| *** ysandeep|out is now known as ysandeep | 05:42 | |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible stable/yoga: Include install_method variables for openrc https://review.opendev.org/c/openstack/openstack-ansible/+/859168 | 07:20 |
|---|---|---|
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible stable/xena: Include install_method variables for openrc https://review.opendev.org/c/openstack/openstack-ansible/+/859169 | 07:20 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible stable/wallaby: Include install_method variables for openrc https://review.opendev.org/c/openstack/openstack-ansible/+/859230 | 07:20 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/ansible-role-python_venv_build stable/yoga: Change default value for venv_wheel_build_enable https://review.opendev.org/c/openstack/ansible-role-python_venv_build/+/859231 | 07:23 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_keystone stable/yoga: Bootstrap when running against last backend https://review.opendev.org/c/openstack/openstack-ansible-os_keystone/+/859232 | 07:29 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_keystone stable/yoga: Bootstrap when running against last backend https://review.opendev.org/c/openstack/openstack-ansible-os_keystone/+/859232 | 07:31 |
| dokeeffe85 | Morning, still a little stuck with the networking on OSA, internal nets are working fine but I can't get an external network working. I can see the ovs bridge br-vlan and it has the right interface but when I spin up a vm and look at the ports they all have br-int as their bridge https://paste.openstack.org/show/btlS1NdYgvzKT4okkwwV/ | 08:20 |
| noonedeadpunk | dokeeffe85: so there are 2 ways of getting external networking - 1. through l3 router and floarting ips, 2. pass public network to each compute | 08:25 |
| *** ysandeep is now known as ysandeep|afk | 08:25 | |
| noonedeadpunk | If you're trying to get public network passed directly to a vm, then it's option 2. And it quite depends on how you have created the network on neutron | 08:26 |
| noonedeadpunk | as then you should explicitly provide network type to be flat for vlan and pass correct vlanid (if it's vlan) | 08:27 |
| dokeeffe85 | Ah ok, it's the second way. So it should be a flat network not a vlan. Ok let me re do that so. Thanks | 08:30 |
| dokeeffe85 | noonedeadpunk thanks a million, all sorted | 08:38 |
| opendevreview | Merged openstack/ansible-role-systemd_mount master: Remove redundant vars line https://review.opendev.org/c/openstack/ansible-role-systemd_mount/+/857861 | 08:54 |
| opendevreview | Merged openstack/ansible-role-uwsgi master: Remove redundant vars line https://review.opendev.org/c/openstack/ansible-role-uwsgi/+/857876 | 08:59 |
| opendevreview | Merged openstack/openstack-ansible-os_masakari master: Remove redundant vars line https://review.opendev.org/c/openstack/openstack-ansible-os_masakari/+/857653 | 09:05 |
| opendevreview | Merged openstack/openstack-ansible-rabbitmq_server master: Remove redundant vars line https://review.opendev.org/c/openstack/openstack-ansible-rabbitmq_server/+/857658 | 09:06 |
| opendevreview | Merged openstack/openstack-ansible-os_cloudkitty master: Remove redundant vars line https://review.opendev.org/c/openstack/openstack-ansible-os_cloudkitty/+/857650 | 09:07 |
| opendevreview | Merged openstack/openstack-ansible-rsyslog_client master: Remove redundant vars line https://review.opendev.org/c/openstack/openstack-ansible-rsyslog_client/+/857837 | 09:09 |
| opendevreview | Merged openstack/openstack-ansible-os_adjutant master: Refactor use of include_vars https://review.opendev.org/c/openstack/openstack-ansible-os_adjutant/+/857827 | 09:17 |
| *** ysandeep|afk is now known as ysandeep | 09:50 | |
| opendevreview | Merged openstack/openstack-ansible-ops master: Correct OVNL3RouterPlugin path https://review.opendev.org/c/openstack/openstack-ansible-ops/+/856569 | 09:51 |
| admin1 | dokeeffe85, here is a doc for the 2nd way => https://www.openstackfaq.com/openstack-add-direct-attached-dhcp-ip/. .. then you just select the network like you select private, but you get direct (public) ips | 09:57 |
| admin1 | not an osa question, but just asking if people know: is it possible to update/change a project from one domain to another ? | 10:00 |
| noonedeadpunk | I don't really think so. | 10:21 |
| noonedeadpunk | But you totally can grant permissions to user from another domain to project in different domain | 10:21 |
| noonedeadpunk | Or hacking DB | 10:22 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Fix cinder-volume AIO ceph scenario https://review.opendev.org/c/openstack/openstack-ansible/+/859263 | 10:54 |
| admin1 | i will try it in hacking a db :) | 11:06 |
| opendevreview | Merged openstack/openstack-ansible-os_ironic master: Replace pxe_append_params with kernel_pxe_params in ironic.conf https://review.opendev.org/c/openstack/openstack-ansible-os_ironic/+/856590 | 11:06 |
| *** ysandeep is now known as ysandeep|afk | 11:20 | |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_neutron master: Avoid ovs restart during package upgrade https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/857652 | 11:21 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_neutron master: Avoid ovs restart during package upgrade https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/857652 | 11:21 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_neutron master: Avoid ovs restart during package upgrade https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/857652 | 11:21 |
| kleini_ | Is it possible with config_overrides to remove some option from an ini config file section? Providing no value gives key = None but I would like to remove key. | 11:24 |
| *** kleini_ is now known as kleini | 11:24 | |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-galera_server master: Use policy_rc_d attribute instead of copy https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/859272 | 11:28 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-lxc_hosts master: Use policy_rc_d attribute instead of copy https://review.opendev.org/c/openstack/openstack-ansible-lxc_hosts/+/859273 | 11:31 |
| noonedeadpunk | kleini: I think it's not possible right now. But it's might be good thing to implement indeed (though quite tricky) | 11:32 |
| noonedeadpunk | best you can do is define option to empty string | 11:33 |
| noonedeadpunk | Highly likely oslo.config will interpret that as variable being not defined | 11:34 |
| noonedeadpunk | But can't remember for sure | 11:34 |
| *** ysandeep|afk is now known as ysandeep | 11:38 | |
| noonedeadpunk | (but can recall some discussions about that) | 11:40 |
| noonedeadpunk | (and I think we edned up with setting empty string is quite same as not define at all for oslo) | 11:41 |
| *** ysandeep is now known as ysandeep|afk | 12:40 | |
| kleini | my current case is about ansible-role-systemd_networkd and I would like to configure the Match section with a MACAddress and not with a Name. But there seems currently no possible solution in my case: https://opendev.org/openstack/ansible-role-systemd_networkd/src/branch/master/templates/systemd-network.j2#L4 | 12:43 |
| kleini | will put that on my TODO list | 12:45 |
| mgariepy | https://opendev.org/openstack/ansible-config_template/src/branch/master/library/config_template#L52-L57 | 12:53 |
| mgariepy | kleini, isn't that suppose to clean up none value from the override? | 12:54 |
| kleini | should be worth a try | 12:57 |
| *** ysandeep|afk is now known as ysandeep | 13:08 | |
| kleini | hmm, default of ignore_none_type is true. I set it to true but I still get "Name = None" in my ens3.network file. So for me this option seems to be broken. | 13:11 |
| mgariepy | hmm :/ sad/ | 13:15 |
| kleini | Got it working. My problem was, that I define input to ansible-role-systemd_networkd as yaml. And the equivalent of None in Yaml is "Name: {}" | 13:17 |
| kleini | very nice! | 13:17 |
| mgariepy | cool | 13:18 |
| *** ysandeep is now known as ysandeep|out | 14:10 | |
| noonedeadpunk | it's worth documenting somewhere.... | 14:12 |
| noonedeadpunk | not too obvious I would say | 14:12 |
| kleini | should I document that within the config_template collection or systemd_networkd role? | 14:23 |
| kleini | maybe I can add both. my example case for systemd_networkd and something general for config_template. | 14:23 |
| noonedeadpunk | kleini: I think it make sense to add it to integrarted repo | 14:42 |
| noonedeadpunk | like https://opendev.org/openstack/openstack-ansible/src/branch/master/doc/source/reference/configuration/using-overrides.rst | 14:43 |
| psymin | I've recently installed openstack ansible aio on a rocky machine and I have a few rather novice questions. For example, I'm not sure how to ssh to the box any more. | 15:09 |
| *** dviroel is now known as dviroel|lunch | 15:10 | |
| noonedeadpunk | psymin: well, aio should not drop your keypair from machine. but I do recall such behaviour in past, and it was related to either some networking issue (when aio-provisioned networks were intersecting with actuall network on machine) | 15:12 |
| noonedeadpunk | as we do hardcode nets, so you can get quite unlucky | 15:12 |
| noonedeadpunk | but other then that you should be able to ssh to VM as ususal | 15:16 |
| noonedeadpunk | *usual | 15:16 |
| psymin | It appears root login w/ password was disabled after the aio install. Resolved. | 15:42 |
| psymin | For the real install I will use keys. | 15:42 |
| noonedeadpunk | oh yes. it's effect of hardening playbook :) | 15:46 |
| noonedeadpunk | for some reason I didn't assume that you had root login with password, sorry | 15:46 |
| noonedeadpunk | you can disable this behaviour if you want to | 15:47 |
| psymin | It is a valid assumption to make. I won't be using aio for production, just getting a feel for the system. After this seems to work we'll be trying it on three baremetal machines. | 15:47 |
| noonedeadpunk | you can set `security_sshd_permit_root_login: yes` if needed | 15:48 |
| noonedeadpunk | (in user_variables.yml) | 15:48 |
| psymin | ahh in the ansible playbook, cool. It isn't needed yet :) I'll use the keys for the next step of install testing. | 15:49 |
| psymin | Is there a good article I could read for learning how to use the default horizon interface in an aio install? | 15:50 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Fix cinder-volume AIO ceph scenario https://review.opendev.org/c/openstack/openstack-ansible/+/859263 | 15:58 |
| psymin | my next issue is that I'm trying to test a 32gb qcow2 file to verify it can work, but "copy image to volume:Image used for creating volume exceeds available space." | 16:04 |
| psymin | so I need to learn how to add existing block devices for space, or something | 16:04 |
| noonedeadpunk | well, for aio we create loop devices and I think we assume having 100gb drive (as it's what we have in CI) | 16:06 |
| noonedeadpunk | I think we use /dev/sdb if any is present, but I need to double-check | 16:07 |
| psymin | all block devices on this baremetal are 1.8tb, I do see some loop devices created /dev/loop2 through /dev/loop6 .. /dev/loop6 is only 11gigs and mapped to /var/lib/machines | 16:09 |
| noonedeadpunk | you can use `export BOOTSTRAP_OPTS="bootstrap_host_data_disk_device=sdb"` before running bootstrap_aio.sh | 16:10 |
| noonedeadpunk | ah, ok, so if it's root device.... Let me check how we do create loop drives | 16:11 |
| noonedeadpunk | we have variable bootstrap_host_loopback_cinder_size for cinder loop device size at very least | 16:11 |
| psymin | gotcha | 16:12 |
| noonedeadpunk | check this out https://opendev.org/openstack/openstack-ansible/src/branch/master/tests/roles/bootstrap-host/defaults/main.yml#L46-L100 | 16:12 |
| psymin | lvdisplay, pvdisplay, df -h https://pastebin.com/raw/775NNPVk | 16:14 |
| noonedeadpunk | ok, yes, so loop device is big enough indeed | 16:14 |
| noonedeadpunk | wait, how have you created volume from image? | 16:15 |
| noonedeadpunk | As if you have not specified min_disk for the image, you would need to provide disk size explicitly | 16:15 |
| noonedeadpunk | and you do have minimal disk created | 16:16 |
| psymin | I uploaded the qcow file using "create image" under "compute" in horizon. then I "create volume" under that image, changed the default size from 32gb to 33gb for a buffer, and clickec "create volume" | 16:17 |
| psymin | It is very likely I'm taking the wrong steps. | 16:17 |
| psymin | if there is a different channel I should be asking this novice questions in, please let me know | 16:18 |
| noonedeadpunk | ok, so you're using horizon. when you're uploading image - you can define properties for it. Among them you can define minimal disk. Also when creating volume you can set volume size. | 16:19 |
| noonedeadpunk | btw, you've isntalled Yoga? | 16:19 |
| psymin | Yep, yoga, rocky 8. Branch 25.1.0 I think. | 16:19 |
| noonedeadpunk | As we were recently reported that horizon appears broken on master | 16:19 |
| psymin | If there is a different path I should be taking I'm all ears. Rocky 9? | 16:20 |
| *** dviroel|lunch is now known as dviroel | 16:20 | |
| noonedeadpunk | psymin: nah, I was jsut wondering :) Rocky 9 is WIP and we want to backport it's support to Yoga quite soon | 16:21 |
| noonedeadpunk | psymin: I guess your issue at the moment is more about data you've filled in either when creating volume or when umploading image | 16:21 |
| noonedeadpunk | *uploading | 16:21 |
| noonedeadpunk | psymin: are you familiar with openstack cli? | 16:22 |
| psymin | noonedeadpunk, not at all familiar. I can learn though. | 16:22 |
| psymin | Perhaps it would be wise for me to configure user_variables.yml to use the eight 1.8 tb drives, and reinstall? | 16:23 |
| noonedeadpunk | ok, so it's pre-installed inside utility container with aio | 16:23 |
| psymin | is there a way to ssh to that container from the host? | 16:24 |
| noonedeadpunk | so you can do lxc-ls to list container and then lxc-attach -n aio_utility_(smth-smth) | 16:24 |
| noonedeadpunk | once you're in - run `source /root/openrc` | 16:24 |
| noonedeadpunk | that will give CLI admin privileges | 16:24 |
| noonedeadpunk | so out of interest - `openstack image list` `openstack image show $UUID` of your image | 16:25 |
| noonedeadpunk | And you can create volume with command `openstack volume create --size 100 --image $UUID somename --bootable` | 16:26 |
| noonedeadpunk | size is in GB | 16:26 |
| psymin | ahh, in the utility container is where I can use the openstack command, thank you so much. | 16:27 |
| noonedeadpunk | well, you can install python-openstackclient anywhere | 16:28 |
| noonedeadpunk | but it's pre-installed there | 16:28 |
| noonedeadpunk | utility container also does have .my.cnf so you can access DB from there as well | 16:28 |
| psymin | status is "error" in `openstack volume list` as well as in horizon after creating the volume from the image. | 16:31 |
| psymin | I used 100 even though the image is 32 | 16:32 |
| psymin | would a path of less resistance be to try using the three baremetal servers as a test as opposed to aio? | 16:33 |
| noonedeadpunk | and what is `openstack volume show` returns you as error reason? | 16:37 |
| noonedeadpunk | also what volume backend do you want to use? | 16:37 |
| psymin | after learning more, I hope to use ceph | 16:38 |
| psymin | I don't see a reason with "openstack volume list <id>", just "status: error", but in horizon "Image used for creating volume exceeds available space." | 16:39 |
| noonedeadpunk | Well, I would assume to spawn aio with ceph scenario but found out today that it's currently broken. I pushed https://review.opendev.org/c/openstack/openstack-ansible/+/859263 in order to fix it | 16:39 |
| noonedeadpunk | "openstack volume show <id>" | 16:40 |
| noonedeadpunk | you meant? | 16:40 |
| psymin | https://pastebin.com/raw/tA2LBRHG | 16:40 |
| psymin | yep, apologies | 16:40 |
| noonedeadpunk | openstack volume service list? | 16:41 |
| noonedeadpunk | eventually I can assume that we've got cinder jsut broken in aio.... | 16:41 |
| noonedeadpunk | as appeared we were not testing cinder properly | 16:41 |
| psymin | https://pastebin.com/raw/0HAmgpFN | 16:41 |
| noonedeadpunk | hm, but it looks quite healthy | 16:42 |
| psymin | I'd assume that the error is on my end rather than with your tools | 16:42 |
| noonedeadpunk | you can enter cinder-volume container and check logs with journalctl | 16:43 |
| noonedeadpunk | But i'm not sure if error happens on cinder or glance side... | 16:43 |
| noonedeadpunk | Anyway you will know more about how to troubleshoot deployments :) | 16:43 |
| noonedeadpunk | eventually I can recall some troubles with iscsi configuration for aio as well, but well. never had time to dig into that | 16:50 |
| noonedeadpunk | psymin: though, you actually don't need volume to spawn a VM from the image | 16:50 |
| noonedeadpunk | You can create server directly from the image. Despite it's not best way to do that - but at least you will be able to test your image right away and see if it works | 16:51 |
| noonedeadpunk | and sort out volume thing later | 16:51 |
| noonedeadpunk | you can do that from horizon as well, on server creation page you will need to select No on question if create new volume | 16:52 |
| psymin | I am seeing issues in journalctl for cinder. More context: this baremetal had eno1 for a nic that would start onboot. After installing aio the nic changed to eth0. This caused some issue. This mornign I set eth0 to come up on boot and rebooted. | 16:52 |
| noonedeadpunk | yeah, I guess we're unifying things with AIO and mess with networking quite a lot | 16:54 |
| noonedeadpunk | we mostly use aio for ci/cd, so need to ensure state let's say | 16:55 |
| psymin | perhaps I should install rocky 8 on these three machines, get them online, create a better user_variables.yml file, and test that way? | 16:56 |
| noonedeadpunk | (and openstack_user_config) | 16:58 |
| noonedeadpunk | well, it depends on goals. that's totally worth doing for production, yes | 16:58 |
| noonedeadpunk | aio totally not designed for any real usage | 16:59 |
| noonedeadpunk | more for checking out concepts and quickly check out some things | 16:59 |
| noonedeadpunk | or run CI | 16:59 |
| psymin | End goal is for production locally for one organization, no untrusted guests. | 16:59 |
| noonedeadpunk | so at the end you will still need to create deployment from scratch ) | 17:00 |
| noonedeadpunk | can you paste errors you've seen in cinder containers? | 17:01 |
| psymin | Sure, they may not be relevant to the immediate issue though, since there were networking issues that caused haproxy to fail, for example. | 17:02 |
| noonedeadpunk | ah | 17:02 |
| noonedeadpunk | you can actually do `journalctl -f -u cinder-*` and try to reproduce issue. Command will follow current logs | 17:03 |
| noonedeadpunk | aio does enable debug by default, so you might want to redirect output... | 17:04 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Bump ansible-core version to 2.13.4 https://review.opendev.org/c/openstack/openstack-ansible/+/857506 | 17:14 |
| *** dviroel is now known as dviroel|dr_appt | 17:20 | |
| psymin | python3.6 is listed in the journalctl (and in the horizon details traceback). Should it be 3.8? | 17:34 |
| psymin | https://pastebin.com/raw/Jse0TuTG | 17:35 |
| noonedeadpunk | well yes, I'd assume it to be 3.8 | 17:41 |
| noonedeadpunk | though it's unlikely root cause of the issue you see | 17:42 |
| noonedeadpunk | psymin: I will spawn up Rocky tomorrow in aio to check what's wrong with python versions there | 17:48 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Bump ansible-core version to 2.13.4 https://review.opendev.org/c/openstack/openstack-ansible/+/857506 | 17:50 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/ansible-role-python_venv_build stable/yoga: Change default value for venv_wheel_build_enable https://review.opendev.org/c/openstack/ansible-role-python_venv_build/+/859231 | 18:02 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Fix cinder-volume AIO ceph scenario https://review.opendev.org/c/openstack/openstack-ansible/+/859263 | 18:30 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Fix cinder-volume AIO ceph scenario https://review.opendev.org/c/openstack/openstack-ansible/+/859263 | 18:36 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Extend NFS scenario to support cinder https://review.opendev.org/c/openstack/openstack-ansible/+/859336 | 18:40 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_cinder master: Add nfs and ceph jobs to templates https://review.opendev.org/c/openstack/openstack-ansible-os_cinder/+/859337 | 18:43 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Run test_minimum_basic_scenario test for cinder https://review.opendev.org/c/openstack/openstack-ansible/+/859339 | 19:19 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_cinder master: Add nfs and ceph jobs to templates https://review.opendev.org/c/openstack/openstack-ansible-os_cinder/+/859337 | 19:20 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Fix cinder-volume AIO ceph scenario https://review.opendev.org/c/openstack/openstack-ansible/+/859263 | 19:31 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Extend NFS scenario to support cinder https://review.opendev.org/c/openstack/openstack-ansible/+/859336 | 19:32 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Run test_minimum_basic_scenario test for cinder https://review.opendev.org/c/openstack/openstack-ansible/+/859339 | 19:32 |
| prometheanfire | is there something 'decided upon' for monitoring an osa deploy? I imagine the elk metrics from osa-ops will be good for metrics but not alerting, at least at first | 20:12 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Fix cinder-volume AIO ceph scenario https://review.opendev.org/c/openstack/openstack-ansible/+/859263 | 21:01 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Extend NFS scenario to support cinder https://review.opendev.org/c/openstack/openstack-ansible/+/859336 | 21:01 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Run test_minimum_basic_scenario test for cinder https://review.opendev.org/c/openstack/openstack-ansible/+/859339 | 21:01 |
| noonedeadpunk | prometheanfire: nah, not really. Well, we do have some prometheus playbook in ops but I have really no idea when and if anybody ever used it https://opendev.org/openstack/openstack-ansible-ops/src/branch/master/prometheus | 21:08 |
| noonedeadpunk | and it really noop rather then any real thing imo | 21:09 |
| noonedeadpunk | but I guess for monitoring it really pretty much depends as it's quite spread across some nrpe/zabbix/prometheus so everybody use own thing as a result | 21:10 |
| prometheanfire | ya, I saw the prometheus thing there, not updated in 4 years | 22:55 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!