Tuesday, 2020-11-24

*** rfolco|ruck has quit IRC00:07
*** tosky has quit IRC00:17
*** spatel has joined #openstack-ansible01:06
*** spatel has quit IRC02:01
*** lemko has quit IRC04:22
*** lemko7 has joined #openstack-ansible04:22
*** evrardjp has quit IRC05:33
*** cshen has quit IRC05:33
*** evrardjp has joined #openstack-ansible05:33
*** shyamb has joined #openstack-ansible06:19
*** klamath_atx has joined #openstack-ansible06:32
*** NewJorg has quit IRC06:40
*** shyamb has quit IRC06:53
*** miloa has joined #openstack-ansible06:58
*** cshen has joined #openstack-ansible07:03
*** NewJorg has joined #openstack-ansible07:12
*** shyamb has joined #openstack-ansible07:31
*** spatel has joined #openstack-ansible07:31
*** spatel has quit IRC07:35
*** rpittau|afk is now known as rpittau07:39
*** pto has joined #openstack-ansible07:43
*** pto_ has joined #openstack-ansible07:52
*** pto has quit IRC07:54
*** luksky has joined #openstack-ansible07:58
*** shyamb has quit IRC08:00
*** jbadiapa has joined #openstack-ansible08:09
*** andrewbonney has joined #openstack-ansible08:42
*** tosky has joined #openstack-ansible08:48
jrossermorning08:54
admin0morning08:58
noonedeadpunko/09:09
*** openstackgerrit has joined #openstack-ansible09:15
openstackgerritDmitriy Rabotyagov proposed openstack/openstack-ansible-os_neutron master: Updated from OpenStack Ansible Tests  https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/75875109:15
noonedeadpunkjrosser: can I summon your vote on https://review.opendev.org/c/openstack/openstack-ansible/+/763048 to see if it solve octavia CI upgrade jobs failure on master?09:17
admin0anyone has the latest AIO for 21.2.0 up and running ? even after 10 rebuilds, the cinder volumes can be created, but not mounted09:27
admin0i want to check if it failed our QA process .. or is something broken in my end09:28
admin0the instance and volumes can be created fine .. but the volumes cannot be mounted to the instances09:28
admin0which beats the purpose ..09:28
jrossernoonedeadpunk: done09:29
jrossernoonedeadpunk: how much more do we have before we can do a V release-candidate?09:31
jrosserperhaps we shoul revisit the etherpad and make a new list09:32
*** shyamb has joined #openstack-ansible09:32
noonedeadpunkeah, probably. I think nothing too critical, except broken magnum because of upgrade jobs, and would be awesome to merge adjutant not to branch it in state that breaks integrated repo09:33
noonedeadpunkand, probably, swift as well...09:33
noonedeadpunkbut I'm about just to branch honestly09:34
jrosseroh yes adjutant is a giant patch but I guess the role is experimental so we should do a quick review, merge and iterate09:34
noonedeadpunkbtw, regarding https://review.opendev.org/c/openstack/openstack-ansible/+/734883/4/scripts/journal_dump.py - I think we won't actually need importing and depending on lxc here in case we did bind mounting right09:36
noonedeadpunksince we should be able to see all container journals on host itself09:37
openstackgerritDmitriy Rabotyagov proposed openstack/openstack-ansible-os_keystone master: Move openstack-ansible-uw_apache centos job to centos-8  https://review.opendev.org/c/openstack/openstack-ansible-os_keystone/+/75412209:41
openstackgerritDmitriy Rabotyagov proposed openstack/openstack-ansible-os_keystone master: Add openstack-ansible-uw_apache focal job  https://review.opendev.org/c/openstack/openstack-ansible-os_keystone/+/75412309:42
jrossernoonedeadpunk: just one comment that i can see on the adjutant patch09:44
noonedeadpunkI think you;re right09:45
openstackgerritDmitriy Rabotyagov proposed openstack/openstack-ansible-os_adjutant master: Make role fit to the OSA standards  https://review.opendev.org/c/openstack/openstack-ansible-os_adjutant/+/75631309:47
openstackgerritDmitriy Rabotyagov proposed openstack/openstack-ansible-os_adjutant master: Make role fit to the OSA standards  https://review.opendev.org/c/openstack/openstack-ansible-os_adjutant/+/75631309:48
*** shyamb has quit IRC09:52
*** shyamb has joined #openstack-ansible09:56
*** SiavashSardari has joined #openstack-ansible10:03
*** spatel has joined #openstack-ansible10:31
*** pto_ has quit IRC10:33
*** pto has joined #openstack-ansible10:33
*** spatel has quit IRC10:36
openstackgerritAndrew Bonney proposed openstack/openstack-ansible-os_zun master: DNM: Update zun role to match current requirements  https://review.opendev.org/c/openstack/openstack-ansible-os_zun/+/76314110:41
SiavashSardariHey guys, I have a question about openstack ansible collections, should I ask here or in the sig channel?10:53
*** macz_ has joined #openstack-ansible11:03
*** macz_ has quit IRC11:07
jrosserSiavashSardari: we are users of the openstack collection here in the openstack-ansible project, but for developer stuff about the collection itself that would be in #openstack-ansible-sig11:45
SiavashSardarijrosser thanks.11:47
openstackgerritMerged openstack/openstack-ansible-os_swift master: Updated from OpenStack Ansible Tests  https://review.opendev.org/c/openstack/openstack-ansible-os_swift/+/75875611:49
*** shyamb has quit IRC12:09
*** rfolco|ruck has joined #openstack-ansible12:11
*** gshippey has joined #openstack-ansible12:12
*** miloa has quit IRC12:13
openstackgerritJonathan Rosser proposed openstack/openstack-ansible master: Added Openstack Adjutant role deployment  https://review.opendev.org/c/openstack/openstack-ansible/+/75631012:48
admin0checking if anyone has an AIO of 21.2.0 to check if the volume created by the test can be mounted on the instance .. or any new volume created can be mounted on the instance for that matter ..12:58
admin0if it cannot be, then we need to add this test to the final checklist as well12:58
openstackgerritMerged openstack/openstack-ansible-os_neutron master: Fix ceilometer constraints  https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/76215113:06
jrosseradmin0: are you able to provide any debug info? like where it actually fails from the service logs?13:34
admin0i nuked it yet again13:35
admin0i will create a new one and provide that13:35
jrosserok thanks13:35
openstackgerritMerged openstack/openstack-ansible-os_barbican master: Cleanup stop handler and barbican_apache_* variables  https://review.opendev.org/c/openstack/openstack-ansible-os_barbican/+/75668913:38
*** dave-mccowan has joined #openstack-ansible13:38
openstackgerritMerged openstack/openstack-ansible master: Add magnum tempest URL  https://review.opendev.org/c/openstack/openstack-ansible/+/76304913:51
*** spatel has joined #openstack-ansible14:02
spatelI loved our old interface https://review.opendev.org/14:03
SiavashSardari+1 on old interface14:04
admin0if i add flat , is it also using br-vlan but without tags ( never used osa flat before)14:17
admin0flat network*14:17
spateladmin0: I think yes br-vlan would be on flat but without tags14:19
admin0thanks14:19
spateladmin0: Hey after disable port security my Trex was happy :)14:19
admin0:D14:19
admin0people forget this part when using trex14:19
spateladmin0: i totally missed that and didn't think about that because my security-group was all allow so i thought it should do that14:20
*** lkoranda has joined #openstack-ansible14:24
openstackgerritMerged openstack/openstack-ansible master: Add default simple key to secrets  https://review.opendev.org/c/openstack/openstack-ansible/+/75908214:24
*** pcaruana has quit IRC14:25
spateljamesdenton: do you have any specific test like to run using Trex?14:28
spatelcurrently i am running all 64k UDP packet test and with 8vCPU/8GB VM reaching 75kpps (with single core multi-queue )14:29
spatelbut if i add 4 core to multi-queue i am almost hitting 175kpps so multi-queue is totally boost performance :)14:30
spatelBRB14:31
*** spatel has quit IRC14:31
*** gouthamr_ has quit IRC14:36
*** SiavashSardari has quit IRC14:37
*** pcaruana has joined #openstack-ansible15:01
*** macz_ has joined #openstack-ansible15:03
*** macz_ has quit IRC15:03
*** macz_ has joined #openstack-ansible15:04
*** klamath_atx has quit IRC15:07
admin0Interface eth12 for physical network flat does not exist15:08
admin0i need a bit of help to understand osa flat15:11
admin0the interface is already on br-vlan ( ens6) from where i am supposed to send public traffic15:11
admin0but neutron dies with " Interface eth12 for physical network flat does not exist"15:12
admin0l3-agent dies with specifically15:12
admin0"neutron-linuxbridge-agent[62822]: 2020-11-24 16:14:16.813 62822 ERROR neutron.plugins.ml2.drivers.linuxbridge.agent.linuxbridge_neutron_agent [-] Interface eth12 for physical network flat does not exist. Agent terminated!"15:14
*** klamath_atx has joined #openstack-ansible15:15
admin0is there a command to check  what port/network a netns is plugged into ?15:24
admin0i see the namespace got qg-4f64bf74-86 which is under brq5296821f-42 .. but this is not under any br-vlan or interface15:27
*** nurdie has quit IRC15:38
jamesdentonunless you plan on using a flat network, i would just get rid of that whole block15:39
*** nurdie has joined #openstack-ansible15:39
admin0first time in almost 8 years of using osa that i have a need to use flat networking :D15:46
jamesdentonoh, hahaha15:46
admin0and i am failing to understand how this works15:46
jamesdentonare you using lxc still for containers?15:46
admin0yep15:46
jamesdentonor, rather, are your neutron agents in containers?15:46
jamesdentonok15:46
admin0but the agents are in metal on a network node15:46
admin0no containers15:46
jamesdentonoh alright15:46
admin0so wat i have is this .. i have  ens5 where is under br-vlan .. and also listed as interface for br-flat15:47
jamesdentonso with lxc, eth12 would have been a veth15:47
jamesdentonyeah, you can't do that15:47
admin0so what exactly do i need to do to use br-flat .. i need to remove br-vlan  ?15:48
jamesdentoni would use host_bind_override15:48
admin0let me pastebin my config15:48
jamesdentonwhich eliminates both br-vlan and br-flat, and uses the interface (ens5) directly15:48
admin0jamesdenton, https://gist.githubusercontent.com/a1git/2d0b68d8c5b14cb234f694d8a458ce2d/raw/fce1b56eec99783c8878f70fe2b406b7901ae3da/gistfile1.txt15:50
jamesdentonwhat happens if you just remove container_interface: "eth12"?15:52
admin0please note that in my setup, br-vlan has the enp5s0 interface15:54
admin0that is ok ?15:54
admin0i will remove and re-run15:54
openstackgerritwu.chunyang proposed openstack/openstack-ansible master: Dep's should be restricted by tox-constraints  https://review.opendev.org/c/openstack/openstack-ansible/+/76399515:54
openstackgerritwu.chunyang proposed openstack/openstack-ansible-apt_package_pinning master: Dep's should be restricted by tox-constraints  https://review.opendev.org/c/openstack/openstack-ansible-apt_package_pinning/+/76399615:55
openstackgerritwu.chunyang proposed openstack/openstack-ansible-ceph_client master: Dep's should be restricted by tox-constraints  https://review.opendev.org/c/openstack/openstack-ansible-ceph_client/+/76399815:56
openstackgerritwu.chunyang proposed openstack/openstack-ansible-galera_client master: Dep's should be restricted by tox-constraints  https://review.opendev.org/c/openstack/openstack-ansible-galera_client/+/76399915:57
openstackgerritwu.chunyang proposed openstack/openstack-ansible-haproxy_server master: Dep's should be restricted by tox-constraints  https://review.opendev.org/c/openstack/openstack-ansible-haproxy_server/+/76400016:00
noonedeadpunk#startmeeting openstack_ansible_meeting16:01
openstackMeeting started Tue Nov 24 16:01:07 2020 UTC and is due to finish in 60 minutes.  The chair is noonedeadpunk. Information about MeetBot at http://wiki.debian.org/MeetBot.16:01
openstackUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.16:01
*** openstack changes topic to " (Meeting topic: openstack_ansible_meeting)"16:01
openstackThe meeting name has been set to 'openstack_ansible_meeting'16:01
gshippeyo/16:01
noonedeadpunko/16:02
jamesdentono/16:02
openstackgerritwu.chunyang proposed openstack/openstack-ansible-nspawn_container_create master: Dep's should be restricted by tox-constraints  https://review.opendev.org/c/openstack/openstack-ansible-nspawn_container_create/+/76400116:03
noonedeadpunk#topic bug triage16:03
*** openstack changes topic to "bug triage (Meeting topic: openstack_ansible_meeting)"16:03
noonedeadpunkso I see 2 new bugs atm16:03
noonedeadpunkhttps://bugs.launchpad.net/openstack-ansible/+bug/190493516:03
openstackLaunchpad bug 1904935 in openstack-ansible "Get swift rings ssh conection" [Undecided,New]16:03
noonedeadpunkI think we should replace these nasty commands with synchronize module16:03
jrossero/ hello16:03
noonedeadpunkit should respect as ansible_ssh_port and ssh config16:04
noonedeadpunkand the second is rgw related https://bugs.launchpad.net/openstack-ansible/+bug/190517416:05
openstackLaunchpad bug 1905174 in openstack-ansible "Using radosgw as a drop-in replacement for Swift in openstack-ansible" [Undecided,New]16:05
jrosserhmm16:07
noonedeadpunkwell, I think we can add endpoints when rgw hosts are set or enabled?16:07
jrosserS3 will never appear in horizon16:07
openstackgerritwu.chunyang proposed openstack/openstack-ansible-nspawn_hosts master: Dep's should be restricted by tox-constraints  https://review.opendev.org/c/openstack/openstack-ansible-nspawn_hosts/+/76400216:07
jrosserthere is a terminology problem there16:07
jrosseryou need swift to make horizon work16:07
jrosseror rgw configured to serve swift16:08
noonedeadpunkwell yes, agree16:08
jrosserand also the change to the name with .rgw on the end looks suspect to me16:08
noonedeadpunkand I thought we have some variables to make this ahppen?16:08
jrosserbecasue  you may have more than one rgw process on the same host16:08
noonedeadpunkwell yes, that's what not very clear to me currentl;y16:09
openstackgerritwu.chunyang proposed openstack/openstack-ansible-openstack_hosts master: Dep's should be restricted by tox-constraints  https://review.opendev.org/c/openstack/openstack-ansible-openstack_hosts/+/76400716:09
jrosseralso these patches ^^16:09
noonedeadpunkbut it's indeed hwat we have in defaults https://opendev.org/openstack/openstack-ansible/src/branch/master/inventory/group_vars/ceph-rgw.yml#L316:09
admin0jamesdenton, didn't worked :(16:09
noonedeadpunkthese patches seem valid for me16:10
jamesdentonadmin0 ok - let's sync back up after this meeting16:10
admin0sure16:10
jrossernoonedeadpunk: are you sure it doesnt inherit install_command from [testenv] ?16:10
openstackgerritwu.chunyang proposed openstack/openstack-ansible-openstack_openrc master: Dep's should be restricted by tox-constraints  https://review.opendev.org/c/openstack/openstack-ansible-openstack_openrc/+/76400816:11
noonedeadpunkah16:11
noonedeadpunkwe did that weird way lol16:11
jrosseri'm fairly well -2 on these unless thats shown to be currently broken16:11
noonedeadpunkagree16:12
openstackgerritwu.chunyang proposed openstack/openstack-ansible-os_aodh master: Dep's should be restricted by tox-constraints  https://review.opendev.org/c/openstack/openstack-ansible-os_aodh/+/76400916:12
openstackgerritwu.chunyang proposed openstack/openstack-ansible-os_barbican master: Dep's should be restricted by tox-constraints  https://review.opendev.org/c/openstack/openstack-ansible-os_barbican/+/76401016:13
noonedeadpunkregarding 1905174 I'm not sure honestly - I never used ceph ansible and deployed rgw with swift only manually, so not huge expert16:13
openstackgerritwu.chunyang proposed openstack/openstack-ansible-os_blazar master: Dep's should be restricted by tox-constraints  https://review.opendev.org/c/openstack/openstack-ansible-os_blazar/+/76401116:14
jrosserwe have a deployment like that here which could be used as a reference16:14
noonedeadpunkwhile only I agree that we probably should double check that we have all set to make it as full swift replacement16:14
jrosserthough we did deploy completely seperate rgw for S3 and swift16:14
openstackgerritwu.chunyang proposed openstack/openstack-ansible-os_ceilometer master: Dep's should be restricted by tox-constraints  https://review.opendev.org/c/openstack/openstack-ansible-os_ceilometer/+/76401216:14
jrosseri think you need to have swift served from / if you want to pass tempest tests16:15
noonedeadpunkyeah would be awesome in case you could take a look16:15
jrosserand also S3 must be served from /16:15
jrosserso there is a bit of a conflict there16:15
openstackgerritwu.chunyang proposed openstack/openstack-ansible-os_cinder master: Dep's should be restricted by tox-constraints  https://review.opendev.org/c/openstack/openstack-ansible-os_cinder/+/76401316:15
noonedeadpunkand there're no options to configure tempest for root?16:16
jrosseri would have to get stuargr to look back at what we did16:17
noonedeadpunkI think for SWIFT it will jsut take endpoint from keystone so it doesn't matter what root it is set?16:18
jrosseryes thats right16:18
*** nurdie has quit IRC16:19
noonedeadpunkso we can set S3 in / and SWIFT in /swift/v1/AUTH_%(project_id)s (at leat that's wjhat I have16:19
noonedeadpunkbut I have `client.rgw.{{ hostvars[inventory_hostname]['ansible_hostname'] }}` and it's working perfect for me....16:20
noonedeadpunkok16:21
noonedeadpunk#topic office hours16:21
*** openstack changes topic to "office hours (Meeting topic: openstack_ansible_meeting)"16:21
gshippeyany high priority patches that I can give a go reviewing?16:21
noonedeadpunkhttps://review.opendev.org/c/openstack/openstack-ansible/+/763063 <- this one breaks bump bot16:22
noonedeadpunkhttps://review.opendev.org/c/openstack/openstack-ansible-os_adjutant/+/756313 <- huge adjutant patch that would be awesome to merge16:23
noonedeadpunkI had smth to discuss but didn't write down and clean forgot now :(16:25
noonedeadpunkI guess the main point was to do some reviews to be able to branch asap16:25
noonedeadpunkas I again failed to released it time :(16:25
noonedeadpunk*to release in time16:25
jrossersure - we should have a big push on reviewing things in order to release16:25
noonedeadpunkI've updated https://etherpad.opendev.org/p/osa-wallaby-ptg but probably worth creating new etherpad16:26
noonedeadpunkalso, I think we never merged spec for ssl16:26
noonedeadpunkand no feedbacck on it16:26
openstackgerritDmitriy Rabotyagov proposed openstack/openstack-ansible stable/ussuri: Add magnum tempest URL  https://review.opendev.org/c/openstack/openstack-ansible/+/76390816:27
noonedeadpunkhttps://review.opendev.org/c/openstack/openstack-ansible-specs/+/758805 <- this one16:27
jrosser:( sorry16:27
noonedeadpunkso everyone interested are welcome to comment out and probably merge it one day16:27
noonedeadpunkoh, well, have everyone heard about change of the ansible versioning process? :)16:31
noonedeadpunkthat they will make 3.0.0 instead of 2.11?16:31
noonedeadpunkand will release in .... february?16:32
noonedeadpunkso I think we should aim to release W with 3.0.016:33
noonedeadpunkthe main challenge will be py3.816:33
noonedeadpunkhttps://www.reddit.com/r/ansible/comments/jwzwwf/ansible300_schedule_and_preview_of_400_schedule/16:33
jrosseris that centos (!) that we will have a challenge with?16:34
noonedeadpunkand partially bionic that we can techically drop16:34
jrosseryeah, though i wonder if we can move to pyenv or something to get the actual python16:34
noonedeadpunkyeah, that was my sggestion as well16:35
noonedeadpunkit will increase deployment time for several minutes, but I think that should be generally ok16:35
noonedeadpunkbut it will be one more thing that we should take care about :(16:36
noonedeadpunkah, well, I placed also https://review.opendev.org/c/openstack/openstack-ansible-os_nova/+/763216 (which is not currentl;y working -had no time to look what specificly rong with it, but there should be something minor)16:37
noonedeadpunkand I have no idea how we technically can backport this...16:38
noonedeadpunkand I really had second thoughts about just masking systemd services to rollback to classic libvirt behaviour16:38
-openstackstatus- NOTICE: The Gerrit service on review.opendev.org is being restarted quickly to troubleshoot an SMTP queuing backlog, downtime should be less than 5 minutes16:41
noonedeadpunkas replacing libvirtd.service provided by system package is no fun....16:41
*** pcaruana has quit IRC16:45
*** shyamb has joined #openstack-ansible16:47
* jrosser looks now gerrit is back16:47
jrossernoonedeadpunk: do we need to replace the service file?16:48
noonedeadpunkwell, not replace, but I'm placing new one in /etc/systemd that should have prescedence over package provided one16:49
noonedeadpunkand I'm trying to make it as much the same as I can16:49
jrosserif there is just things like socket activation we could use a systemd dropin16:49
jrosserrather than replace16:49
noonedeadpunkyeah, sorry, it's really not replace but drop in16:50
*** shyam89 has joined #openstack-ansible16:50
noonedeadpunkif I got you right16:50
noonedeadpunk(I think we can't just partially override that way)16:50
*** shyamb has quit IRC16:51
*** spatel has joined #openstack-ansible16:53
jrossera systemd dropin can (i think) be used to selectively override the one that comes with the package16:54
jrosser"Along with a unit file foo.service, a "drop-in" directory foo.service.d/ may exist. All files with the suffix ".conf" from this directory will be parsed after the unit file itself is parsed."16:55
jrosserso if there is one thing that we need to add, like a [socket] section, then it can be carried in a seperate file and we just leave alone the package on, if it doesnt otherwise need changing16:56
* jrosser not looked at the detail of this though16:56
noonedeadpunkyeah, ofc I added sockets as a separate files. but I think that service itself needs changing to add requirement of these new sockets?16:57
*** lkoranda has quit IRC16:58
jrosseryeah so if you want to just change one thing out of the original config you can put foo.service.d/bar.conf [section] key=value16:59
*** shyam89 has quit IRC16:59
jrosserand that will change the setting from the original service unit16:59
noonedeadpunkhm....17:00
*** timburke has quit IRC17:00
jrosserit's the systemd native way of doing overrides of distro provided unit files17:00
jrosserso that things dont vanish when you update packages for example17:00
*** rpittau is now known as rpittau|afk17:01
noonedeadpunkI'm strugling to do it properly then....17:01
noonedeadpunkok, yes, I think it's better approach that mess I was trying to do17:02
noonedeadpunk*then17:02
noonedeadpunkbut how to place socket's itself... I thought that adding sockets deployment to systemd-service role should be perfect17:03
noonedeadpunkbut the problem is that sockets should be part of the service (and reference it)17:03
noonedeadpunkAnd I have no idea what the right structure should be in case we don't define service17:04
noonedeadpunkThat what I come up with for systemd role https://review.opendev.org/c/openstack/ansible-role-systemd_service/+/76319417:04
noonedeadpunkbut maybe I should do sockets just indepenently from service....17:04
jrosseri think they should be seperate .socket units?17:06
admin0are we still on meeting ?17:06
noonedeadpunkyeah, but they should have `Service`  in `[Socket]`17:07
noonedeadpunk#endmeeting17:07
*** openstack changes topic to "Launchpad: https://launchpad.net/openstack-ansible || Weekly Meetings: https://wiki.openstack.org/wiki/Meetings/openstack-ansible || Review Dashboard: http://bit.ly/osa-review-board-v3"17:07
openstackMeeting ended Tue Nov 24 17:07:20 2020 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)17:07
openstackMinutes:        http://eavesdrop.openstack.org/meetings/openstack_ansible_meeting/2020/openstack_ansible_meeting.2020-11-24-16.01.html17:07
openstackMinutes (text): http://eavesdrop.openstack.org/meetings/openstack_ansible_meeting/2020/openstack_ansible_meeting.2020-11-24-16.01.txt17:07
openstackLog:            http://eavesdrop.openstack.org/meetings/openstack_ansible_meeting/2020/openstack_ansible_meeting.2020-11-24-16.01.log.html17:07
admin0:) thanks17:07
jrosserprobably with Accept=no and Service=libvirtd, or something17:07
noonedeadpunkand let users to define Service correctly?17:08
admin0jamesdenton, so i want to use br-flat on enp5s0 - and failing .. what do i need to do in the configs to make it work17:08
noonedeadpunkthen maybe we should do separate role for sockets? as we did for mounts and networkd?17:08
jrosserAccept=no says that the socket should be passed by systemd to the service rather than used to spawn the service on connection17:08
jrosserand then Service= says which service to hand them off to if then name of the .socket file does not exactly match the service .conf file17:09
jrosserwhich if we have two sockets might need to be explicitly done17:09
jrossernoonedeadpunk: i just had a look at our radosgw, and we needed 'rgw swift url prefix = /' for swift in order to get refstack to pass better17:11
jrosserand you can't have the swift prefix as /, and be serving S3 at the same time as S3 is implicitly at /17:11
jrosserso we have two instances of rgw, one for swift and one for S3 all behind haproxy17:12
noonedeadpunkwell, `Accept=no,` is default behavior anyway17:13
noonedeadpunkyeah, I see, makes sense then. I just didn't do refstack yet :(17:13
jrossertbh i have no idea if that is broken swift in rgw or refstack/tempest error, or it really has to be that way.....17:14
noonedeadpunkwhat I was trying to do is to link socket to the systemd service, since eventually socket can't live without systemd service17:14
noonedeadpunkand systemd service can have several sockets17:14
noonedeadpunkbut in our structure that meant that we should define systemd service as well if we want socket17:14
noonedeadpunkwich makes sense for most of the usecases where we could define services, except libvirt...17:15
noonedeadpunkwhich is package provided and we should just slightly modify it17:16
noonedeadpunkbut yeah, probably we should really add there sockets independently from service structure17:17
noonedeadpunkand add creation of the foo.service.d for overrides17:17
noonedeadpunkok, thanks jrosser that is really great input17:18
jrosserok np hope i have understood it properly :)17:19
noonedeadpunkyeah, you did)17:19
jrosserhopefully we can use the existing systemd_service role to make .socket files17:19
jrosserbecause really they are just another unit file17:19
noonedeadpunkwell, like systemd-mount actually...17:20
jrosserthere are also implicit dependancies between a service and the .socket17:22
jrosserso it may not be necessary to specify the sockets at all in the service unit17:23
noonedeadpunktrue17:23
noonedeadpunkthat was the reason why I put socket as part of the service definition....17:23
noonedeadpunkwhich is not really handy in terms of the libvirt17:24
*** klamath_atx has quit IRC17:57
spateljamesdenton: around?18:04
*** jbadiapa has quit IRC18:05
*** yann-kaelig has joined #openstack-ansible18:08
admin0i am waiting for him as well18:09
jamesdentoni'm here but have to bounce for about an hour. whats up?18:09
admin0 i want to use br-flat on enp5s0 -- what config option do i need to do ?18:10
admin0and how does enp5s0 appear in the network nodes/hypervisors .. just an interface without any bridge or ip .18:10
admin0or how18:10
admin0another question .. is it possible to use br-vlan but without any segmentation ip ( no tagged ips at all)18:11
jamesdentonwhen you use host_bind_override, the agent will configure networks using enp5s0 directly rather than using an intermediate bridge (like br-vlan). So if you create a VLAN (50) provider network, you get brq937592fj-29 -> enp5s0.50 rather than brq937592fj-29 -> br-vlan.50 -> br-vlan -> enp5s018:11
jamesdentonthat make sense?18:11
admin0yes .. on vlans i have done this before .. this is flat ( the datacenter does not accept tagged traffic)18:12
jamesdentonif you want a flat network, you will need to use host_bind_overide, as the host cannot connect a bridge to a bridge (it would try to plug br-vlan into brq bridge18:12
jamesdentonwhat error did you get this time?18:12
admin0https://gist.github.com/a1git/4812121051d32f698cd7b309abc533aa -- this is what I have now ..18:13
admin0the interface does not have any IPs at all and is not a part of any bridge18:13
admin0its just a bridge18:13
admin0the bridge gets its ip .. but the bridge is just an island bridge .. its not plugged into any interface or other bridges18:14
admin0so the ip is there, but its not going out anywhere18:14
admin0is there a command to see how a brq interface is linked to a ens interface ?18:14
jamesdentonbrctl show brqXXX18:15
jamesdentonyou may also need to set host_bind_override: "enp5s0" on the other network block, too.18:15
jamesdentoni don't use flat networks so i can't recall offhand18:16
admin0 on the other network block -- which ones18:17
admin0same here .. its the first time in 8 years using osa that i have the need to use flat networking18:18
jamesdentonthe other br-vlan18:18
admin0maybe i will spin up a vyos to create a tagged vlan between this and vyos and use the vyos to route to the DC18:18
jamesdentonwhen my lab infra node comes back up i will test it18:19
admin0ok18:19
*** andrewbonney has quit IRC18:19
jamesdentonsome time this afternoon18:19
admin0that is fine18:19
admin0many thanks18:19
spateljamesdenton: do you have any specific load-test scenario to try out?18:23
spatelI am using 64 bytes UDP only18:23
spatelI am going to try couple TCP scenario also18:24
*** jamesdenton has quit IRC18:26
*** jamesden_ has joined #openstack-ansible18:26
spateladmin0: how can run datacenter in flat network?18:28
spatelmy home network i can say flat [ISP]-----[Wireless Router]--------[my_lan]18:30
admin0i got a new client who wants me to config a cluster in a datacenter provided hardware ..  all is up and working .. except the exsternal network18:30
spatelbut datacenter can't be like that right?18:30
admin0in my own home, i got br-vlan :)18:30
spatellol18:31
admin0mikrotik routers are wonderful18:31
spateli love their software.. very nice and easy to use18:31
admin0its what is used to run a whole isp in asia and africa --very low cost, but featureful devices18:31
spatelindeed, its very cheap and easy to turn you PC to high performance router/firewall/vpn :)18:32
spatelI used them a lot specially in wireless network18:32
admin0i have the isp modem in bridge mode so that the router gets direct IP .. and have a archer c7 in bridge mode also so that the router controls everything + dual 2 and 5 ghz18:33
spatelnext year i am planning to turn my basement in mid-size datacenter, i am collecting all equipments for that.18:33
spatelnice!18:34
spateladmin0: what kind of home lab you have?18:37
admin0its 2 servers only ..18:37
admin0but its an openstack kind ..18:37
admin0meaning i use the openstack qcow2 files as base images to spawn up vms18:37
spatelI have lots of server in my datacenter on floor which i am planning to bring it home to build good size lab18:38
admin0so between the 2 servers, they have around 20-25 virual machines running18:38
admin0it takes me less than a minute to spawn up18:38
spatelwhat is the configuration of those server ?18:38
admin0i use cloud-init and base files exactly like openstack does18:38
admin0256gb of ram, 20 thread cpus and  2x 1gb nvme on raid 018:38
spatelthat is monster machine18:39
admin0my laptop has 8 threads i7 and 64gb ram  :D18:39
spateldo you keep them powered on all the time (what about electricity bills?)18:39
admin0i use a aio18:39
admin0i don't keep them powered all the time .. only when i have to test something or teach my son18:39
spatelgood18:40
admin0i do have tst lab in office, but with  sata and ssd .. and when you need to etst stuff.. they are sloe18:40
admin0like running 2 virtual ceph clusters18:40
admin0i had to test osa + dual ceph18:40
admin0so for those, nvme on raid0 works awesome18:40
spatelhmm18:41
spatelI'm also thinking to build lab like that + network lab18:42
spatelI have couple of nexus switch + Cisco ASA firewall18:42
spatelMostly people use EVE-NG or Cisco CML for networking lab18:43
*** cshen has quit IRC18:43
spatelHow do you guys rollout compute nodes? using kickstart + scripts or ironic style image deployment?18:45
admin0pxe+kickstart18:45
admin0i don't like ironic or maas where someone ( interns or anyone ) accidently cliking a button and destroying clients workload18:46
admin0so someone needs to be on idrac to physically slect a os-reinstall18:46
spatel+118:46
admin0all servers are set to boot from pxe, and the first menu that auto loads in pxe is to boot from hard-disk18:47
admin0so it works nicely18:47
spatelI am using pxe+ks18:47
admin0same18:47
admin0i use vyos for all my lab network needs18:48
spatelI have set default image to compute so whenever new machine join datacenter it default turn into compute node18:48
admin0also vyos for customers who want to do ipsec between the cloud and their office and run their workload internally18:48
spatelwhy not mikoritk ?18:49
spatelmikrotik*18:49
*** cshen has joined #openstack-ansible18:50
admin0i have not used a virtual mikrotik yet18:51
*** ajg20 has joined #openstack-ansible18:51
admin0i even have a small mikrotik that is 2ghz that has the same AP name as my home .. so whenever i go on holiday .. i just plugin to eth0 .. and then i don't have to reconfig my phones, tables, laptops etc18:51
admin0works fine .. like in airbnb vacations and some hotels that have eth plugs18:52
admin0which automatically starts a vpn session to home .. so home ip everywhere18:52
spatelwow! that is cool18:53
admin0is mikrotik free for virtual image ? like we don't have to license it ?18:53
openstackgerritMerged openstack/openstack-ansible master: Fix networking_nsx_git_track_branch  https://review.opendev.org/c/openstack/openstack-ansible/+/76306318:53
spatelIts free software :)18:54
admin0upto which level ?18:54
admin0is level6 also free ?18:54
spatelhttps://mikrotik.com/download18:54
spatelI didn't know you have to pay because many years back i was using it free18:55
admin0https://wiki.mikrotik.com/wiki/Manual:License -- check the license levels and what you get per level18:56
spatelI was working for small ISP and they were buy standard desktop and turning into them in mikrotik router18:56
admin0i guess they had a keygen18:57
spateladmin0: haha.. i used that in past :)18:57
admin0for the first 8 years of my IT carrier, i worked on isps also :)18:58
spatelsame with me but only 4 years.18:58
spatelWe had big RAS server where all telephone lines terminated :)18:59
spatelIt was dialup ISP18:59
admin0my workspace in the 1st isp :)18:59
spatelFreeRadius was our billing + AAA18:59
admin0we used radiator18:59
spateloh boy we did used radiator18:59
admin0freeradius did not had good oracle support that time ( 1997 - 2001)18:59
spatelCGI + Perl based18:59
spatelinterface was crapy but it was good19:00
admin0i like radiator .. its like a framework to design your own radius19:00
admin0works awesome19:00
spateltotally19:00
spatelsquid proxy was our CDN.. haha19:01
admin0yep19:01
admin095% filled up with porn19:01
admin0that was what 2000s was all about .. cybercafes19:02
spatelhaha.. we put some bandwidth policy around that19:02
admin0https://www.softether.org/  is  the VPN i am sticking to now19:02
admin0works nicely19:02
admin0especially for iphones and my android phone19:03
spatelman!! cybercafe where my IT life started, I was paying $2 per hour.. haha  (and keep asking for owner don't throttle my bandwidth...)19:03
admin0here .. look at this site (which totally is doing business fine so far) - https://www.lingscars.com/ to take you back to your cybercafe times19:04
spatel56kbps like 10 people was using internet connection using wingate proxy :)19:05
spatelgood old days.19:05
admin0yep19:06
spatellol19:06
admin0upgrading customers from windows 3.11 -> 95 was where most money was made19:06
spateloh man!! i remembered now, i had 6 floppy disk to install 3.11 windows19:07
spatelboot from DOS 6.2 to type > win  to boot GUI :)19:07
admin0and played price of persia i guess :)19:12
admin0prince*19:12
* persia is relieved, not having remembered any bribes from admin0 in that time period19:12
* admin0 orders a pizza to persia 19:13
admin0persia, did you played price of persia :)19:13
admin0in DOS or mac19:13
admin0pre 2000 AD19:13
persiaSadly, that title postdates my time immersed in Broderbund software19:16
admin0:)19:19
admin0checking if anyone here got kubernetes working in osa19:19
openstackgerritRafael Folco proposed openstack/openstack-ansible-os_tempest master: Fix stackviz for failed tempest runs  https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/76405519:22
*** ajg20 has quit IRC19:53
*** nurdie has joined #openstack-ansible20:17
*** nurdie has quit IRC20:17
*** cshen has quit IRC20:37
*** cshen has joined #openstack-ansible20:38
*** rfolco|ruck has quit IRC21:32
*** spatel has quit IRC22:35
*** cshen has quit IRC22:35
djhankbadmin0: I am still stuck on magnum as well :-( I was hoping you had made some progress22:41
admin0i will .. its in my "must be done" todo list :)22:41
*** gshippey has quit IRC22:49
jamesden_admin0 i still wait for my ssd clone to finish22:50
*** jamesden_ is now known as jamesdenton22:50
jamesdentonadmin0 you might be able to do away with that block in openstack_user_config and instead override with this in user_variables: http://paste.openstack.org/show/800393/22:56
*** rfolco|ruck has joined #openstack-ansible22:57
*** klamath_atx has joined #openstack-ansible23:28
*** yann-kaelig has quit IRC23:28
*** cshen has joined #openstack-ansible23:32
*** luksky has quit IRC23:33
*** cshen has quit IRC23:37
*** rfolco|ruck has quit IRC23:50
*** rfolco|ruck has joined #openstack-ansible23:50
admin0jamesdenton, so remove both blocks for br-vlan and br-flat from the config23:51
admin0i will try this tomorrow23:51
* admin0 sends jamesdenton a pizza :) 23:51
*** rfolco|ruck has quit IRC23:55
*** tosky has quit IRC23:57
*** openstackgerrit has quit IRC23:59

Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!