| *** amoralej|off is now known as amoralej | 07:04 | |
| jingvar | Morining | 07:26 |
|---|---|---|
| opendevreview | Pierre Riteau proposed openstack/kolla-ansible master: Move project_name to role vars https://review.opendev.org/c/openstack/kolla-ansible/+/818714 | 07:27 |
| holtgrewe | Hi. I'm getting "MariaDB cluster is not synced. Please wait for WSREP sync before proceeding." for quite some time. Can I somehow enforce the wsrep? | 08:01 |
| holtgrewe | oh, just restart mariadb container that says not synced. | 08:03 |
| frickler | holtgrewe: which version are you using? yoctozepto had some fix for this recently. is this during initial deployment or later? | 08:05 |
| jingvar | I've deployed Xena with Masakari and having a lot of issues ... | 08:05 |
| holtgrewe | frickler, later | 08:05 |
| jingvar | disbaled TLS , enabled fqdn, and hostmontor tryes use fqdn name to connenct to pacemaker crmadmin -S ctl01.mycloud -fails, but works for "ctl01" | 08:08 |
| jingvar | I've changed /etc/kolla/hacluster-corosync/corosync.conf to use fqdn for nodes - now hostmonitor is working | 08:10 |
| jingvar | but can't send notifications - bacause it send short names in request | 08:11 |
| osmanlicilegi | morning | 08:13 |
| holtgrewe | How do I "neutron_external_interface" setup properly. Would I set `kolla_neutron_external_interfaces`, e.g., to [private_net_interface, public_net_interface] in group_vars/controllers/network-interfaces? | 09:05 |
| opendevreview | Merged openstack/kolla master: Use MariaDB 10.6 https://review.opendev.org/c/openstack/kolla/+/811015 | 09:57 |
| hrw | morning | 10:01 |
| frickler | holtgrewe: you need to specify both multiple interfaces and multiple bridges to attach them to https://docs.openstack.org/kolla-ansible/latest/reference/networking/neutron.html#example-multiple-interfaces | 10:06 |
| holtgrewe | frickler, I got it to work... by default bridges are added as physical interfaces | 10:06 |
| opendevreview | Michal Arbet proposed openstack/kolla master: Add kolla_git_commit LABEL for kolla images https://review.opendev.org/c/openstack/kolla/+/818727 | 10:08 |
| kevko | ^^ :) can someone check ? it's trivial patch | 10:09 |
| mnasiadka | Maybe it's trivial, but isn't it possible to get that from pip today? One could argue if you need a label | 10:11 |
| kevko | Well, I explained it in commit message .. | 10:12 |
| kevko | as a downstream user I would like to use my git where I have several changes ... | 10:12 |
| kevko | not merged features, bugs, etc ..and this is normally weeks or months | 10:13 |
| kevko | also, as downstream user I shouldn't change version of kolla in setup.cfg (or where it is) ..because I can easily be in conflict with upstream | 10:14 |
| kevko | so, if I have a git with my patches locally applied .. I will definitely not use kolla from pip .... but also need to know from which "version" - my git commit head images were built | 10:15 |
| hrw | kevko: "kolla-build --tag `git rev-parse --short HEAD`" does similar and is more visible | 10:16 |
| kevko | mnasiadka: so, no It's not possible to get kolla from pip ..because it's upstream ..this is feature for downstream user with changes .. | 10:16 |
| hrw | or 'git tag just-testing-11;kolla-build --tag just-testing-11' even | 10:17 |
| kevko | hrw: nope, what if I will building two times in a day ? from same commit .. but upstream changed upper constraints for example ..so i can't give same tag ...because images are different - I know it because of upstream contraints changed | 10:18 |
| kevko | so i want to have possibility to tag my images for example --tag wallaby-22-11-2021-11-18 and --tag wallaby-22-11-2021-15-30 | 10:19 |
| kevko | if i check both images, i can easily know if images are built from same commit or no .. | 10:20 |
| hrw | kevko: still - it is left for user how they handle marking images | 10:20 |
| hrw | kevko: in your case: --tag 'wallaby-`git rev-parse --short HEAD`-`date %options`' | 10:21 |
| hrw | done | 10:21 |
| hrw | and cleanly visible without going to logs or 'docker image inspect' | 10:21 |
| hrw | kevko: still kolla git 2579ea can give you different images than kolla git 2579ea 2 months later | 10:22 |
| hrw | and by default both will have same kolla_git_commit, tag and version | 10:22 |
| hrw | but distro packages inside will change, requirements will bump some u-c in source images | 10:22 |
| kevko | I would still like it | 10:24 |
| hrw | that's why we repeat 'build your own images for deployments and keep them' | 10:24 |
| hrw | kevko: inject it into tag in your build? | 10:24 |
| hrw | kevko: 'pip install kolla;kolla-build' and you do not have git info. are those images worse? | 10:25 |
| kevko | nope, their are just not built from git :) | 10:26 |
| hrw | and still can be just random version | 10:26 |
| mnasiadka | For me it's just a case that you're too lazy to tag your builds properly. | 10:26 |
| hrw | +2 for that | 10:27 |
| kevko | :D | 10:27 |
| kevko | ok, doesn't matter ..maybe you are right ..i will leave it there .. | 10:28 |
| kevko | I'm going to the sea, I'm on vacation :) :D | 10:28 |
| hrw | kevko: https://hub.docker.com/r/linaro/debian-source-neutron-base/tags | 10:28 |
| hrw | ldc-BRANCH-BUILDNUMBER is all what was needed to handle order for deployment | 10:29 |
| hrw | we do that way for ~4 years now | 10:29 |
| *** EugenMayer3 is now known as EugenMayer | 10:31 | |
| hrw | build your images with sensible tag, document somewhere which git tag/branch/commit, push to registry, use to deploy. 2 months later when you are adding new nodes, use the same images. | 10:31 |
| kevko | btw, guys, are u someone facing this issue https://bugs.launchpad.net/nova/+bug/1944779 ? | 10:32 |
| hrw | update code to latest stable, build, update deployment. then migrate to stable+1(+2,+3) same way | 10:32 |
| opendevreview | Dr. Jens Harbott proposed openstack/kolla-ansible master: Add a variable to set dns_servers for OVN https://review.opendev.org/c/openstack/kolla-ansible/+/818362 | 11:20 |
| opendevreview | Maksim Malchuk proposed openstack/kayobe master: Adds support for custom Placement configuration. https://review.opendev.org/c/openstack/kayobe/+/818755 | 11:23 |
| opendevreview | Adrian Andreias proposed openstack/kolla-ansible stable/xena: docs: Install openstack-client with upper constraints https://review.opendev.org/c/openstack/kolla-ansible/+/817975 | 11:37 |
| opendevreview | Adrian Andreias proposed openstack/kolla-ansible stable/wallaby: docs: Install openstack-client with upper constraints https://review.opendev.org/c/openstack/kolla-ansible/+/817976 | 11:46 |
| opendevreview | Adrian Andreias proposed openstack/kolla-ansible stable/victoria: docs: Install openstack-client with upper constraints https://review.opendev.org/c/openstack/kolla-ansible/+/817977 | 11:50 |
| admin1 | hi all.. is this also the room for kolla-ansible .. how to get around/fix Ensure hostname does not point to 127.0.1.1 in /etc/hosts .. where there is no 127.0.1.1. entry in /etc/hosts .. | 12:13 |
| frickler | admin1: what action are you doing to trigger this and how does your log look like? use paste.opendev.org to show it | 12:20 |
| jingvar | sed -i 's/gather_facts: false/gather_facts: true/g' /opt/kayobe/venvs/kolla-ansible/share/kolla-ansible/ansible/site.yml | 12:48 |
| jingvar | can someone to show anisble_hostname from centos | 12:59 |
| jingvar | with fqdn on host | 13:00 |
| frickler | mnasiadka: why is neutron_dnsmasq_dns_servers defaulted in ansible/group_vars/all.yml instead of ansible/roles/neutron/defaults/main.yml ? likely I should move neutron_ovn_dns_servers then, too- | 13:19 |
| mnasiadka | frickler: if it's not used in any other role, that's probably an error | 13:20 |
| mnasiadka | frickler: we could also move to one variable... | 13:21 |
| frickler | mnasiadka: at least not in the kolla-ansible repo. I'm also wondering whether ... exactly | 13:22 |
| frickler | mnasiadka: but that would be a new variable then that works as default for both? | 13:22 |
| mnasiadka | frickler: yes, but we need a reno that the variable has been renamed ;-) | 13:23 |
| frickler | or just reuse neutron_dnsmasq_dns_servers although the "dnsmasq" part is misleading for ovn? | 13:23 |
| mnasiadka | that's also an option | 13:23 |
| frickler | hmm, I'll think about it a bit, thx for now | 13:24 |
| *** amoralej is now known as amoralej|lunch | 13:25 | |
| frickler | probably I need to test what happens when dhcp-agent is deployed in addition to ovn. there might be an obscure use case to actually use different nameservers in that scenario | 13:27 |
| mnasiadka | we have that (ironic bare metal + ovn case), but wouldn't think of setting different resolvers :) | 13:32 |
| mnasiadka | you can always override the config template | 13:32 |
| admin1 | frickler this is a brand new install .. and the error comes in "kolla-ansible -i multinode bootstrap-servers" | 13:56 |
| jingvar | can I run kayobe overcloud service deploy for particular services? | 13:58 |
| *** amoralej|lunch is now known as amoralej | 14:04 | |
| parallax | jingvar: kayobe overcloud service deploy -kt {service_name} | 14:09 |
| jingvar | thanks | 14:14 |
| JohnB | A quick question. The nodes in the overcloud. Do they need BIOS boot or UEFI boot? We would prefer UEFI boot as this helps us with NVMe RAID, but don't know if this is possible. | 14:21 |
| jingvar | I think yes, it just Ironic options | 14:27 |
| opendevreview | Merged openstack/kolla-ansible stable/xena: docs: Install openstack-client with upper constraints https://review.opendev.org/c/openstack/kolla-ansible/+/817975 | 14:28 |
| opendevreview | Merged openstack/kolla-ansible stable/wallaby: docs: Install openstack-client with upper constraints https://review.opendev.org/c/openstack/kolla-ansible/+/817976 | 14:32 |
| opendevreview | Merged openstack/kolla-ansible stable/victoria: docs: Install openstack-client with upper constraints https://review.opendev.org/c/openstack/kolla-ansible/+/817977 | 14:42 |
| admin1 | frickler .. https://paste.opendev.org/show/811226/ | 15:28 |
| admin1 | all servers .. none of them have 127.0.1.1 in thier /etc/hosts | 15:28 |
| admin1 | could it be some facts caching somewhere ? | 15:28 |
| admin1 | frickler even if i blank the /etc/hosts file, that error is not going away | 15:45 |
| admin1 | what could i be missing ? | 15:45 |
| admin1 | "The task includes an option with an undefined variable. The error was: 'dict object' has no attribute 'hostname'" -- maybe hosts need to return some sort of hostname ? | 15:46 |
| mgoddard | admin1: you aren't gathering facts | 16:03 |
| mgoddard | admin1: and looks like your inventory is broken | 16:03 |
| admin1 | mgoddard, how to validate ? | 16:05 |
| admin1 | i followed https://docs.openstack.org/kolla-ansible/latest/user/quickstart.html | 16:05 |
| mgoddard | admin1: you won't get all the warnings if valid | 16:05 |
| admin1 | let me pastebin my configs | 16:06 |
| admin1 | well the warning is misleading isn't it ? | 16:06 |
| opendevreview | Verification of a change to openstack/kayobe stable/victoria failed: CI: always return host configure test results https://review.opendev.org/c/openstack/kayobe/+/818571 | 16:06 |
| admin1 | if inventory file is bad, it should not say 127.0.1.1 exists in /etc/hosts ? | 16:07 |
| jingvar | admin1: if inventory is bad it does'nt wok | 16:16 |
| jingvar | wok- work | 16:16 |
| jingvar | it is ok [WARNING]: Invalid characters were found in group names but not replaced, use -vvvv to see details | 16:17 |
| jingvar | I think it is bad - [WARNING]: * Failed to parse /etc/kolla/multinode with yaml plugin | 16:18 |
| *** amoralej is now known as amoralej|off | 16:20 | |
| yoctozepto | oh, it seems you were analysing the same thing I did for admin0 in https://bugs.launchpad.net/kolla-ansible/+bug/1951849 | 16:30 |
| jingvar | Yes, I faced with it when play with Xena and Maskari, just enable gather fact via sed :) | 16:32 |
| jingvar | mgoddard: https://github.com/openstack/kolla-ansible/blob/stable/xena/ansible/roles/hacluster/tasks/bootstrap_service.yml | 16:34 |
| jingvar | why there is ansible_facts.hostname instead of ansible_facts.fqdn | 16:34 |
| jingvar | Hostmonitor checks crmadmin -S ctl01.mycluser - and fails , because resources added with their short names. | 16:37 |
| jingvar | I've replaced in bootstrap_service.yml and hacluster_corosync.conf.j2 to ansible_facts.fqdn - it works . Hostmonitor uses import socket ; socket.hostanme - it returns fqdn | 16:41 |
| jingvar | yoctozepto: You are one of contributors :) | 16:45 |
| jingvar | masakarimonitors/hostmonitor/host_handler/handle_host.py, utils.execute('crmadmin', '-S', self.my_hostname, self.my_hostname = socket.gethostname() | 16:52 |
| opendevreview | Dr. Jens Harbott proposed openstack/kolla-ansible master: Add a variable to set dns_servers for OVN https://review.opendev.org/c/openstack/kolla-ansible/+/818362 | 17:06 |
| opendevreview | Merged openstack/kayobe stable/wallaby: CI: always return host configure test results https://review.opendev.org/c/openstack/kayobe/+/818570 | 18:06 |
| opendevreview | Merged openstack/kayobe stable/xena: CI: always return host configure test results https://review.opendev.org/c/openstack/kayobe/+/818489 | 18:06 |
| ozzzo_work | What's the best way to find out which groups an openstack user is in? | 18:58 |
| ozzzo_work | I'm looking here but I don't see a command to list group membership for a user: https://docs.openstack.org/python-openstackclient/train/cli/command-objects/group.html | 19:01 |
| frickler | ozzzo_work: "openstack group list --user $username" | 19:25 |
| ozzzo_work | ty frickler | 20:05 |
| guesswhat | guys? is possible to autocreate external network in neuron if i am using neutron_external_interface in globals.yaml ? or do i need to create it manually? thanks | 21:27 |
| guesswhat | is there any chance to support DNS01 in kolla ? or do i need to run separate service like certbot? | 22:31 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!