzaachi has quit IRC | 17:27 | |
*** henriqueof has joined #openstack-kolla | 17:32 | |
*** unicell has joined #openstack-kolla | 17:41 | |
*** igordc has joined #openstack-kolla | 17:41 | |
*** dasp has joined #openstack-kolla | 17:42 | |
*** igordc has quit IRC | 17:43 | |
*** igordc has joined #openstack-kolla | 17:46 | |
yoctozepto | mgoddard: reported ~> https://github.com/ansible/ansible/issues/58835 | 17:48 |
mgoddard | yoctozepto: nice! | 17:48 |
mgoddard | yoctozepto: now for another weird ansible issue, our discovery wait is broken in ansible <2.8 | 17:48 |
yoctozepto | lolz! | 17:49 |
yoctozepto | how so? | 17:49 |
yoctozepto | mgoddard: need failure example | 17:51 |
mgoddard | yoctozepto: https://github.com/markgoddard/ansible-experiments/tree/master/05-referencing-registered-var-do-until | 17:52 |
mgoddard | will push a fix | 17:52 |
mgoddard | will pause the release :( | 17:52 |
yoctozepto | mgoddard: :-( | 17:53 |
*** cgrosjean has quit IRC | 17:57 | |
yoctozepto | mgoddard: maybe we should seriously consider supporting only the latest ansible - setting up a virtualenv for users is a blink of an eye and supporting so many ansible releases makes it impractical | 17:58 |
kplant | yeah virtualenv is magic | 17:59 |
mgoddard | yoctozepto: perhaps. if we add ansible to requirements.txt we could control it | 17:59 |
kplant | openstack-ansible doesn't even give the end-user a choice | 17:59 |
yoctozepto | mgoddard: the lol thing about it is that my test plugin I abandoned would avoid this issue, life is brutal ;p | 17:59 |
kplant | i think that's a pretty good approach imo | 17:59 |
mgoddard | still, the version will be different for different releases | 18:00 |
yoctozepto | kplant: it's because ansible is pretty unstable stills | 18:00 |
mgoddard | but this is what we do with kayobe | 18:00 |
yoctozepto | mgoddard: yeah but we could then pin one and test one | 18:00 |
kplant | yoctozepto: yeah but forcing the version and accompanying requirements would help with that | 18:00 |
mgoddard | yoctozepto: yeah | 18:00 |
mgoddard | we're stuck for stein and earlier though | 18:00 |
mgoddard | testing min and max will help a lot | 18:00 |
yoctozepto | mgoddard: yeah, unfortunately | 18:00 |
kplant | not hoping joe operator is going to pip install ansible==2.7.1.3.5.2.rc15 | 18:01 |
yoctozepto | +1 | 18:01 |
mgoddard | would have caught this | 18:01 |
yoctozepto | yeah, test | 18:01 |
yoctozepto | but also think it over for future | 18:01 |
yoctozepto | releases | 18:01 |
yoctozepto | mgoddard: I think the best approach now is to make it uglier by throwing the expansion in 'util' | 18:02 |
yoctozepto | and be happy with the new knowledge | 18:02 |
mgoddard | yoctozepto: yeah, working on it | 18:02 |
yoctozepto | mgoddard: and remember the <2.8 compat comment | 18:03 |
yoctozepto | it seriously makes reading friendlier ;D | 18:03 |
*** BjoernT_ has joined #openstack-kolla | 18:05 | |
*** BjoernT has quit IRC | 18:06 | |
openstackgerrit | Mark Goddard proposed openstack/kolla-ansible master: Fix nova deploy with Ansible<2.8 https://review.opendev.org/669730 | 18:08 |
mgoddard | yoctozepto: seems we also didn't deprecate ceph: https://review.opendev.org/#/c/669214/ | 18:09 |
openstackgerrit | Mark Goddard proposed openstack/kolla-ansible master: Test minimum supported and latest versions of Ansible https://review.opendev.org/668413 | 18:09 |
*** cgrosjean has joined #openstack-kolla | 18:19 | |
*** whoami-rajat has quit IRC | 18:20 | |
*** priteau has joined #openstack-kolla | 18:22 | |
yoctozepto | mgoddard: I deprecate | 18:25 |
yoctozepto | as soon as it merges there is a backport to stein | 18:25 |
yoctozepto | no kidding | 18:25 |
*** hamzaachi has joined #openstack-kolla | 18:34 | |
*** cgrosjea_ has joined #openstack-kolla | 18:52 | |
*** cgrosjea_ has quit IRC | 18:52 | |
*** cgrosjean has quit IRC | 18:55 | |
*** whoami-rajat has joined #openstack-kolla | 18:57 | |
*** BjoernT_ has quit IRC | 19:05 | |
*** cloudnull is now known as cloudkiller | 19:25 | |
*** BjoernT has joined #openstack-kolla | 19:28 | |
*** factor has quit IRC | 19:30 | |
*** factor has joined #openstack-kolla | 19:30 | |
*** Sravan has quit IRC | 19:35 | |
*** BjoernT_ has joined #openstack-kolla | 19:41 | |
*** unicell has quit IRC | 19:44 | |
*** BjoernT has quit IRC | 19:44 | |
*** icarusfactor has joined #openstack-kolla | 19:45 | |
*** factor has quit IRC | 19:47 | |
openstackgerrit | Mark Goddard proposed openstack/kolla-ansible master: Test minimum supported and latest versions of Ansible https://review.opendev.org/668413 | 19:57 |
openstackgerrit | Mark Goddard proposed openstack/kolla-ansible master: Fix nova deploy with Ansible<2.8 https://review.opendev.org/669730 | 19:58 |
openstackgerrit | Mark Goddard proposed openstack/kolla-ansible master: Test minimum supported and latest versions of Ansible https://review.opendev.org/668413 | 19:59 |
*** factor has joined #openstack-kolla | 20:01 | |
*** icarusfactor has quit IRC | 20:03 | |
*** kplant has quit IRC | 20:05 | |
*** luksky11 has joined #openstack-kolla | 20:25 | |
*** icarusfactor has joined #openstack-kolla | 20:25 | |
*** factor has quit IRC | 20:27 | |
*** priteau has quit IRC | 20:44 | |
*** priteau has joined #openstack-kolla | 20:49 | |
*** pcaruana has quit IRC | 20:50 | |
*** whoami-rajat has quit IRC | 21:00 | |
*** Sravan has joined #openstack-kolla | 21:06 | |
*** BjoernT_ has quit IRC | 21:07 | |
*** goldyfruit has quit IRC | 21:14 | |
*** priteau has quit IRC | 21:27 | |
*** priteau has joined #openstack-kolla | 21:27 | |
*** cloudkiller is now known as cloudnull | 21:27 | |
*** kplant has joined #openstack-kolla | 21:28 | |
*** priteau has quit IRC | 21:29 | |
*** Sravan has quit IRC | 21:31 | |
*** henriqueof has quit IRC | 21:35 | |
*** Sravan has joined #openstack-kolla | 21:36 | |
*** ivve has quit IRC | 21:39 | |
*** hamzaachi has quit IRC | 21:41 | |
*** Sravan has quit IRC | 21:41 | |
*** hamzaachi has joined #openstack-kolla | 21:42 | |
*** Sravan has joined #openstack-kolla | 21:45 | |
*** hamzaachi has quit IRC | 21:49 | |
*** Sravan has quit IRC | 21:50 | |
*** Sravan has joined #openstack-kolla | 21:51 | |
*** cgrosjean has joined #openstack-kolla | 22:01 | |
*** Sravan has quit IRC | 22:09 | |
*** Sravan has joined #openstack-kolla | 22:12 | |
*** Sravan has quit IRC | 22:13 | |
*** Sravan has joined #openstack-kolla | 22:13 | |
*** k_mouza has joined #openstack-kolla | 22:15 | |
*** k_mouza has quit IRC | 22:18 | |
*** luksky11 has quit IRC | 22:19 | |
openstackgerrit | Merged openstack/kolla-ansible master: Deprecate Ceph deployment https://review.opendev.org/669214 | 22:22 |
*** cgrosjean has quit IRC | 22:31 | |
*** goldyfruit has joined #openstack-kolla | 22:56 | |
*** jonaspaulo has quit IRC | 23:14 | |
*** goldyfruit has quit IRC | 23:19 | |
*** Sravan has quit IRC | 23:30 | |
*** BjoernT has joined #openstack-kolla | 23:48 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!