| *** dviroel|out is now known as dviroel | 00:33 | |
| *** dviroel is now known as dviroel|out | 01:00 | |
| *** ysandeep|out is now known as ysandeep | 01:52 | |
| opendevreview | Merged openstack/openstack-ansible stable/wallaby: Fix facts gathering for zun https://review.opendev.org/c/openstack/openstack-ansible/+/849631 | 01:56 |
|---|---|---|
| *** ysandeep is now known as ysandeep|afk | 03:23 | |
| noonedeadpunk | mornings | 07:58 |
| noonedeadpunk | and also bump mariadb version as https://zuul.opendev.org/t/openstack/build/97fee39bdbff41868a608475b50dc3db/log/job-output.txt#12439 | 08:04 |
| noonedeadpunk | hm, seems my previous message was lost | 08:04 |
| noonedeadpunk | and, I was talking that if we want to backport https://review.opendev.org/c/openstack/openstack-ansible-rabbitmq_server/+/850350 we also need to backport rally | 08:05 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_rally stable/ussuri: Control rally-openstack installed version https://review.opendev.org/c/openstack/openstack-ansible-os_rally/+/850477 | 08:06 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_rally stable/ussuri: Control rally-openstack installed version https://review.opendev.org/c/openstack/openstack-ansible-os_rally/+/850477 | 08:08 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_rally stable/ussuri: Control rally-openstack installed version https://review.opendev.org/c/openstack/openstack-ansible-os_rally/+/850477 | 08:09 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible stable/xena: Set minimum and maximum microversions for manila api https://review.opendev.org/c/openstack/openstack-ansible/+/850478 | 08:23 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible stable/xena: Set minimum and maximum microversions for manila api https://review.opendev.org/c/openstack/openstack-ansible/+/850478 | 08:23 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_manila stable/xena: Create backends when running against manila_share https://review.opendev.org/c/openstack/openstack-ansible-os_manila/+/847993 | 08:23 |
| jrosser | morning | 08:49 |
| damiandabrowski | hi! | 08:51 |
| opendevreview | Merged openstack/openstack-ansible-galera_server stable/xena: Bump mariadb version. https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/847435 | 09:27 |
| opendevreview | Merged openstack/openstack-ansible-os_tempest stable/xena: Drop tripleo c7 job definition https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/849977 | 09:28 |
| opendevreview | Merged openstack/openstack-ansible-os_tempest stable/wallaby: Remove tripleo jobs from stable/wallaby https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/849970 | 09:28 |
| opendevreview | Merged openstack/openstack-ansible-os_tempest stable/victoria: Remove tripleo jobs from stable/victoria branch https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/849971 | 09:34 |
| opendevreview | Merged openstack/openstack-ansible stable/wallaby: Switch ARA report type from html to database for openstack-ansible jobs https://review.opendev.org/c/openstack/openstack-ansible/+/849812 | 09:53 |
| opendevreview | Merged openstack/openstack-ansible stable/victoria: Switch ARA report type from html to database for openstack-ansible jobs https://review.opendev.org/c/openstack/openstack-ansible/+/849798 | 10:05 |
| *** ysandeep|afk is now known as ysandeep | 10:14 | |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_rally stable/ussuri: Control rally-openstack installed version https://review.opendev.org/c/openstack/openstack-ansible-os_rally/+/850477 | 10:23 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_rally stable/ussuri: Control rally-openstack installed version https://review.opendev.org/c/openstack/openstack-ansible-os_rally/+/850477 | 10:24 |
| *** anbanerj is now known as frenzy_friday | 11:33 | |
| *** dviroel|out is now known as dviroel | 11:34 | |
| *** anbanerj is now known as frenzy_friday | 11:58 | |
| opendevreview | Marc Gariépy proposed openstack/openstack-ansible-rabbitmq_server stable/ussuri: Use cloudsmith repo for rabbit and erlang https://review.opendev.org/c/openstack/openstack-ansible-rabbitmq_server/+/850350 | 12:05 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-haproxy_server stable/xena: Don't restrict haproxy tunable options https://review.opendev.org/c/openstack/openstack-ansible-haproxy_server/+/850480 | 13:09 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_horizon stable/xena: Fix ALLOWED_HOSTS https://review.opendev.org/c/openstack/openstack-ansible-os_horizon/+/848004 | 13:09 |
| noonedeadpunk | I'm not sure if it's good idea to backport https://review.opendev.org/c/openstack/openstack-ansible-haproxy_server/+/850480 | 13:11 |
| noonedeadpunk | but defining tune.maxrewrite is required | 13:11 |
| noonedeadpunk | maybe I should instead just push simple patch to add tune.maxrewrite 1280 as a default thing... | 13:12 |
| mgariepy | jrosser, with gluster if i started aio with 1 repo and want to scale it to 3, was that supposed to work ? | 13:29 |
| jrosser | thats an interesting question | 13:29 |
| mgariepy | haha | 13:29 |
| jrosser | so it doesnt need gluster at all in a single repo server setup, but it always deploys it even with just one | 13:30 |
| mgariepy | https://paste.openstack.org/show/bdcpnRs8LlRUTjaTprPC/ | 13:30 |
| jrosser | i've tried replacing a repo server when there are 3 and that should be ok | 13:31 |
| mgariepy | yeah but adding one after is kinda.. broken now. | 13:31 |
| jrosser | well it may never have worked | 13:31 |
| jrosser | the gluster role i wrote is very simple | 13:32 |
| mgariepy | yeah, seems like we cannot add peers when there is already a volume. | 13:32 |
| jrosser | there is some data about the cluster stored in /var/<somewhere> | 13:32 |
| jrosser | and there is a uuid that represents the cluster | 13:32 |
| jrosser | as the error is "is either already part of another cluster" i might look at that first | 13:33 |
| jrosser | tbh i expect that the playbook just can't handle this situation | 13:34 |
| jrosser | but it doesnt look too difficult https://www.cyberciti.biz/faq/howto-add-new-brick-to-existing-glusterfs-replicated-volume/ | 13:34 |
| *** ysandeep is now known as ysandeep|afk | 13:37 | |
| mgariepy | ok i'll take a look. | 13:38 |
| jrosser | mgariepy: you also need to be careful googling for how to do gluster stuff because it's changed over time in the later versions | 13:38 |
| mgariepy | well it's AIO ahha | 13:38 |
| jrosser | oh well i mean the commands may have "modern" and old ways to do things | 13:39 |
| jrosser | it took a lot of effort to find proper docs for this https://github.com/openstack/openstack-ansible-plugins/blob/master/roles/glusterfs/tasks/main.yml#L129-L141 | 13:39 |
| mgariepy | it was only missing the correct peer probe i think | 13:40 |
| jrosser | right - looking at the code it feels like it might be really close already | 13:40 |
| mgariepy | yeah | 13:40 |
| mgariepy | needs to do peer probe from the working node intead of the other. | 13:41 |
| mgariepy | not too far :D | 13:41 |
| jrosser | delegate_to: <any-host-except-this-one> | 13:41 |
| jrosser | ^ hmm | 13:41 |
| mgariepy | haha yep | 13:42 |
| mgariepy | it needs a host already in the cluster to probe the other. | 13:42 |
| mgariepy | thanks for your help. i'll patchup the role to fix this usecase | 13:43 |
| jrosser | excellent - no problem | 13:44 |
| jrosser | you can probably find that reliably in the output of https://github.com/openstack/openstack-ansible-plugins/blob/master/roles/glusterfs/tasks/main.yml#L122-L127 | 13:44 |
| *** ysandeep|afk is now known as ysandeep | 14:13 | |
| mgariepy | jrosser, was there a reason why you used the last_play_host? | 15:03 |
| mgariepy | since all the tasks are ran on all node one by one, the first one would have been quite equivalent | 15:03 |
| jrosser | are they run one by one? | 15:04 |
| jrosser | i think its not serial | 15:04 |
| mgariepy | they run in parrallel i thing | 15:04 |
| mgariepy | think* | 15:04 |
| mgariepy | my point being that when executing the task that add gluster peers they are all at the same step./ | 15:06 |
| mgariepy | either one would be ok to use at this point. you only want the ones with the service up. | 15:07 |
| mgariepy | tasks are run on each node one tasks a the time** | 15:10 |
| opendevreview | Marc Gariépy proposed openstack/openstack-ansible-plugins master: Fix gluster play_hosts https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/850540 | 15:14 |
| mgariepy | jrosser, something like this ^^ | 15:14 |
| mgariepy | this way i don't have to parse the output of the commands and it works to grow from 1 to 3 in affinity, i guess it would also work if you scale up your infra servers from 1 to 3. since the inventory would probably keep the order they were added. | 15:20 |
| jrosser | oh sorry i'm confusing what you said there | 15:27 |
| jrosser | i suspect what happened here was i was originally writing the role when the gluster deployment happened in the same play as the rest of the repo server deployment | 15:27 |
| jrosser | which is always serial=1 | 15:27 |
| jrosser | but now it's not like that at all | 15:28 |
| jrosser | so it should not matter at all if it's first or last host in the play, you are right | 15:28 |
| mgariepy | would it matter on switching from not gluster to gluster? | 15:29 |
| jrosser | i would guess not - as the gluster stuff is done as one play before the repo stuff starts another play | 15:51 |
| jrosser | unfortunatley the upgrade job that tests that will be on stable/yoga | 15:52 |
| mgariepy | we should backport anyway :) | 15:52 |
| jrosser | jamesdenton: have you ever used networking-generic-switch with nxos? | 15:55 |
| jamesdenton | i think so, but been a while. actively using with a catalyst | 15:55 |
| jamesdenton | now that i think about it, might've only been arista and not nexus | 15:56 |
| jamesdenton | but either way, i think cisco driver will handle nxos | 15:56 |
| jrosser | yeah, i was looking at the lldp autodiscovery | 15:58 |
| jrosser | and that looks kind of nice for auto-enrolment | 15:58 |
| jrosser | however /o\ nxos gives a unique chassis id per port which is super unhelpful | 15:59 |
| jrosser | so setting ngs_mac_address isnt possible https://docs.openstack.org/networking-generic-switch/latest/configuration.html | 16:00 |
| jamesdenton | i hadn't seen anything about lldp auto discovery | 16:02 |
| jrosser | i think that ngs_mac_address lets you set the switch mac address, which inspector then finds and reports back | 16:03 |
| jrosser | so you don't have to initially configure that at all when making a new node | 16:03 |
| jamesdenton | ahhh gotcha | 16:04 |
| jrosser | or rather, if you've defned a bunch of switches up front by their mac address, it can select the right one based on the info from inspector | 16:04 |
| *** dviroel is now known as dviroel|lunch | 16:06 | |
| jamesdenton | as in, it would auto-populate local-link-connection based on lldp info during discovery | 16:06 |
| admin1 | noonedeadpunk, cherry pick of that patch worked perfectly for the mysql upgrade, and the issue of not enough hosts is gone | 16:06 |
| jamesdenton | i haven't looked at that; we haven't used inspector much up to this point | 16:06 |
| jrosser | jamesdenton: yes i think thats what it should do | 16:09 |
| jamesdenton | seems like there could be a chicken/egg situation there, especially if you're using NGS to tweak the VLAN and drop it into the 'inspection' vlan. the baremetal port would need to have those switchport properties defined already (i would think) | 16:11 |
| *** ysandeep is now known as ysandeep|out | 16:11 | |
| jrosser | well - lldp should work regardless of the VLAN configuration i think? | 16:13 |
| jrosser | it's a different protocol so kind of 'out of band' | 16:14 |
| jamesdenton | true, but you still need to be in the PXE vlan | 16:14 |
| jrosser | hmm yes | 16:15 |
| jamesdenton | FWIW, this Cisco catalyst also has hardware MAC per switchport, but the switch itself has a base MAC identified in 'show ver'. That's the one I use; I need to see what LLDP returns | 16:18 |
| jrosser | well this is super early days - we are just starting to mess with this | 16:18 |
| jrosser | `sudo tcpdump -i ens1f0 ether proto 0x88cc -vv` | 16:18 |
| jrosser | compare that between two hosts | 16:18 |
| anskiy | jrosser: is cisco's proprietary LLDP known as CDP dead? Or you're just not all cisco? | 16:20 |
| jrosser | it has both LLDP and CDP | 16:21 |
| jamesdenton | @jrosser Looking at a 9372PX here, I see what you mean | 16:49 |
| jamesdenton | chassis id is unique per port. | 16:50 |
| jrosser | it’s odd isn’t it | 16:50 |
| jamesdenton | I could see that being problematic if you use inspector to auto-populate that | 16:50 |
| jamesdenton | yes, seems odd | 16:52 |
| jamesdenton | https://www.cisco.com/c/en/us/td/docs/dcn/nx-os/nexus9000/102x/configuration/system-management/cisco-nexus-9000-series-nx-os-system-management-configuration-guide-102x/m-configuring-lldp-10x.html | 16:53 |
| jamesdenton | Beginning with Cisco NX-OS Release 10.2(3)F, the Correctly Advertise LLDP Chassis-ID feature introduces a new global configuration command—lldp chassis-id switch —to advertise the switch chassis MAC address instead of the port MAC address, that is, all the ports will publish only the MAC address of the switch chassis. This feature is supported on all Cisco Nexus 9000 series platform switches. | 16:54 |
| jamesdenton | there you go | 16:54 |
| jamesdenton | Just Upgrade™ :D | 16:54 |
| jrosser | argh | 17:10 |
| jrosser | what they actually meant to say is "This feature is supported on all Cisco Nexus 9000 series platform switches which can run nxos 10.x" | 17:11 |
| jrosser | i wonder if thats fixed in a newer 9.x | 17:12 |
| *** dviroel|lunch is now known as dviroel | 17:23 | |
| *** dviroel is now known as dviroel|afk | 19:34 | |
| opendevreview | Merged openstack/openstack-ansible-os_neutron stable/xena: Allow to provide custom configuration for VPNaaS https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/848066 | 21:01 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!