| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-haproxy_server master: Update SHA for hatop download https://review.opendev.org/c/openstack/openstack-ansible-haproxy_server/+/872261 | 07:59 |
|---|---|---|
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-haproxy_server master: Update hatop to latest release, 0.8.2 https://review.opendev.org/c/openstack/openstack-ansible-haproxy_server/+/872262 | 08:03 |
| opendevreview | Jonathan Rosser proposed openstack/ansible-role-uwsgi master: Allow extra arguments to be passed to the uwsgi application through pyargv https://review.opendev.org/c/openstack/ansible-role-uwsgi/+/872195 | 08:04 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-os_neutron master: Ensure that all neutron config files are used when deployed with uwsgi https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/872199 | 08:04 |
| opendevreview | Jonathan Rosser proposed openstack/ansible-role-uwsgi master: Allow extra arguments to be passed to the uwsgi application through pyargv https://review.opendev.org/c/openstack/ansible-role-uwsgi/+/872195 | 08:06 |
| anskiy | noonedeadpunk: about mixed group (some hosts with `no_containers: True`): this doesn't seem to work. If I don't put `no_containers` hosts last, all other hosts would be with it implicitly set. And in /etc/hosts LXC hostname for the host without `no_containers` resolves to IP-address of physical host, not the container one (still not sure if this is gonna be a problem later, tho). I haven't yet had time to check the `osa_toolk | 08:42 |
| opendevreview | Jonathan Rosser proposed openstack/ansible-role-uwsgi master: Allow extra arguments to be passed to the uwsgi application through pyargv https://review.opendev.org/c/openstack/ansible-role-uwsgi/+/872195 | 08:51 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-os_neutron master: Ensure that all neutron config files are used when deployed with uwsgi https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/872199 | 08:51 |
| opendevreview | Jonathan Rosser proposed openstack/ansible-role-uwsgi master: Allow extra arguments to be passed to the uwsgi application through pyargv https://review.opendev.org/c/openstack/ansible-role-uwsgi/+/872195 | 10:11 |
| *** dviroel|out is now known as dviroel|rover | 11:29 | |
| ThiagoCMC | Morning! When I upload small RAW images to Glance (like Ubuntu 22.04 or 20.04), it works, however, when I try to upload a bigger image, like CentOS Stream 9 converted to RAW, it fails, it throws "Internal Server Error 500". Glance's default is 1T, so I'm guessing that this limitation is on haproxy side. How to increase it only for Glance in OSA inventory? | 11:31 |
| ThiagoCMC | For some reason, I tried a few times and ir worked! 10G image uploaded. lol | 11:49 |
| ThiagoCMC | Sorry about the buzz! | 11:49 |
| moha7 | I don't which boning mode I should choose! --> https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux/8/html/configuring_and_managing_networking/configuring-network-bonding_configuring-and-managing-networking#upstream-switch-configuration-depending-on-the-bonding-modes_configuring-network-bonding | 12:32 |
| moha7 | May you give me some hints for a production env! I'm going to bond 2 10G ports for Ceph access transmit and 2 other 10G port for VXLAN network | 12:33 |
| moha7 | ?* | 12:33 |
| moha7 | there are 6 modes | 12:35 |
| moha7 | 0 - balance-rr | 2 - balance-xor | 3 - broadcast | 4 - 802.3ad | 5 - balance-tlb | 6 - balance-alb | 12:36 |
| jrosser | moha7: i have mode 4 and also set transmit hashing to layer3+4 | 12:41 |
| jrosser | if you don't do the second part and your traffic on some interface is predominantly heading to a router gateway (for example) that has a constant mac address and all the traffic will end up on one side | 12:42 |
| jrosser | moha7: but like that redhat article says it really depends how your switch(es) are configured | 12:48 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-os_neutron master: Ensure that all neutron config files are used when deployed with uwsgi https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/872199 | 12:53 |
| admin1 | moha7: 4 - 802.3ad | 13:06 |
| admin1 | you need to config your switches for this as well | 13:06 |
| mgariepy | jrosser, https://github.com/orgs/community/discussions/45830 | 13:10 |
| mgariepy | they did revert the change :D | 13:11 |
| mgariepy | https://github.com/git/git/commit/4f4be00d302bc52d0d9d5a3d4738bb525066c710 | 13:22 |
| moha7 | admin1: Thanks | 13:27 |
| opendevreview | James Denton proposed openstack/openstack-ansible-os_nova master: Add authentication for [cinder] section of nova.conf https://review.opendev.org/c/openstack/openstack-ansible-os_nova/+/872279 | 13:53 |
| spatel | https://ubuntu.com/pro | 14:07 |
| jamesdenton | spatel time to move to debian | 14:40 |
| admin1 | debian is a good choice | 14:47 |
| spatel | if you start using it then i am in :) | 14:52 |
| *** dviroel|rover is now known as dviroel|rover|lunch | 15:08 | |
| admin1 | anyone tried any database as service using osa ? | 15:47 |
| admin1 | or magnum and deployed k8s successfully | 15:47 |
| opendevreview | Damian Dąbrowski proposed openstack/openstack-ansible master: Define some temporary vars for haproxy https://review.opendev.org/c/openstack/openstack-ansible/+/872328 | 16:09 |
| *** dviroel|rover|lunch is now known as dviroel|rover | 16:14 | |
| moha7 | spatel: ubuntu.com/pro --> it's been for a while but not under the name 'pro'; LTS is free for 5 years, isn't it? | 16:44 |
| opendevreview | Damian Dąbrowski proposed openstack/openstack-ansible master: Define some temporary vars for haproxy https://review.opendev.org/c/openstack/openstack-ansible/+/872328 | 16:59 |
| opendevreview | Damian Dąbrowski proposed openstack/openstack-ansible-haproxy_server master: Prepare haproxy role for separated haproxy config https://review.opendev.org/c/openstack/openstack-ansible-haproxy_server/+/871188 | 17:01 |
| opendevreview | Damian Dąbrowski proposed openstack/openstack-ansible-haproxy_server master: Prepare haproxy role for separated haproxy config https://review.opendev.org/c/openstack/openstack-ansible-haproxy_server/+/871188 | 17:03 |
| opendevreview | Damian Dąbrowski proposed openstack/openstack-ansible master: Define some temporary vars for haproxy https://review.opendev.org/c/openstack/openstack-ansible/+/872328 | 17:06 |
| opendevreview | Damian Dąbrowski proposed openstack/openstack-ansible master: Prepare service roles for separated haproxy config https://review.opendev.org/c/openstack/openstack-ansible/+/871189 | 17:06 |
| spatel | moha7 LTS has 5 year hardware and 5 more year for security | 17:07 |
| opendevreview | Damian Dąbrowski proposed openstack/openstack-ansible-specs master: Blueprint for separated haproxy service config https://review.opendev.org/c/openstack/openstack-ansible-specs/+/871187 | 17:22 |
| opendevreview | Damian Dąbrowski proposed openstack/openstack-ansible-specs master: Switch sphinx language to en https://review.opendev.org/c/openstack/openstack-ansible-specs/+/872335 | 17:22 |
| opendevreview | Damian Dąbrowski proposed openstack/openstack-ansible-specs master: Switch sphinx language to en https://review.opendev.org/c/openstack/openstack-ansible-specs/+/872335 | 17:49 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-os_neutron master: Ensure that all neutron config files are used when deployed with uwsgi https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/872199 | 17:50 |
| opendevreview | Damian Dąbrowski proposed openstack/openstack-ansible-specs master: Blueprint for separated haproxy service config https://review.opendev.org/c/openstack/openstack-ansible-specs/+/871187 | 18:08 |
| opendevreview | Damian Dąbrowski proposed openstack/openstack-ansible-specs master: Blueprint for separated haproxy service config https://review.opendev.org/c/openstack/openstack-ansible-specs/+/871187 | 18:37 |
| ElDuderino | Hi all, seeing some strange behavior and wonder if you’ve ever encountered it. We make use use HA l3_agents, and have a fairly standard build except we have the controllers and network nodes separated. I can crate thousands of vms, and they usually work, but I always have a fair amount that fail/error. If I push, say, 6 x 50 vm create requests I’ll get http errors. At times the ui/cli becomes completely non-re | 18:50 |
| ElDuderino | Looking at the neutron-server-container logs, I always see http failures (as soon as the create request is furnished) but for the most part, build requests will get through. ‘neutron.pecan_wsgi_hooks.transation POST’ failure messages, followed by neutron agents being down messages. | 18:51 |
| ElDuderino | Anyone else ever see that? We're on a dinosaur I know (Rocky) but I've been charged to push forward and 'fix' it. | 18:51 |
| ElDuderino | Looking into keepalived_ping_intervals - if perhaps we're flopping (I suspect that, as the http endpoints are at times unavailable). What is a 'default' setting? I see some RH documentation that used '10' for that, but IDK what makes sense. (https://docs.openstack.org/openstack-ansible-haproxy_server/rocky/configure-haproxy.html) | 18:58 |
| admin1 | ElDuderino, can you tunnel and access to check your queue in rabbitmq | 20:14 |
| admin1 | if you have a lot of queue on neutron, then your number of workers is not enough | 20:14 |
| ElDuderino | @admin1 will do. Thank you. I haven't tried to tunnel to check via browser, b/c I do everything through putty through multi-hops. since this is older rabbit, do you mean `rabbitmqctl list_vhosts` and then find the associated queue with those vhosts? | 20:42 |
| admin1 | check for the number of pending messages in the neutron queue | 20:43 |
| admin1 | i have worked on platforms that did 1000+ vms an hour .. and this used to be where we find bottlenecks | 20:44 |
| admin1 | and 3 neutron servers were not enough, so we just had to expand neutron server containers based on load | 20:44 |
| moha7 | Each new router that goes has a hand in the provider network uses a public IP which means at least one public IP per customer. Is there a solution to not pick a public IP up per tenant router and move all tenant traffic from one IP? | 20:58 |
| *** dviroel|rover is now known as dviroel|uot | 21:00 | |
| moha7 | Every new router that wants to provide the Internet for its tenant network, assigns an IP from the external subnet. Each customer probably creates at least one router, which means at least one public IP per customer; Is it possible to change this procedure so that the internet traffic of all routers passes through a public IP? | 21:05 |
| admin1 | moha7, https://www.openstackfaq.com/openstack-add-direct-attached-dhcp-ip/ | 21:19 |
| admin1 | you then do not need to create network and router | 21:20 |
| admin1 | just select the direct network, the IP will come directly | 21:20 |
| moha7 | it means eliminating the self-service network! | 21:24 |
| moha7 | FYI: During each new deploy, this usually happens: https://i.postimg.cc/90sBhfnK/Screenshot-2023-02-01-004924.png | 21:27 |
| jrosser | moha7: you asked about that before - the only way to find out what has failed there is to look in the log file on the repo host | 21:41 |
| jrosser | an example would be slow access again to opendev git servers, but really that can only be a guess without examining the log file | 21:42 |
| jrosser | moha7: if you want your tenants to have self service networks, multi-tenancy, and be able to accept incoming connections to their vm with floating IP then routers and their external ip are needed | 21:52 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-os_neutron master: Ensure that all neutron config files are used when deployed with uwsgi https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/872199 | 22:13 |
| moha7 | Oh, you're right; I forgot it! | 22:26 |
| admin1 | moha7, the networks can be side by side | 22:41 |
| admin1 | so you have one ext-direct and one ext-pub, where ext-direct = get ips directly, ext-pub = floating ip as normal | 22:41 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
