| *** b1tsh1ft3r has joined #openstack-ansible | 00:39 | |
| *** b1tsh1ft3r has quit IRC | 00:53 | |
| *** b1tsh1ft3r has joined #openstack-ansible | 02:16 | |
| *** b1tsh1ft_ has joined #openstack-ansible | 02:20 | |
| *** b1tsh1ft3r has quit IRC | 02:24 | |
| *** spatel has joined #openstack-ansible | 02:26 | |
| *** frenzy_friday has quit IRC | 02:54 | |
| *** frenzy_friday has joined #openstack-ansible | 02:54 | |
| *** furkance has quit IRC | 03:32 | |
| *** furkance has joined #openstack-ansible | 03:33 | |
| *** spatel has quit IRC | 03:38 | |
| *** spatel has joined #openstack-ansible | 03:41 | |
| *** spatel has quit IRC | 03:43 | |
| *** b1tsh1ft_ has quit IRC | 04:11 | |
| *** b1tsh1ft3r has joined #openstack-ansible | 04:42 | |
| *** b1tsh1ft3r has quit IRC | 04:50 | |
| *** gokhani has joined #openstack-ansible | 05:59 | |
| *** gokhani has quit IRC | 06:03 | |
| *** gokhani has joined #openstack-ansible | 06:03 | |
| *** luksky has joined #openstack-ansible | 06:13 | |
| noonedeadpunk | mornings | 07:13 |
|---|---|---|
| noonedeadpunk | it seems I missed all fun | 07:16 |
| *** andrewbonney has joined #openstack-ansible | 07:18 | |
| *** tosky has joined #openstack-ansible | 07:20 | |
| *** rpittau|afk is now known as rpittau | 07:34 | |
| jrosser | noonedeadpunk: yes, looks like another upgrade gone strange | 07:42 |
| noonedeadpunk | I hope we will be able to figure out at least used version... | 07:43 |
| jrosser | and try to unpick if its no repo server / stale facts / not upgraded controllers first / ... | 07:43 |
| noonedeadpunk | explicitly disabled building wheels | 07:44 |
| jrosser | right | 07:45 |
| jrosser | i still do wonder if we need to add some sort of circuit breaker behaviour that requires a -e break_the_opendev_servers_for_everyone=yes to avoid | 07:45 |
| jrosser | from the chat in #opendev it was failing on the nova repo which is massive, and sounded like compute nodes | 07:46 |
| noonedeadpunk | not sure tbh. and even if we add, it will be effective in years for ppl that face this right now | 07:47 |
| noonedeadpunk | but yeah, we might add some logic to python_venv_build and check for length of play hosts | 07:48 |
| noonedeadpunk | and fail if wheels are not going to be used or smth | 07:49 |
| noonedeadpunk | I wonder if it can be side effect of broken venv_rebuild and the way ppl workarounded it | 07:49 |
| jrosser | potentially yes, and if the fix for that is not in their upgrade then the behaviour will be like that forever | 07:51 |
| noonedeadpunk | btw we have never merged "fix" of venv_rebuild to U | 07:53 |
| noonedeadpunk | https://review.opendev.org/c/openstack/ansible-role-python_venv_build/+/773971 | 07:53 |
| jrosser | oh yes that, nginx there is totally just for the functional test | 07:59 |
| jrosser | was questioning again why that wasnt apache instead given it would be straight from the distro packages | 08:00 |
| noonedeadpunk | not to define different package name? hahaha | 08:02 |
| noonedeadpunk | I think because we also use nginx for repo container.... | 08:03 |
| noonedeadpunk | should be easy fix though | 08:05 |
| opendevreview | likui proposed openstack/openstack-ansible-os_panko master: Deprecate support for Panko https://review.opendev.org/c/openstack/openstack-ansible-os_panko/+/794457 | 08:06 |
| jrosser | noonedeadpunk: there is one obvious difference though | 08:07 |
| jrosser | this https://github.com/openstack/openstack-ansible-repo_server/blob/master/defaults/main.yml#L30-L32 | 08:07 |
| jrosser | vs. https://github.com/openstack/ansible-role-python_venv_build/blob/master/tests/test.yml#L51 | 08:07 |
| noonedeadpunk | yeah... | 08:08 |
| noonedeadpunk | to be fair, we also install epel there as well | 08:08 |
| jrosser | but only for lsyncd i think | 08:09 |
| noonedeadpunk | oh, hm.. yeah | 08:11 |
| * noonedeadpunk already launched functional test locally | 08:11 | |
| opendevreview | Jonathan Rosser proposed openstack/ansible-role-python_venv_build stable/ussuri: Do not drop all wheels with venv_rebuild https://review.opendev.org/c/openstack/ansible-role-python_venv_build/+/773971 | 08:16 |
| jrosser | ^ that uses the same repo for nginx as repo_server | 08:16 |
| opendevreview | likui proposed openstack/openstack-ansible-os_placement master: Update index.rst for stable/victoria https://review.opendev.org/c/openstack/openstack-ansible-os_placement/+/794462 | 08:17 |
| jrosser | i guess part of the trouble here is that the fix we made is good enough to prevent venv_rebuild=true causing a problem | 08:21 |
| jrosser | but it won't rescue a repo_server which is already in the broken state | 08:21 |
| noonedeadpunk | yeah, it won't.... | 08:21 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: DNM: exercise base-test in ansible deploy jobs https://review.opendev.org/c/openstack/openstack-ansible/+/794332 | 08:23 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: DNM: exercise base-test in ansible deploy jobs https://review.opendev.org/c/openstack/openstack-ansible/+/794332 | 08:24 |
| opendevreview | likui proposed openstack/openstack-ansible-os_placement master: These should be indented four spaces to match the other things in this block https://review.opendev.org/c/openstack/openstack-ansible-os_placement/+/794464 | 08:24 |
| noonedeadpunk | yep, it;s working (in regards to nginx | 08:25 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-rsyslog_client master: Add support for debian bullseye https://review.opendev.org/c/openstack/openstack-ansible-rsyslog_client/+/791288 | 08:29 |
| noonedeadpunk | jrosser: any point for -W in https://review.opendev.org/c/openstack/ansible-role-uwsgi/+/791290 ? | 08:29 |
| jrosser | no - i think other than it was maybe one of the first bullseye patches i made and it couldnt be shown to be OK without a bunch of others | 08:31 |
| *** luksky has quit IRC | 08:31 | |
| jrosser | which we now have | 08:31 |
| noonedeadpunk | ok, fair | 08:32 |
| *** luksky has joined #openstack-ansible | 08:32 | |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Add PKI role to integrated repo https://review.opendev.org/c/openstack/openstack-ansible/+/788031 | 08:38 |
| opendevreview | Merged openstack/ansible-hardening master: Switch hardening to integrated tests https://review.opendev.org/c/openstack/ansible-hardening/+/792640 | 09:06 |
| *** odyssey4me has quit IRC | 09:51 | |
| opendevreview | Merged openstack/openstack-ansible-os_placement master: These should be indented four spaces to match the other things in this block https://review.opendev.org/c/openstack/openstack-ansible-os_placement/+/794464 | 09:54 |
| opendevreview | Merged openstack/openstack-ansible master: setup.cfg: Replace dashes with underscores https://review.opendev.org/c/openstack/openstack-ansible/+/788312 | 09:57 |
| *** gokhani has quit IRC | 10:03 | |
| *** admin1 has joined #openstack-ansible | 10:06 | |
| *** odyssey4me has joined #openstack-ansible | 10:10 | |
| *** gilou_ has joined #openstack-ansible | 10:13 | |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible stable/victoria: Fix sahara backend URL https://review.opendev.org/c/openstack/openstack-ansible/+/794444 | 10:13 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible stable/ussuri: Fix sahara backend URL https://review.opendev.org/c/openstack/openstack-ansible/+/794445 | 10:14 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible stable/ussuri: Fix sahara backend URL https://review.opendev.org/c/openstack/openstack-ansible/+/794445 | 10:16 |
| *** gokhani has joined #openstack-ansible | 10:17 | |
| *** Gilou has quit IRC | 10:20 | |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-openstack_hosts master: Use PKI role to install CA certificates https://review.opendev.org/c/openstack/openstack-ansible-openstack_hosts/+/790431 | 10:29 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_manila master: Updated from OpenStack Ansible Tests https://review.opendev.org/c/openstack/openstack-ansible-os_manila/+/780362 | 10:46 |
| *** gokhani has quit IRC | 11:00 | |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-openstack_hosts master: Use PKI role to install CA certificates https://review.opendev.org/c/openstack/openstack-ansible-openstack_hosts/+/790431 | 11:02 |
| *** gokhani has joined #openstack-ansible | 11:02 | |
| noonedeadpunk | Feels like we can merge https://review.opendev.org/q/topic:%22osa%252Fbullseye%22+(status:open) and https://review.opendev.org/q/topic:%22osa%252Fpki%22+(status:open) and do RC1 this week! | 11:03 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-openstack_hosts master: Use PKI role to install CA certificates https://review.opendev.org/c/openstack/openstack-ansible-openstack_hosts/+/790431 | 11:06 |
| jrosser | ^ fixed the tags | 11:06 |
| noonedeadpunk | I think both things might be left | 11:07 |
| noonedeadpunk | but not sure what makes sense here | 11:07 |
| noonedeadpunk | always is must have for sure though | 11:07 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/ansible-config_template master: Use ansible_facts[] instead of fact variables https://review.opendev.org/c/openstack/ansible-config_template/+/780753 | 11:10 |
| *** halali_ has quit IRC | 11:11 | |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Disable fact variables https://review.opendev.org/c/openstack/openstack-ansible/+/778396 | 11:12 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_ceilometer master: Fix u-c filter regex https://review.opendev.org/c/openstack/openstack-ansible-os_ceilometer/+/783812 | 11:13 |
| opendevreview | Merged openstack/openstack-ansible-haproxy_server master: Update documentation for keepalived ping addresses https://review.opendev.org/c/openstack/openstack-ansible-haproxy_server/+/782375 | 11:25 |
| *** halali_ has joined #openstack-ansible | 11:32 | |
| *** halali_ has quit IRC | 11:40 | |
| *** akahat has joined #openstack-ansible | 11:48 | |
| *** ioni has joined #openstack-ansible | 11:52 | |
| ioni | hi guys | 11:55 |
| ioni | does anyone have example on how to configure network on centos 7? | 11:55 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-tests stable/victoria: Add is_metal variable https://review.opendev.org/c/openstack/openstack-ansible-tests/+/794450 | 11:58 |
| noonedeadpunk | sorry, don't have any centos7 handy nowadays :( | 12:00 |
| admin1 | ioni, is it for bridge or ovs or ovn ? | 12:02 |
| ioni | bridge | 12:04 |
| admin1 | its fairly easy .. copy the eth0/ens0 file to br-vlan etc and modify ..and then put BRIDGE=br-mgmt etc in the interface files | 12:05 |
| ioni | done that | 12:06 |
| ioni | but it tells me that I need to add an ip on bond0, bond0.10 | 12:06 |
| ioni | but the ip is on br-mgmt | 12:06 |
| admin1 | what tells you to add those ips ? | 12:06 |
| ioni | bond0 bond0.10 doesn't come up | 12:06 |
| ioni | ifcfg ? | 12:07 |
| noonedeadpunk | if interfaces are part of bond there should be no ips | 12:07 |
| admin1 | what tells you add the ips != ifcfg .. valid answers are => "the voice in my head", "ansible documentation" ,"some guy who said it should be like that" etc | 12:07 |
| ioni | https://paste.xinu.at/6n6sN/ | 12:07 |
| ioni | noonedeadpunk, indeed but it fails to come up with that error :) | 12:08 |
| admin1 | so you have 1 bond and you want everything to be under that bond ? | 12:08 |
| ioni | i know i shouldn't have ips on those, i'm running OSA for 5 years on ubuntu | 12:08 |
| ioni | admin1, yes | 12:09 |
| admin1 | if you have bond0 and you want to use bond0 for all, then you need 3 or 4 vlan tags .. for br-mgmt, br-vxlan, br-storage ( and maybe br-ssh) .. and then the bond0 itself will be under br-vlan | 12:09 |
| admin1 | if the port is hybrid and carries a vlan tag, then bond0 itself can have an ip address to be used for ssh | 12:10 |
| admin1 | so in your case, are these already done and are the interfaces up and in the correct bridges | 12:10 |
| admin1 | once those are up, then we move to configuring the ip part | 12:10 |
| ioni | ok, have you seen the screenshot? | 12:11 |
| ioni | admin1, again, i know how to layout should be, i've done it 1000 times on ubuntu | 12:11 |
| mgariepy | what is are interface configuration ? | 12:12 |
| admin1 | the screenshot is not an output of nmcli or brctl show status to validate that | 12:12 |
| mgariepy | like what interface are part of the bound ? | 12:12 |
| admin1 | ioni, ok .. in the bond0, you need to put BRIDGE=br-vlan | 12:12 |
| admin1 | but there is no br-vlan in the screenshot | 12:12 |
| ioni | https://paste.xinu.at/s7mL/ | 12:12 |
| mgariepy | `cat /proc/net/bonding/bond0` | 12:13 |
| mgariepy | is the bonding module loaded ? | 12:13 |
| admin1 | ioni, modinfo 8021q | 12:13 |
| *** halali_ has joined #openstack-ansible | 12:13 | |
| ioni | i loaded it now. | 12:14 |
| ioni | seems that centos 7 is not really a modern os | 12:15 |
| mgariepy | 2014.. | 12:15 |
| mgariepy | why centos btw? | 12:16 |
| admin1 | it should work now | 12:17 |
| admin1 | config looks fine | 12:17 |
| admin1 | cat /proc/net/vlan/config will show if the tags are added and in what interface | 12:17 |
| noonedeadpunk | And it has only maintenance updates atm | 12:18 |
| ioni | i'm trying to get rid on some legacy tehnology and the "new one" works only centos 7 | 12:19 |
| admin1 | ioni, did it worked after the module is loaded ? | 12:20 |
| ioni | admin1, well, at least now i see some errors | 12:21 |
| ioni | brctl not available :) | 12:21 |
| admin1 | you have to go to rpmfind.net, search for the bridge-utils rpm and download and install | 12:21 |
| admin1 | maybe its in epel .. | 12:21 |
| admin1 | have been a while since i have used centos .. its all ubuntu/debian now | 12:21 |
| ioni | i'll manage it somehow | 12:22 |
| admin1 | ioni => rpm -ivvh https://rpmfind.net/linux/centos/7.9.2009/os/x86_64/Packages/bridge-utils-1.5-9.el7.x86_64.rpm | 12:23 |
| ioni | yum install bridge-utils worked | 12:24 |
| opendevreview | Andrew Bonney proposed openstack/openstack-ansible master: Add PKI role to integrated repo https://review.opendev.org/c/openstack/openstack-ansible/+/788031 | 12:34 |
| opendevreview | Merged openstack/openstack-ansible-openstack_hosts master: Use appropriate osbpo repo for Bullseye https://review.opendev.org/c/openstack/openstack-ansible-openstack_hosts/+/794161 | 12:37 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Add PKI role to integrated repo https://review.opendev.org/c/openstack/openstack-ansible/+/788031 | 12:38 |
| *** gokhani has quit IRC | 12:47 | |
| opendevreview | Merged openstack/openstack-ansible-lxc_container_create master: Add Debian Bullseye support https://review.opendev.org/c/openstack/openstack-ansible-lxc_container_create/+/793896 | 12:49 |
| admin1 | ioni, all working fine now ? | 12:54 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-tests stable/victoria: Workaround linting issue https://review.opendev.org/c/openstack/openstack-ansible-tests/+/792906 | 12:57 |
| ioni | admin1, yes. bond0 and all interfaces are coming up as expected. looking into bond1 | 12:58 |
| *** spatel has joined #openstack-ansible | 12:59 | |
| opendevreview | Merged openstack/openstack-ansible-lxc_container_create master: Add upgrade jobs https://review.opendev.org/c/openstack/openstack-ansible-lxc_container_create/+/794256 | 13:47 |
| *** odyssey4me has quit IRC | 13:49 | |
| opendevreview | Merged openstack/openstack-ansible-lxc_hosts master: Add upgrade jobs https://review.opendev.org/c/openstack/openstack-ansible-lxc_hosts/+/794255 | 13:52 |
| admin1 | https://thelounge.chat/ is nice ( i am using now ) . no more znc or bouncers . | 13:53 |
| *** odyssey4me has joined #openstack-ansible | 14:00 | |
| *** lowercase has joined #openstack-ansible | 14:01 | |
| noonedeadpunk | looks suuuper nice indeed | 14:05 |
| jrosser | noonedeadpunk: do you have a moment to look at the upgrade trouble lowercase was having yesterday? i have a meeting unfortunatley just now | 14:06 |
| noonedeadpunk | sure | 14:06 |
| noonedeadpunk | btw, regarding policy issue being reported - http://paste.openstack.org/show/806326/ | 14:07 |
| opendevreview | Merged openstack/openstack-ansible-os_panko master: Fix wsgi_venv path for panko-api https://review.opendev.org/c/openstack/openstack-ansible-os_panko/+/794195 | 14:07 |
| noonedeadpunk | (it's pure aio, so single container/node) | 14:07 |
| noonedeadpunk | and it randomly fails. Which doesn't happen with yaml btw | 14:07 |
| mgariepy | not all the thread refreshed ? | 14:08 |
| noonedeadpunk | feels like this, yes... | 14:09 |
| noonedeadpunk | but um... | 14:09 |
| noonedeadpunk | why only for json then... | 14:11 |
| mgariepy | if you restart keystone. | 14:18 |
| mgariepy | does it do the same thing? | 14:19 |
| mgariepy | err.. nova.. lol | 14:20 |
| mgariepy | just to see if it's a dynamic loading issue or another issue. | 14:21 |
| spatel | noonedeadpunk in ansible i have to use shell because command module doesn't support complex operation | 14:23 |
| spatel | I tried command and playbook failed but it works with shell: module | 14:23 |
| noonedeadpunk | mgariepy: haven't tried tbh with other service | 14:24 |
| noonedeadpunk | also to add context, it's about https://bugs.launchpad.net/openstack-ansible/+bug/1930276 | 14:24 |
| opendevmeet | Launchpad bug 1930276 in openstack-ansible "Nova API not restarted when nova policy is updated" [Undecided,Triaged] - Assigned to Dmitriy Rabotyagov (noonedeadpunk) | 14:24 |
| noonedeadpunk | spatel: you haven't used anything complex where I commented | 14:24 |
| spatel | shell: ovs-appctl -t /var/run/ovn/ovnnb_db.ctl cluster/status OVN_Northbound | sed 's/ //g' | grep -oP '(?<=Role:).*' | 14:25 |
| spatel | this line doesn't work with command: module | 14:25 |
| noonedeadpunk | but I commented L20 | 14:25 |
| noonedeadpunk | not L39 | 14:25 |
| spatel | :) | 14:25 |
| spatel | you are goddam right... i thought you prefer command over shell | 14:26 |
| spatel | i was overthinking.. sorry | 14:26 |
| spatel | all good then, let me check my handler part and then we i can resolve these changes | 14:27 |
| spatel | noonedeadpunk what is the deal with config_template ? | 14:27 |
| spatel | what is the difference between template and config_template ? | 14:28 |
| noonedeadpunk | because you can override anything you want with config_template? | 14:29 |
| noonedeadpunk | like we do with all config overrides | 14:29 |
| noonedeadpunk | I prefer command over shell when it's possible to use command :) | 14:30 |
| noonedeadpunk | And it's not like me, but more like ansible-lint is also thinking | 14:30 |
| noonedeadpunk | https://github.com/ansible-community/ansible-lint/blob/master/src/ansiblelint/rules/UseCommandInsteadOfShellRule.py#L102-L106 | 14:31 |
| spatel | Thank you for explanation. | 14:33 |
| mgariepy | noonedeadpunk, added a comment on https://review.opendev.org/c/openstack/openstack-ansible/+/783606 | 14:42 |
| opendevreview | Merged openstack/openstack-ansible-os_neutron master: Use neutron_conf_dir for absent policy removal https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/788499 | 14:43 |
| opendevreview | Merged openstack/openstack-ansible master: Gather minimal facts in CI https://review.opendev.org/c/openstack/openstack-ansible/+/790042 | 14:51 |
| opendevreview | Merged openstack/openstack-ansible master: Don't collect virtual facts https://review.opendev.org/c/openstack/openstack-ansible/+/789926 | 14:52 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Add Debian Bullseye support https://review.opendev.org/c/openstack/openstack-ansible/+/783606 | 15:18 |
| noonedeadpunk | jrosser: so it was 20.04 hosts and 18.04 controllers what caused the issue | 15:39 |
| jrosser | ahha | 15:40 |
| jrosser | "upgrade one of the controllers" is tricky too - it has to be the one running lsyncd i think | 15:40 |
| jrosser | i was thinking if it was possible to introduce just one host of the required OS and use that for wheel build | 15:41 |
| noonedeadpunk | and I have some ideas how to handle that | 15:42 |
| jrosser | in principle its easy, boot a host (or VM) and override the build host for a ansible group or smth | 15:43 |
| jrosser | but then the LB will not pick the one you want | 15:43 |
| noonedeadpunk | I mean overall situatuion during venv_build | 15:43 |
| * noonedeadpunk need to run away | 15:43 | |
| jrosser | ok | 15:43 |
| jrosser | i wonder if thats a threading issue for the oslo.policy bug | 15:51 |
| mgariepy | i do wonder if the restart fix the issue or not. | 16:06 |
| jrosser | i was wondering if reloading the policy only updated what was possibly local to one (some?) threads | 16:10 |
| jrosser | and restarting would guarantee that they all got the new policy | 16:11 |
| jrosser | fsvo threads though, it is python after all | 16:12 |
| *** rpittau is now known as rpittau|afk | 16:12 | |
| mgariepy | https://github.com/openstack/oslo.policy/blob/master/oslo_policy/policy.py#L417-L424 | 16:13 |
| mgariepy | not the same parser ! | 16:13 |
| *** sshnaidm is now known as sshnaidm|afk | 16:32 | |
| opendevreview | Merged openstack/ansible-role-uwsgi master: Add support for debian bullseye https://review.opendev.org/c/openstack/ansible-role-uwsgi/+/791290 | 16:39 |
| noonedeadpunk | mgariepy: restart fixes it, yes | 16:40 |
| noonedeadpunk | and that's good catch... | 16:41 |
| noonedeadpunk | jrosser: mgariepy: - would be great to merge that to unblock V functional tests https://review.opendev.org/c/openstack/openstack-ansible-tests/+/792906 | 16:44 |
| jrosser | done | 16:51 |
| noonedeadpunk | btw regarding pki - https://review.opendev.org/c/openstack/openstack-ansible-openstack_hosts/+/790431 fail during upgrade. I think it's because we don't have https://review.opendev.org/c/openstack/openstack-ansible/+/788031/10/tests/roles/bootstrap-host/files/user_variables_pki.yml when upgrading | 16:55 |
| noonedeadpunk | which means we should add some upgrade hook, or move that to defaults? | 16:56 |
| noonedeadpunk | I think we were discussing that some time ago...? | 16:56 |
| jrosser | oh right i think user_variables_pki needs to be removed completely to group_vars | 16:59 |
| jrosser | just which group_vars maybe a good question | 16:59 |
| noonedeadpunk | all/infra? | 17:00 |
| noonedeadpunk | oh! | 17:00 |
| noonedeadpunk | all/ssl | 17:00 |
| jrosser | easy answer is all | 17:00 |
| noonedeadpunk | https://opendev.org/openstack/openstack-ansible/src/branch/master/inventory/group_vars/all/ssl.yml | 17:00 |
| noonedeadpunk | super proper place :) | 17:00 |
| jrosser | but i was thinking that maybe the variables only actually do anything on the host that makes the certs | 17:00 |
| jrosser | however i was unsure of that | 17:01 |
| noonedeadpunk | we don't really delegate everything... | 17:03 |
| noonedeadpunk | I think we can add to all and move later on | 17:03 |
| noonedeadpunk | Also would be great to have https://review.opendev.org/c/openstack/openstack-ansible/+/769974 | 17:04 |
| noonedeadpunk | (offtop_ | 17:04 |
| noonedeadpunk | but we will less mess up things and more clear view where does things come from | 17:04 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Add PKI role to integrated repo https://review.opendev.org/c/openstack/openstack-ansible/+/788031 | 17:06 |
| opendevreview | Satish Patel proposed openstack/openstack-ansible-os_neutron master: Add ovn clustering support https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/794266 | 18:18 |
| opendevreview | Satish Patel proposed openstack/openstack-ansible-os_neutron master: Add ovn clustering support https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/794266 | 18:19 |
| opendevreview | Satish Patel proposed openstack/openstack-ansible-os_neutron master: Add ovn clustering support https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/794266 | 18:22 |
| opendevreview | Merged openstack/openstack-ansible-os_ceilometer master: Fix u-c filter regex https://review.opendev.org/c/openstack/openstack-ansible-os_ceilometer/+/783812 | 18:51 |
| *** dmsimard4 has joined #openstack-ansible | 19:14 | |
| *** dmsimard has quit IRC | 19:20 | |
| *** dmsimard4 is now known as dmsimard | 19:20 | |
| *** lowercase has quit IRC | 19:26 | |
| *** lowercase has joined #openstack-ansible | 19:29 | |
| opendevreview | Merged openstack/openstack-ansible master: Cleanup after service variables merged https://review.opendev.org/c/openstack/openstack-ansible/+/769974 | 19:39 |
| *** djhankb has joined #openstack-ansible | 19:55 | |
| *** spotz has quit IRC | 20:24 | |
| *** lowercase has quit IRC | 20:28 | |
| *** spotz has joined #openstack-ansible | 20:47 | |
| spatel | Folks, I have question | 20:55 |
| spatel | I want to build nginx or haproxy loadbalancer which support 1 million connection | 20:56 |
| spatel | my load-testing saying i running out of local port range so i need to add more IPs in backend | 20:57 |
| spatel | but found linux kernel doesn't do load-balancing with multiple IPs | 20:59 |
| *** andrewbonney has quit IRC | 21:14 | |
| admin1 | for 1 million connects, you use dns to round-robin the connections to multiple servers | 21:20 |
| jrosser | spatel: are you running out of ports between haproxy and the backends? | 21:22 |
| spatel | yes | 21:23 |
| spatel | i have 3 web server | 21:23 |
| jrosser | you can specify source <ip> for each server in the backend | 21:24 |
| admin1 | you can try to do ip_local_port_range from 2000 : 65000 | 21:24 |
| admin1 | so that you get a big range | 21:24 |
| spatel | jrosser something like this ? - https://www.irccloud.com/pastebin/0167bHKh/ | 21:24 |
| jrosser | https://stackoverflow.com/questions/26886641/haproxy-connect-to-backend-with-source-ip | 21:25 |
| spatel | interesting | 21:25 |
| admin1 | spatel, found an article for 2 million connections -> https://www.freecodecamp.org/news/how-we-fine-tuned-haproxy-to-achieve-2-000-000-concurrent-ssl-connections-d017e61a4d27/ | 21:25 |
| jrosser | depending on how many connections each backend can support you may even be able to have multiple IP on each backend | 21:26 |
| jrosser | and then list them as multiple server in the haproxy config each with a unique source / destination address | 21:26 |
| spatel | lets say i have 10 IPs in haproxy i can distribute them between 3 web right? like manual mapping | 21:29 |
| spatel | let me build haproxy in my lab and give it a try | 21:29 |
| jrosser | something like that, yes | 21:29 |
| jrosser | another thing to look as is using bgp and ecmp and just don’t use haproxy at all | 21:30 |
| jrosser | route-to-the-host | 21:30 |
| spatel | what about TLS/SSL termination :) | 21:30 |
| jrosser | well distribute that too :) don’t have a single bottleneck | 21:31 |
| jrosser | just saying there’s more than one approach here - funnelling the traffic through one thing isn’t the only way | 21:31 |
| admin1 | if your single haproxy goes wrong, you cause a 100% downtime for those million (clients) | 21:35 |
| admin1 | distributing to say 5-10 will minimize a total downtime | 21:36 |
| spatel | started building lab - https://ibb.co/tsBRDfq | 21:41 |
| spatel | my haproxy is in keepalive | 21:41 |
| jrosser | spatel: this is an excellent article (check out the whole blog too) https://vincent.bernat.ch/en/blog/2018-multi-tier-loadbalancer | 21:51 |
| spatel | wow! very nice work by that guy.. so much details | 21:52 |
| *** spatel has quit IRC | 22:15 | |
| *** tosky has quit IRC | 23:00 | |
| *** luksky has quit IRC | 23:07 | |
| *** furkance has quit IRC | 23:37 | |
| *** furkance has joined #openstack-ansible | 23:38 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!