| *** ysandeep|out is now known as ysandeep | 02:04 | |
| opendevreview | OpenStack Proposal Bot proposed openstack/openstack-ansible master: Imported Translations from Zanata https://review.opendev.org/c/openstack/openstack-ansible/+/847704 | 03:40 |
|---|---|---|
| *** ysandeep is now known as ysandeep|afk | 05:15 | |
| *** ysandeep|afk is now known as ysandeep | 05:56 | |
| jrosser_ | good morning | 07:03 |
| *** frenzyfriday|PTO is now known as frenzy_friday | 07:13 | |
| admin1 | \o | 07:30 |
| opendevreview | likui proposed openstack/openstack-ansible-os_adjutant master: Replace deprecated UPPER_CONSTRAINTS_FILE variable https://review.opendev.org/c/openstack/openstack-ansible-os_adjutant/+/847721 | 08:28 |
| *** vkmc_ is now known as vkmc | 09:10 | |
| *** ysandeep is now known as ysandeep|afk | 09:50 | |
| Adri2000 | hello noonedeadpunk, I can propose a mariadb bump for wallaby. any "process" to follow when bumping mariadb version in a stable release? is a release note needed? or can I just change the version in the variable and that's it? | 10:08 |
| Adri2000 | noonedeadpunk: actually you mentioned the CVEs, those seem to impact <<10.5.16 as well, so I guess I should cherry-pick your patch from master and keep the release note about the CVEs. | 10:12 |
| jrosser_ | Adri2000: i think there might be some bad things happen if we backport releasenotes | 10:22 |
| Adri2000 | ah right the release note is not in the master branch patch (I suppose because it's the developement release, a relnote is not needed for that), it's only in the xena patch. so I'll create a new relnote for wallaby, and put the same content inside? | 10:34 |
| *** ysandeep|afk is now known as ysandeep | 10:38 | |
| admin1 | when i run octavia playbook, i don't see it creating the ssh key in the util folder .. did something changed on this behavior for the latest patch ? | 11:22 |
| admin1 | util container. | 11:22 |
| jrosser_ | are you sure it ever did that? | 11:23 |
| jrosser_ | show me the code :) | 11:23 |
| jrosser_ | admin1: it kind of depends how you have the service setup host configured really https://codesearch.opendev.org/?q=octavia_keypair_setup_host&i=nope&literal=nope&files=&excludeFiles=&repos= | 11:26 |
| jrosser_ | admin1: here is the code that does it https://github.com/openstack/openstack-ansible-os_octavia/blob/stable/wallaby/tasks/octavia_keypair.yml | 11:28 |
| jrosser_ | find the relevant part of your logging | 11:28 |
| *** dviroel|out is now known as dviroel | 12:01 | |
| jrosser_ | dmsimard: are you around? | 12:10 |
| *** arxcruz is now known as arxcruz|brb | 12:42 | |
| opendevreview | Adrien Cunin proposed openstack/openstack-ansible-galera_server stable/wallaby: Bump mariadb version. https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/847782 | 13:06 |
| *** arxcruz|brb is now known as arxcruz | 13:48 | |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Set zuul post-timeout to 3 hours https://review.opendev.org/c/openstack/openstack-ansible/+/847797 | 14:38 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible stable/yoga: Set zuul post-timeout to 3 hours https://review.opendev.org/c/openstack/openstack-ansible/+/847668 | 14:38 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible stable/yoga: Switch ARA report type from html to database for openstack-ansible jobs https://review.opendev.org/c/openstack/openstack-ansible/+/847652 | 14:38 |
| admin1 | jrosser_, found something .. in variables I have: octavia_ssh_enabled: true , but the logs say : | 15:14 |
| admin1 | Task "Write private key locally" has been omitted from the job because the conditional "['_octavia_keypair is changed', 'octavia_ssh_enabled']" was evaluated as "False" | 15:14 |
| admin1 | should true, be True ? (caps) | 15:14 |
| jrosser_ | both of those things would have to be true | 15:16 |
| jrosser_ | was this task 'changed'? https://github.com/openstack/openstack-ansible-os_octavia/blob/stable/wallaby/tasks/octavia_keypair.yml#L16-L32 | 15:17 |
| admin1 | me ? no .. my goal of using osa is as-is .. i don't modify anything | 15:19 |
| admin1 | changed true to True and running again .. if it sees it this time | 15:21 |
| admin1 | i recall ( not on osa but another application where true was not True ) and has to be caps | 15:25 |
| jrosser_ | you're going to have to interpret what happens with what the code says | 15:26 |
| jrosser_ | if it had not just created the keypair then it will never got copied | 15:27 |
| *** dviroel is now known as dviroel|lunch | 15:27 | |
| jrosser_ | i'm not suggesting that you need to modify anything, but to understand what does/doesnt happen it is necessary to dive into the code | 15:27 |
| jrosser_ | specifically 'changed' is one of the possible status-es of an ansible task | 15:28 |
| jrosser_ | i'm not suggesting that you have edited it | 15:28 |
| admin1 | understood . | 15:28 |
| jrosser_ | thats why there is "_octavia_keypair is changed" in the confitional that you pasted earlier | 15:29 |
| jrosser_ | the keypair is only copied if octavia_ssh_enabled=True and the status of the previous task was "changed" | 15:29 |
| jrosser_ | so if you run it a second time for exmaple, the key exists so that task is never "changed", so the key will not be copied | 15:29 |
| admin1 | technically i can login as octavia, delete the existing octavia_key , generate a new one with the same name, and save that one and its the same output right ? | 15:37 |
| admin1 | i mean the end goal will be the same | 15:37 |
| jrosser_ | i'm still pretty confused really | 15:44 |
| jrosser_ | what are you expecting to happen that is not happening? | 15:44 |
| *** ysandeep is now known as ysandeep|PTO | 16:09 | |
| admin1 | root issue: amphora instances not getting ip from dhcp | 16:09 |
| jrosser_ | and you want to ssh to them? | 16:11 |
| jrosser_ | admin1: ^ | 16:12 |
| admin1 | i need to fix the ip first .. and then do the ssh second | 16:12 |
| jrosser_ | it should dhcp from neutron | 16:12 |
| jrosser_ | just like a normal instance | 16:12 |
| admin1 | does the network node also need a br-lbaas ? | 16:14 |
| jrosser_ | thats up to you | 16:14 |
| jrosser_ | look at spatel blog post https://satishdotpatel.github.io/openstack-ansible-octavia/ | 16:15 |
| jrosser_ | if the network is a neutron vlan network and you already have something like br-vlan, there is nothing to do | 16:16 |
| admin1 | i have the exact same setup .. its working in 3 clusters, not working in the 4th one .. and so troubleshooting .. it has the same setup .. just diff tags | 16:18 |
| admin1 | and i don't think tags are an issue | 16:18 |
| spatel | br-lbass only required on controller node (compute node just use br-vlan to tag and talk to octavia manager) | 16:22 |
| admin1 | who serves dhcp ? | 16:22 |
| spatel | neutron-agent-dhcp | 16:22 |
| admin1 | they run in a network node .. which have br-vlan | 16:22 |
| spatel | you are going to create mgmt network for octavia which will wire up with namespace | 16:22 |
| admin1 | that is done .. controllers have br-lbaas on vlan 509 | 16:23 |
| spatel | check namespace | 16:23 |
| spatel | if you have dedicated network node then DHCP NS will be there | 16:24 |
| admin1 | how to figure out what dhcp is for ampora :D | 16:24 |
| spatel | This is what i have https://paste.opendev.org/show/bu2GqAZued5HLjjq7A6e/ | 16:27 |
| spatel | lbaas-mgmt which is DHCP going to provide IP | 16:27 |
| admin1 | yeah .. found it | 16:28 |
| spatel | https://paste.opendev.org/show/bIUBoS70jaiCHcH3ixvf/ | 16:29 |
| admin1 | now this dhcp should be via br-vlan connected to the vlan tag ? | 16:29 |
| spatel | that is my lbass DHCP namespace | 16:29 |
| spatel | Yes, you need to create VLAN on all your switch to tag from controller to compute | 16:29 |
| admin1 | i have controllers ( where br-lbaas is connected to br-vlan using the script ) , i have br-vlan on the computes .. and then i have a network node which has the namespaces ( ovs) .. do i need to do anything extra on the network node ? | 16:30 |
| admin1 | how to check ( ovs command) that the port ( of the dhcp namespace) is listening to the correct vlan tag ? | 16:32 |
| spatel | network node job is to just provide DHCP IP so i don't think you any anything else except br-vlan with correct vlan tag | 16:32 |
| spatel | ovs-vsctl show should tell you everything | 16:33 |
| admin1 | my network node has eno1 on br-vlan ...(ovs port) . .when I do tcpdump -ni eno1 -e vlan 509 .. i dont get the dhcp request that the amphora instance is requesting for | 16:34 |
| admin1 | it shows the port is in br-int | 16:34 |
| admin1 | but i don't know the ovs command to get the vlan id | 16:34 |
| spatel | why don't you post ovs-vsctl show command output here | 16:35 |
| *** dviroel|lunch is now known as dviroel | 16:37 | |
| spatel | OSA create lbaas-mgmt based on provided input in user_variable.yml file something like this - https://paste.opendev.org/show/bkgiclc3Vb6WoOuVZ39E/ | 16:40 |
| admin1 | spatel https://gist.githubusercontent.com/a1git/272357e5fafb7a78f8eda1d1515b420c/raw/401afa5acef6857581d7d277e46221c8dfe020cc/gistfile1.txt | 16:41 |
| admin1 | oh | 16:42 |
| admin1 | i see the issue | 16:42 |
| admin1 | | provider:segmentation_id | 4007 | -- is diff from the vlan i am using :( | 16:42 |
| admin1 | i think a config issue | 16:42 |
| admin1 | yeah .. found the issue . | 16:43 |
| admin1 | thanks spatel, jrosser_ | 16:44 |
| admin1 | is anyone running ironic ? | 16:50 |
| admin1 | how easy/hard is it to set it up | 16:50 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Exclude /var/log/journal from log collection https://review.opendev.org/c/openstack/openstack-ansible/+/847836 | 16:52 |
| admin1 | octavia is up now .. i had a diff vlan in the config ( copy/paste ) vs what I thought it should be and had the patches on ( 4007 on br-lbaas) vs 509 on everything else | 16:54 |
| spatel | cool | 16:56 |
| spatel | why are you running ovs just curious :) | 16:56 |
| admin1 | i have half half i think ..40% on lb , 60% on ovs | 17:02 |
| spatel | oh so some compute running on OVS and some on LXB | 17:02 |
| admin1 | no no .. these are different clusters | 17:03 |
| admin1 | some clusters on lb, some on ovs | 17:03 |
| spatel | oh okie | 17:03 |
| admin1 | i don't have a mix on the same env .. never tried that | 17:03 |
| spatel | I am running it :) | 17:04 |
| spatel | just for experiment | 17:04 |
| spatel | i did blog related that how to run in mix design | 17:04 |
| admin1 | have you tried ironic ? | 17:06 |
| jrosser_ | we are starting to try it in the lab | 17:06 |
| jrosser_ | but it's going to take a while | 17:07 |
| jrosser_ | lots to consider | 17:07 |
| admin1 | those using ceph, how do you backup ceph ? | 17:08 |
| admin1 | or consider DR | 17:08 |
| spatel | I didn't use ironic because it required some network changes. I have plan to do POC in lab | 17:08 |
| jrosser_ | admin1: i have cinder-backup with a ZFS/NFS backend, but it requires users to opt-in that the user experience is pretty poor | 17:09 |
| spatel | admin1 can't you create two ceph in two distinct location and tell crush to create geo replication between AZ | 17:09 |
| admin1 | spatel, setting that up in the lab this week | 17:13 |
| jrosser_ | rbd mirroring https://www.youtube.com/watch?v=ZifNGprBUTA | 17:13 |
| spatel | This is also good resource to start https://www.sebastien-han.fr/blog/2013/01/28/ceph-geo-replication-sort-of/ | 17:20 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Exclude /var/log/journal from log collection https://review.opendev.org/c/openstack/openstack-ansible/+/847836 | 17:21 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Fix facts gathering for zun https://review.opendev.org/c/openstack/openstack-ansible/+/846799 | 17:43 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Do not try to run etcd in serial mode https://review.opendev.org/c/openstack/openstack-ansible/+/847362 | 17:48 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-galera_server stable/wallaby: Bump mariadb version. https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/847782 | 17:49 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-galera_server stable/wallaby: Bump mariadb version. https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/847782 | 17:49 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-os_horizon stable/yoga: Fix ALLOWED_HOSTS https://review.opendev.org/c/openstack/openstack-ansible-os_horizon/+/847653 | 17:50 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible stable/yoga: Fix facts gathering for zun https://review.opendev.org/c/openstack/openstack-ansible/+/847261 | 17:50 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-os_ironic stable/yoga: Updated from OpenStack Ansible Tests https://review.opendev.org/c/openstack/openstack-ansible-os_ironic/+/846777 | 17:51 |
| opendevreview | Merged openstack/openstack-ansible stable/yoga: Update doc index page https://review.opendev.org/c/openstack/openstack-ansible/+/847397 | 18:13 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: DNM - display tree of collected logs https://review.opendev.org/c/openstack/openstack-ansible/+/844817 | 20:21 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: DNM - display tree of collected logs https://review.opendev.org/c/openstack/openstack-ansible/+/844817 | 20:22 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible stable/yoga: Return facts gathering to pre_tasks for masakari https://review.opendev.org/c/openstack/openstack-ansible/+/847260 | 21:18 |
| *** dviroel is now known as dviroel|out | 21:40 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!