| *** tosky has quit IRC | 00:10 | |
| *** rfolco has quit IRC | 00:19 | |
| *** rfolco has joined #openstack-ansible | 00:19 | |
| *** rfolco has quit IRC | 00:25 | |
| spotz | jrosser: You hear? | 01:16 |
|---|---|---|
| spotz | here even | 01:16 |
| *** macz_ has quit IRC | 01:37 | |
| *** macz_ has joined #openstack-ansible | 01:38 | |
| *** maharg101 has joined #openstack-ansible | 01:58 | |
| *** maharg101 has quit IRC | 02:02 | |
| *** cshen has quit IRC | 02:06 | |
| *** cp- has quit IRC | 02:50 | |
| *** cp- has joined #openstack-ansible | 02:51 | |
| *** kplant has joined #openstack-ansible | 03:12 | |
| *** gshippey has quit IRC | 03:29 | |
| *** lemko3 has joined #openstack-ansible | 03:35 | |
| *** lemko has quit IRC | 03:38 | |
| *** lemko3 is now known as lemko | 03:38 | |
| *** maharg101 has joined #openstack-ansible | 03:59 | |
| *** maharg101 has quit IRC | 04:03 | |
| *** raukadah is now known as chandankumar | 04:38 | |
| *** rohit02 has joined #openstack-ansible | 05:11 | |
| *** brad[] has quit IRC | 05:20 | |
| *** shyamb has joined #openstack-ansible | 05:21 | |
| *** shyam89 has joined #openstack-ansible | 05:21 | |
| rohit02 | noonedeadpunk: openstackansible centos 8 ussuri deployment failed with ""Error: No matching repo to modify: PowerTools." | 05:21 |
| rohit02 | in setup-hosts.yml playbook | 05:21 |
| rohit02 | @Team openstackansible centos 8 ussuri deployment failed with ""Error: No matching repo to modify: PowerTools." in setup-hosts.yml playbook | 05:23 |
| *** shyamb has quit IRC | 05:25 | |
| *** brad[] has joined #openstack-ansible | 05:26 | |
| *** evrardjp has quit IRC | 05:33 | |
| *** evrardjp has joined #openstack-ansible | 05:33 | |
| *** shyam89 has quit IRC | 05:59 | |
| *** shyamb has joined #openstack-ansible | 06:19 | |
| *** macz_ has quit IRC | 06:38 | |
| *** rohit02 has quit IRC | 06:38 | |
| *** shyamb has quit IRC | 07:05 | |
| *** shyamb has joined #openstack-ansible | 07:06 | |
| *** cshen has joined #openstack-ansible | 07:14 | |
| *** gyee has quit IRC | 07:21 | |
| *** spatel has joined #openstack-ansible | 07:33 | |
| *** spatel has quit IRC | 07:38 | |
| *** shyamb has quit IRC | 07:50 | |
| *** rohit02 has joined #openstack-ansible | 07:52 | |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible stable/ussuri: Bump SHAs for stable/ussuri https://review.opendev.org/c/openstack/openstack-ansible/+/766860 | 07:56 |
| *** rpittau|afk is now known as rpittau | 08:06 | |
| *** maharg101 has joined #openstack-ansible | 08:11 | |
| *** MickyMan77 has joined #openstack-ansible | 08:16 | |
| *** andrewbonney has joined #openstack-ansible | 08:17 | |
| *** MickyMan77 has quit IRC | 08:21 | |
| openstackgerrit | Dr. Jens Harbott proposed openstack/openstack-ansible master: Fix galera link in aio quickstart guide https://review.opendev.org/c/openstack/openstack-ansible/+/767092 | 08:22 |
| *** shyamb has joined #openstack-ansible | 08:34 | |
| *** macz_ has joined #openstack-ansible | 08:35 | |
| *** macz_ has quit IRC | 08:39 | |
| jrosser | morning | 08:39 |
| *** tosky has joined #openstack-ansible | 08:41 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_octavia master: [doc] Adjust octavia docs https://review.opendev.org/c/openstack/openstack-ansible-os_octavia/+/766833 | 08:47 |
| admin0 | morning | 08:49 |
| admin0 | what is the best-practice on setting up osa+ceph as to where to install the mons ? is it a good idea to give c1/c2/c3 as mons to ceph-ansible .. or create a vm inside c1/c2/c3 for mon ? | 08:49 |
| *** shyamb has quit IRC | 08:52 | |
| *** shyamb has joined #openstack-ansible | 08:53 | |
| kleini | admin0: I use now ceph-ansible currently only in test environments and there I just install MONs along with OSDs. for production best approach would be to have three dedicated machines for MONs and hard disc monster machines providing as much hard disc as possible and OSDs on them, of course | 09:07 |
| admin0 | osd's are there .. just wondering if people are usig controllers as mons also .. or doing 3 separate mons | 09:08 |
| kleini | furthermore we put optane SSDs into OSD machines for non-collocated journal to have writes returns as fast as possible | 09:09 |
| admin0 | and also checking if there is a potential conflict in future where osa will install package X while (if controllers on metal are mons) ceph-ansible will override them | 09:10 |
| kleini | depending how large your Ceph cluster is, MONs can have some load | 09:10 |
| jrosser | admin0: my ceph deployment is totally seperate from my openstack | 09:10 |
| jrosser | but you can do it however you like :) | 09:10 |
| jrosser | imho having some decoupling between the two is good, you have ceph clients and rbd libraries for your openstack on the controllers, on metal for cinder potentially | 09:11 |
| jrosser | and you don't necessarily want to introduce too much version / co-installation coupling with your ceph server side stuff | 09:11 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_octavia master: [doc] Adjust octavia docs https://review.opendev.org/c/openstack/openstack-ansible-os_octavia/+/766833 | 09:19 |
| admin0 | thanks jrosser kleini | 09:26 |
| kleini | will try now to extend my staging deployment with octavia. I saw https://review.opendev.org/c/openstack/openstack-ansible-os_octavia/+/766833. is there anything else, that should be considered? | 09:41 |
| noonedeadpunk | andrewbonney: uh, zun seems to be instantly failing tempest for centos now... not failing, but timeouting, which is even more annoying... | 09:44 |
| noonedeadpunk | maybe it needs more modern docker - dunno.... | 09:44 |
| andrewbonney | noonedeadpunk: it's definitely passed with the older version which is confusing me. I'll have to try spinning up an AIO again | 09:46 |
| noonedeadpunk | but it was passing before 8.3 release ? haha | 09:46 |
| jrosser | train looks broken https://zuul.opendev.org/t/openstack/build/252172c84e604d62a5313aa7317c2d56/log/job-output.txt#13126 | 09:47 |
| noonedeadpunk | ubuntu upgrade job? | 09:47 |
| noonedeadpunk | yeah | 09:47 |
| jrosser | yes | 09:47 |
| jrosser | i tried cloning the repo and checking out that SHA by hand and its just fine | 09:47 |
| jrosser | cant reproduce what ansible is doing there | 09:47 |
| jrosser | but we've seen this same thing may times before | 09:48 |
| jrosser | same on the train sha bump patch i think | 09:48 |
| noonedeadpunk | the most frustarting is that centos 7 is good | 09:49 |
| *** rohit02 has quit IRC | 09:50 | |
| noonedeadpunk | I'm wondering if we should backport fix of the upgrade jobs to train | 09:50 |
| jrosser | though it's a git checkout error isnt it? | 09:52 |
| noonedeadpunk | well, we shouldn't to git checkout there, so :) | 09:52 |
| noonedeadpunk | I thin kwe're supposed to get zuul checked out versions | 09:52 |
| jrosser | oh i see - right | 09:53 |
| noonedeadpunk | I mean this one https://review.opendev.org/c/openstack/openstack-ansible/+/746366 | 09:53 |
| jrosser | given that is a CI only change then backporting is really OK | 09:53 |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible stable/train: Make upgrade jobs get role relevant versions https://review.opendev.org/c/openstack/openstack-ansible/+/767048 | 09:54 |
| noonedeadpunk | at least let's see result... | 09:54 |
| jrosser | ok, so what blocks us on release candidate for V right now | 09:55 |
| admin0 | octavia by itself is not complete .. ( it only does http ) .. babrican is also needed for https | 09:57 |
| noonedeadpunk | jrosser: I was waiting only for zun patch | 09:58 |
| noonedeadpunk | Can push a branching patch right now otherwise | 09:58 |
| noonedeadpunk | was just under impression we can get it working as it seemed absolutely ready... | 09:59 |
| *** rohit02 has joined #openstack-ansible | 10:02 | |
| andrewbonney | noonedeadpunk: I'll try pushing a patch with a newer docker version for centos just in case. If it works that'll be quicker than I can get an AIO ready | 10:05 |
| noonedeadpunk | worth trying in the meanwhile | 10:06 |
| openstackgerrit | Andrew Bonney proposed openstack/openstack-ansible-os_zun master: Update zun role to match current requirements https://review.opendev.org/c/openstack/openstack-ansible-os_zun/+/763141 | 10:08 |
| openstackgerrit | Merged openstack/openstack-ansible master: Fix galera link in aio quickstart guide https://review.opendev.org/c/openstack/openstack-ansible/+/767092 | 10:18 |
| *** rpittau is now known as rpittau|bbl | 10:27 | |
| *** spatel has joined #openstack-ansible | 10:39 | |
| *** spatel has quit IRC | 10:44 | |
| *** shyamb has quit IRC | 10:46 | |
| frickler | trying a simple aio setup fails for me with http://paste.openstack.org/show/801043/ . possibly triggered by the first playbook run having been interrupted due to network issues, this now happens on every retry. any hint how to clean this up? this is stable/ussuri head, btw. | 11:00 |
| jrosser | frickler: is that failing during installing neutron? | 11:02 |
| jrosser | if so you can try openstack-ansible playbooks/os-neutron-install.yml -e venv_rebuild=yes | 11:03 |
| jrosser | that will force it to start over on the venv build completely for neutron | 11:03 |
| frickler | jrosser: yes, trying that now, thx | 11:04 |
| *** rohit02 has quit IRC | 11:19 | |
| *** rohit02 has joined #openstack-ansible | 11:19 | |
| admin0 | on a ceph + osa, where osa cannot ssh to ceph mons . how do i pass cinder and glance auth keys to osa ? | 11:25 |
| jrosser | admin0: documentation here https://docs.openstack.org/openstack-ansible-ceph_client/latest/config-from-file.html | 11:28 |
| kleini | admin0: and for ussuri the list of ceph_mons is still required otherwise ceph_client role is not executed | 11:31 |
| jrosser | didnt we patch that? | 11:32 |
| kleini | yes, it is patched but not released | 11:32 |
| admin0 | will provide the list .. just that it is not able to ssh to that | 11:32 |
| kleini | 21.2.0 does not contain the backport | 11:32 |
| *** rfolco has joined #openstack-ansible | 11:46 | |
| *** alvinstarr has joined #openstack-ansible | 11:47 | |
| *** SecOpsNinja has joined #openstack-ansible | 11:48 | |
| *** mgariepy has quit IRC | 11:59 | |
| *** shyamb has joined #openstack-ansible | 12:12 | |
| openstackgerrit | Merged openstack/openstack-ansible master: Cleanup group_vars https://review.opendev.org/c/openstack/openstack-ansible/+/766001 | 12:24 |
| *** rohit02 has quit IRC | 12:47 | |
| *** shyamb has quit IRC | 12:49 | |
| frickler | jrosser: o.k., your command worked fine and the complete run finished fine after that. would it make sense to add a note about this to the quickstart guide or is there some general "known pitfalls" guide that I did not yet find? | 12:55 |
| jrosser | well, i'm not sure, in some sense it is a bug | 12:56 |
| jrosser | the files that it couldnt find get deleted if something bad happens during the venv build, which sounds like your network trouble may have caused | 12:57 |
| jrosser | ideally re-running things should have worked out better than it did | 12:58 |
| jrosser | in a way a lot of the documentation is in the code, so heres that var i suggested https://github.com/openstack/ansible-role-python_venv_build/blob/master/defaults/main.yml#L100-L107 | 12:58 |
| *** cloudnull has quit IRC | 12:59 | |
| *** cloudnull has joined #openstack-ansible | 13:00 | |
| jrosser | there is a general troubleshooting guide here https://docs.openstack.org/openstack-ansible/latest/admin/troubleshooting.html | 13:00 |
| *** mgariepy has joined #openstack-ansible | 13:07 | |
| *** macz_ has joined #openstack-ansible | 13:07 | |
| *** macz_ has quit IRC | 13:07 | |
| *** macz_ has joined #openstack-ansible | 13:08 | |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible stable/train: Make upgrade jobs get role relevant versions https://review.opendev.org/c/openstack/openstack-ansible/+/767048 | 13:08 |
| noonedeadpunk | andrewbonney: nah, 20.10 docker for centos fails the same way as ubuntu focal does... | 13:10 |
| andrewbonney | Yeah, which is odd as it has worked before with what I assume was 20.10. I'll let you know what I find in my own AIO | 13:10 |
| noonedeadpunk | sure - sorry for pinging :) | 13:11 |
| andrewbonney | No problem | 13:11 |
| admin0 | does barbican have anything in the horizon ui ? | 13:11 |
| admin0 | or is it just purely api solution ? | 13:11 |
| *** rpittau|bbl is now known as rpittau | 13:12 | |
| *** SiavashSardari has joined #openstack-ansible | 13:22 | |
| frickler | jrosser: thx, quite a bit of interesting stuff in that doc. but also links to ocata docs and ask.o.o. and more disfunctional links like in my earlier patch | 13:33 |
| frickler | and I'll try to reproduce and create a bug report | 13:33 |
| openstackgerrit | Merged openstack/openstack-ansible-os_octavia master: [doc] Adjust octavia docs https://review.opendev.org/c/openstack/openstack-ansible-os_octavia/+/766833 | 13:41 |
| *** spatel has joined #openstack-ansible | 13:56 | |
| *** jbadiapa has joined #openstack-ansible | 13:58 | |
| *** spatel has quit IRC | 13:59 | |
| *** spatel has joined #openstack-ansible | 14:01 | |
| *** mensis has joined #openstack-ansible | 14:12 | |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_monasca master: dnm https://review.opendev.org/c/openstack/openstack-ansible-os_monasca/+/767159 | 14:16 |
| spatel | jrosser: I have successfully able to PXE boot Ubuntu 20.04 using autointall :) | 14:16 |
| admin0 | nice | 14:17 |
| spatel | blog is on the way... | 14:17 |
| jrosser | awesome, this is the yaml based one? | 14:21 |
| spatel | Yes | 14:22 |
| jrosser | seems a good improvement over the previous preseed method, especially for disk partitioning | 14:22 |
| spatel | jrosser: disk section is little complex | 14:22 |
| jrosser | you can do things like say "install OS to the two smallest disks" rather than | 14:23 |
| jrosser | just hardwiring /dev/sda which then maybe changes letters in the future | 14:23 |
| jrosser | spatel: previously it was *horrible* :) | 14:24 |
| spatel | jrosser: i found on centos it was very simple and just 4 like snippet currently its like this - http://paste.openstack.org/show/801051/ | 14:24 |
| spatel | Do you have your user-data file. i would like to see what magic you have in it :) | 14:24 |
| spatel | but i like yaml way because easy to read :) | 14:25 |
| spatel | Do you guys prefer netplan for latest version of ubuntu? | 14:27 |
| admin0 | i like netplan .. though i am struggling to create ovs mappings :) | 14:27 |
| kleini | netplan is the first package I deinstall in Ubuntu. I prefer systemd-networkd | 14:29 |
| jrosser | spatel: something like this, templated out per host on the tftp server http://paste.openstack.org/show/801052/ | 14:29 |
| kleini | And the openstack-ansible-systemd-networkd role is great to configure systemd-networkd | 14:29 |
| spatel | jrosser: thank you! this is neat! | 14:30 |
| jrosser | early-commands and late-commands in our case download and execute shell scripts which do things that the installer cant | 14:31 |
| spatel | kleini: how to run network playbook when you are on same network? | 14:31 |
| jrosser | some specific sshd config and also those scripts hit some API endpoints with http so we know at what point the install has got to | 14:32 |
| spatel | jrosser: its like PRE and POST script | 14:32 |
| kleini | I configure only br-mgmt port for target machine during install and use the Ansible to deploy the other ports | 14:33 |
| SiavashSardari | spatel I like netplan too, specially now they added support for OVS | 14:33 |
| spatel | jrosser: is this your local cache uri: http://mirror.test.mist.rd.bbc.co.uk/gb.archive.ubuntu.com/ubuntu | 14:34 |
| spatel | SiavashSardari: do you have example snippet of netplan? | 14:34 |
| jrosser | yes, its http://<cache-server>/<original-url>/.... | 14:35 |
| spatel | kleini: Got it. my kickstart file does all bridge/vlan and IP address setup so after kickstart my machine is ready to run openstack playbook | 14:36 |
| SiavashSardari | spatel I'm using maas for the netplan setup but maas does not support ovs for now, so I added ovs part myself. let me find my controller netplan config | 14:37 |
| spatel | My compute-XXX number is my last octet of compute IP. | 14:37 |
| spatel | jrosser: why did you hardcoded macaddress: 24:8a:07:58:07:f4 ? | 14:38 |
| SiavashSardari | spatel this is our controller netplan config for our lab installation http://paste.openstack.org/show/801056/ | 14:39 |
| jrosser | spatel: that is a 'match' so that it is totally clear to the installer which interface is used for installation | 14:40 |
| jrosser | no guessing needed | 14:40 |
| *** fanfi has quit IRC | 14:41 | |
| spatel | jrosser: make sense | 14:41 |
| spatel | SiavashSardari: Thank you! why do you have nameserver on both interface how does that work? | 14:42 |
| spatel | also why did you create tun: interface? | 14:43 |
| SiavashSardari | I'm not sure if you guys are maas users or not, but I'm in the middle of setuping a new region and maas 2.9 has added some greate new features. you just need to setup machines and connectivity and start the machine. maas will take care of almost everithing after that | 14:44 |
| SiavashSardari | @spat | 14:44 |
| SiavashSardari | spatel The dns part is maas misconfiguration, we played too much with our lab setup. but it does not break anything | 14:45 |
| SiavashSardari | I'm using tun interface for neutron vxlan mesh | 14:46 |
| spatel | SiavashSardari: thank you, i heard about maas but never tried | 14:49 |
| SiavashSardari | maas has it's own downsides too but if you just want BM management and using ubuntu, maas will make your life muuuuch easier | 14:53 |
| admin0 | i am looking for netplan + ovs used on compute | 14:55 |
| SiavashSardari | admin0 take a look at http://paste.openstack.org/show/801056/ I hope it can help | 14:56 |
| SiavashSardari | BTW make sure you have the latest version of netplan | 14:57 |
| admin0 | SiavashSardari, the old 18.04 had this in /etc/network/interfaces -- https://gist.github.com/a1git/e731033a580dd7a20c22850df3d9c27f .. i am trying to find an equivalent setting in netplan | 14:59 |
| *** mensis has quit IRC | 14:59 | |
| SiavashSardari | admin0 sorry but I can't open your link due to sanctions. could you please use paste.openstack.org? | 15:06 |
| admin0 | SiavashSardari, http://paste.openstack.org/show/801058/ | 15:07 |
| admin0 | this was 18.04 way of creating a ovs bridge and also plugging a port into it | 15:08 |
| SiavashSardari | anyone here had a scenario with ceph-ansible which different class of osds on osd-nodes are in different order? I mean in one node sda and sdb are SSDs and in other node they are HDDs and two other disk are in the SSD class. and I need to use crush_classes too | 15:09 |
| openstackgerrit | Andrew Bonney proposed openstack/openstack-ansible-os_zun master: Update zun role to match current requirements https://review.opendev.org/c/openstack/openstack-ansible-os_zun/+/763141 | 15:09 |
| mgariepy | can i get an extra vote on this one ? https://review.opendev.org/c/openstack/openstack-ansible-haproxy_server/+/766504 | 15:13 |
| SiavashSardari | admin0 it would be something like this http://paste.openstack.org/show/801059/ | 15:14 |
| *** gshippey has joined #openstack-ansible | 15:16 | |
| spatel | mgariepy: this is great! | 15:17 |
| mgariepy | spatel, needs haproxy 2.0 tho. | 15:18 |
| spatel | +1 why don't we force osa to use haproxy 2.0 (haproxy 2.0 improved performance overall multi-processing/multi-threading) | 15:19 |
| mgariepy | it's in focal | 15:20 |
| spatel | nice! i am soon going to deploy it | 15:21 |
| mgariepy | hehe. | 15:21 |
| SiavashSardari | now that you mentioned it I rememberd long ago I added something to install haproxy 2.0 in our own repo | 15:22 |
| spatel | Question related IPv6 what subnet mask i should ask to my ISP /48 or /64 ? | 15:22 |
| SiavashSardari | I'll be happy if I can help with haproxy. we are already use that in production | 15:23 |
| SiavashSardari | spatel if you need any help with haproxy. I'll be happy to help. | 15:26 |
| spatel | SiavashSardari: I am using BigIP F5 on production (haproxy we used for lab only) | 15:27 |
| admin0 | SiavashSardari, so saying openvswitch: {} means its a ovs bridge ? | 15:27 |
| spatel | but i am deploying octavia and i may trying to dig into more to boost haproxy 2.0 performance number | 15:27 |
| admin0 | spatel, i have seen that mostly you add /48 as a network in your side | 15:28 |
| admin0 | like in a vlan and the dc keeps the .1 of that /48 | 15:28 |
| spatel | so /48 from ISP and create /64 for VLAN/LAN start with .1 | 15:28 |
| SiavashSardari | admin0 pretty much yeah. if you are interested about more details, there is more info on netplan reference website | 15:29 |
| *** jbadiapa has quit IRC | 15:40 | |
| *** jbadiapa has joined #openstack-ansible | 15:41 | |
| noonedeadpunk | #startmeeting openstack_ansible_meeting | 16:01 |
| openstack | Meeting started Tue Dec 15 16:01:20 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 |
| openstack | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 16:01 |
| *** openstack changes topic to " (Meeting topic: openstack_ansible_meeting)" | 16:01 | |
| openstack | The meeting name has been set to 'openstack_ansible_meeting' | 16:01 |
| noonedeadpunk | #topic office hours | 16:01 |
| *** openstack changes topic to "office hours (Meeting topic: openstack_ansible_meeting)" | 16:01 | |
| noonedeadpunk | we don't have new bugs, so we can just talk about current status of things | 16:01 |
| noonedeadpunk | I still haven't fixed ubuntu focal bug with live migrations and new way of libvirt should be configured to listen on port... | 16:02 |
| *** mgariepy has quit IRC | 16:02 | |
| noonedeadpunk | never have time to finish what I've started despite it's really critical | 16:02 |
| noonedeadpunk | hope will have some cycles to finish it during this week | 16:03 |
| noonedeadpunk | also I think we're super ready to branch, except zun patch, which went out of hands because of new docker release.. | 16:03 |
| noonedeadpunk | https://review.opendev.org/c/openstack/openstack-ansible/+/767048 seems helped upgrade jobs for ubuntu - thay failed anyway, but later | 16:05 |
| noonedeadpunk | https://zuul.opendev.org/t/openstack/build/db7fe82903284461b938e19a243847f5/log/job-output.txt#16772 | 16:06 |
| noonedeadpunk | and it's during second run, so I think we're good here | 16:07 |
| jrosser | o/ hello | 16:09 |
| noonedeadpunk | \o/ | 16:10 |
| noonedeadpunk | have anything to discuss?:) | 16:10 |
| jrosser | lets make a release! | 16:10 |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible stable/train: Bump SHAs for stable/train https://review.opendev.org/c/openstack/openstack-ansible/+/766859 | 16:11 |
| noonedeadpunk | Should we wait for https://review.opendev.org/c/openstack/openstack-ansible-os_zun/+/763141 ? | 16:11 |
| noonedeadpunk | I'm super rady honestly... We were merging and rechecking all kind of stuff the whole weekends for that... | 16:12 |
| noonedeadpunk | s/rady/ready/ | 16:12 |
| jrosser | at some point we need RC for people to test | 16:12 |
| noonedeadpunk | we can always backport though | 16:12 |
| jrosser | backports can be done and especially as zun is basically broken anyway for a long time, perhaps better to move forward | 16:13 |
| andrewbonney | Certainly if this build doesn't fix it I wouldn't wait any longer | 16:13 |
| noonedeadpunk | but we won't be able to merge patches that will be issued right after branching (like change to .gitreview) | 16:13 |
| jrosser | oh becasue the role is broken anyway | 16:13 |
| noonedeadpunk | yeah | 16:14 |
| jrosser | what is focal/live migration bug? | 16:14 |
| noonedeadpunk | the one I worked for sockets | 16:14 |
| noonedeadpunk | let me find it... | 16:14 |
| noonedeadpunk | https://bugs.launchpad.net/openstack-ansible/+bug/1903846 | 16:15 |
| openstack | Launchpad bug 1903846 in openstack-ansible " libvirt fails to start because --listen parameter is not permitted with systemd activation socket " [Critical,In progress] - Assigned to Dmitriy Rabotyagov (noonedeadpunk) | 16:15 |
| jrosser | ah yes and we had some patches to make .socket files | 16:15 |
| noonedeadpunk | https://review.opendev.org/c/openstack/ansible-role-systemd_service/+/765815 and https://review.opendev.org/c/openstack/ansible-role-systemd_service/+/763194 should cover usecases on the systemd_service part | 16:15 |
| noonedeadpunk | just need to make use os_nova this functionality... | 16:15 |
| jrosser | and thats some more vars in here? https://github.com/openstack/openstack-ansible-os_nova/blob/master/defaults/main.yml#L425-L435 | 16:21 |
| jrosser | oh hmm well its libvirtd.socket | 16:21 |
| jrosser | i wonder if those patches let us create just the .socket and leave distro systemd unit alone | 16:30 |
| noonedeadpunk | well, since I've followed your tip they should | 16:32 |
| noonedeadpunk | Idea was to set `systemd_overrides_only: true` for libvirtd service | 16:33 |
| noonedeadpunk | and we need to make small adjustments, so that it want sockets | 16:33 |
| jrosser | right, so that from the first patch then the socket made with the second | 16:33 |
| noonedeadpunk | yeah | 16:34 |
| noonedeadpunk | that was the plan kind of | 16:34 |
| *** jamesdenton has quit IRC | 16:36 | |
| *** jamesden_ has joined #openstack-ansible | 16:36 | |
| * jrosser another meeting | 16:37 | |
| noonedeadpunk | #endmeeting | 16:38 |
| *** 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" | 16:38 | |
| openstack | Meeting ended Tue Dec 15 16:38:00 2020 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 16:38 |
| openstack | Minutes: http://eavesdrop.openstack.org/meetings/openstack_ansible_meeting/2020/openstack_ansible_meeting.2020-12-15-16.01.html | 16:38 |
| openstack | Minutes (text): http://eavesdrop.openstack.org/meetings/openstack_ansible_meeting/2020/openstack_ansible_meeting.2020-12-15-16.01.txt | 16:38 |
| openstack | Log: http://eavesdrop.openstack.org/meetings/openstack_ansible_meeting/2020/openstack_ansible_meeting.2020-12-15-16.01.log.html | 16:38 |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_glance stable/ussuri: Trigger uwsgi restart https://review.opendev.org/c/openstack/openstack-ansible-os_glance/+/767144 | 16:44 |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_glance stable/train: Trigger uwsgi restart https://review.opendev.org/c/openstack/openstack-ansible-os_glance/+/767145 | 16:45 |
| andrewbonney | noonedeadpunk: Zun looking promising this time. Just got past all the tempest tests | 17:02 |
| noonedeadpunk | nice! | 17:02 |
| noonedeadpunk | awesome | 17:03 |
| noonedeadpunk | one of the best news for the last several weeks xD | 17:03 |
| andrewbonney | :) | 17:03 |
| noonedeadpunk | ubuntu seems good as well | 17:04 |
| *** mgariepy has joined #openstack-ansible | 17:15 | |
| *** SiavashSardari has quit IRC | 17:25 | |
| * jrosser votes | 17:37 | |
| jrosser | theres probably a whole bunch of patches to the zun repo we can abandon now | 17:37 |
| jrosser | noonedeadpunk: could you take a look at this too? https://review.opendev.org/c/openstack/openstack-ansible/+/766457 | 17:38 |
| noonedeadpunk | has haproxy patch already merged? | 17:39 |
| noonedeadpunk | (just can't find it) | 17:40 |
| jrosser | did this need a haproxy patch? | 17:40 |
| jrosser | there was a different one for prometheus exporter support to haproxy iirc | 17:41 |
| noonedeadpunk | hm. ok | 17:41 |
| jrosser | i think this is all the needed ones for security.txt https://review.opendev.org/q/topic:%22security-txt%22 | 17:41 |
| noonedeadpunk | btw it's worthy a reno imo | 17:42 |
| jrosser | i can do one later/tomorrow, need to head out for a bit now | 17:43 |
| *** rpittau is now known as rpittau|afk | 17:46 | |
| noonedeadpunk | yeah, we can do it as a follow up for sure | 17:51 |
| * noonedeadpunk will head out as well | 17:51 | |
| *** maharg101 has quit IRC | 17:58 | |
| openstackgerrit | Merged openstack/openstack-ansible-haproxy_server master: Add haproxy_frontend_only and haproxy_raw feature. https://review.opendev.org/c/openstack/openstack-ansible-haproxy_server/+/766504 | 18:07 |
| spatel | noonedeadpunk: sorry trying to catching up but any words on Victoria release? | 18:08 |
| spatel | soon i am going to deploy new cloud on ubuntu 20.04 so thought check with you. | 18:09 |
| SecOpsNinja | spatel, i have reinstalled all the rabbitmq cluster but im still having problems on vm creation. could it be a database problem? trying to check other components logs to see if can find the cause of rthis... | 18:18 |
| spatel | i don't think its database | 18:27 |
| SecOpsNinja | so the problem should be in the comunication of nova-scheduler and nova-compute nodes | 18:27 |
| spatel | did you re-deploy nova-* | 18:27 |
| SecOpsNinja | where are the logs in compute nodes? the journal of nova-compute doesn show anything usefull | 18:27 |
| spatel | journalctl -u nova-compute -f | 18:28 |
| SecOpsNinja | i reconfigured nova to use the new configurations of rabbitmq but i didn't recreated them | 18:28 |
| spatel | you need to run in debug to see some goodness | 18:28 |
| SecOpsNinja | i suposse you are talking about nova sapi and noot the nova in compute nodes | 18:28 |
| SecOpsNinja | yep im enabling in nova api and will try to enable debug in compute nodes | 18:29 |
| mgariepy | spatel, with the haproxy_raw you can add any feature you might want or needs in the service configs. | 18:29 |
| SecOpsNinja | let see try again | 18:29 |
| spatel | what happening when you creating vm ? (forget about logs, are you able to spin up vm?) | 18:29 |
| spatel | mgariepy: for example ? | 18:29 |
| SecOpsNinja | it get stukk in build forever | 18:30 |
| mgariepy | spatel, https://www.haproxy.com/blog/four-examples-of-haproxy-rate-limiting/ | 18:31 |
| spatel | when nova-scheduler not able to find something then it get stuck or rabbitMQ not happy | 18:31 |
| mgariepy | whatever option you need in the frontend. | 18:31 |
| spatel | very interesting.. | 18:32 |
| spatel | rate-limiting is really good feature | 18:34 |
| mgariepy | any feature can be added there. | 18:34 |
| spatel | any good tool to benchmark haproxy? | 18:34 |
| mgariepy | i don't know any | 18:35 |
| spatel | no problem will find out from google. I want to run some good load-test to certify deployment | 18:36 |
| jrosser | spatel: imho the backend API will be limiting *way* before you run out of steam on haproxy | 18:46 |
| spatel | jrosser: this load-testing is for my octavia haproxy (amphora) | 18:47 |
| jrosser | we got excellent use out of this https://locust.io/ | 18:47 |
| jrosser | it's distributed so you can stand up as many locust workers as you need | 18:48 |
| spatel | thanks, i will add that in my list :) | 18:48 |
| SecOpsNinja | humm im detecting some errors in open-iscsi.service in compute node that is trying to connected to older isscsi lvm volumes that where already deletect... let me see if i can find where is this info to delete it | 18:49 |
| jrosser | we ended up hitting radosgw with over 70gbit/sec with that | 18:49 |
| spatel | nice! | 18:49 |
| jrosser | it was a L4 haproxy in tcp mode in front of 6 radosgw physical servers | 18:50 |
| johnsom | Tsung is pretty good as well: http://tsung.erlang-projects.org | 18:50 |
| spatel | I heard from couple of people about Tsung | 18:51 |
| admin0 | on a ceph backed storage , can a instance downgrade also to a lower disk flavor ? | 18:51 |
| spatel | Thanks johnsom | 18:51 |
| *** kleini has quit IRC | 18:52 | |
| admin0 | say the flavor is 4 cpu, 4gb ram, 40gb disk .. his os takes 10gb .. can he downgrade to 2 cpu, 2gb ram, 20gb disk flavor safely / | 18:52 |
| jrosser | a downside with locust is you have to be prepared to write some python to pretend to be your clients | 18:52 |
| spatel | admin0: i think if you don't have room then you should be able to do that (not seeing any issue, but method is not good practically) | 18:54 |
| spatel | I meant if you have room then * | 18:54 |
| *** gyee has joined #openstack-ansible | 19:00 | |
| *** kleini has joined #openstack-ansible | 19:10 | |
| jawad_axd | Hi ! Error 43, windows 10 with gpu passthrough. Added img_hide_hypervisor_id=True to image property. Is there anything I need to do to load nvidia drivers in windows instance? | 19:15 |
| SecOpsNinja | from what im seeing in nova api logs, the scheduler selects an compute-host and aftert nova-conductor gets the selected compute hosts, tries to create the block_device_mapping amnd some locks regaring acquired by "nova.context.set_target_cell.<locals>.get_or_set_cached_cell_and_set_connections" and after tetting the locks it startes returning ERROR oslo.messaging._drivers.impl_rabbit [Er | 19:15 |
| SecOpsNinja | rno 113] EHOSTUNREACH. I allso locked to nova-compute service and i dont see any revelenat logs regarding this nyw vm | 19:15 |
| jawad_axd | spatel: yesterday was switch blocking traffic for some reason. compute was not reaching nova-api. Its fixed! | 19:16 |
| admin0 | jawad_axd, you need to identify the "some reason" to not let it happen again | 19:17 |
| *** andrewbonney has quit IRC | 19:17 | |
| jawad_axd | Actually my colleague fixed that issue, it was wrong MTU set on vlan. | 19:18 |
| SecOpsNinja | is there and easy what toi destroy all the openstack cluster with ansible or i need to reinstall the the machines and run openstack-ansible again? | 19:21 |
| spatel | jawad_axd: glad to here, it must be because of vfio-pci | 19:21 |
| jawad_axd | spatel: You mentioned, you have done passthrough successfully, Did u try windows and try to load nvidia drivers? ery famous error 43? | 19:23 |
| jawad_axd | *Very famous | 19:23 |
| spatel | I am running standard SRIOV network instance using PIC passthrough. not fancy stuff | 19:23 |
| jawad_axd | Ok | 19:26 |
| mgariepy | jawad_axd, the famous error can probably be fixed via a property. | 19:26 |
| jawad_axd | img_hide_hypervisor_id=True didt fix it. What else? | 19:27 |
| mgariepy | hmm it was that haha | 19:28 |
| mgariepy | jamesden_, what gpu do you have issue with ? | 19:29 |
| mgariepy | oops.. jawad_axd ^^ | 19:31 |
| jawad_axd | GeForce GTX 780 | 19:32 |
| mgariepy | which os ? | 19:42 |
| mgariepy | and which release of openstack | 19:42 |
| jawad_axd | I have setup stein with centos7. And I am doing passthrough with windows 10. | 19:43 |
| *** maharg101 has joined #openstack-ansible | 19:54 | |
| mgariepy | jawad_axd, have you tried with linux ? | 19:56 |
| openstackgerrit | Merged openstack/openstack-ansible-os_horizon master: Add ability to configure ALLOWED_HOSTS for horizon. https://review.opendev.org/c/openstack/openstack-ansible-os_horizon/+/765998 | 19:56 |
| mgariepy | just to make sure it's not a windows 10 quirk ? | 19:56 |
| jawad_axd | mgariepy: No, because I need to use it with windows for one use case. | 19:57 |
| *** maharg101 has quit IRC | 19:59 | |
| openstackgerrit | Merged openstack/openstack-ansible master: Add security.txt to haproxy frontend https://review.opendev.org/c/openstack/openstack-ansible/+/766457 | 20:01 |
| mgariepy | yeah i can understand taht but it can git you an idea if it's a windows setting or something else to look for. | 20:01 |
| jawad_axd | I ll give it a shot. | 20:03 |
| jawad_axd | Thnx | 20:04 |
| mgariepy | the kvm hidden was in the libvirt xml file ? | 20:04 |
| jawad_axd | img_hide property added kvm hidden true in xml file. Its there. | 20:06 |
| *** cshen has quit IRC | 20:09 | |
| SecOpsNinja | what could be the reason to take to long time to disable os a compute service using openstack compute service set --disable ? it returns an 504 but if i liust the service its disabled... | 20:09 |
| mgariepy | jawad_axd, you probably saw that already but maybe you need extra options in the libvirt file : https://passthroughpo.st/apply-error-43-workaround/ | 20:10 |
| *** cshen has joined #openstack-ansible | 20:15 | |
| *** jawad_axd has quit IRC | 20:59 | |
| *** jbadiapa has quit IRC | 21:06 | |
| ThiagoCMC | Hey guys! I'm curious about something that might be easy to fix... When I launch Windows Instances, it doesn't come with a virtual audio device. Is there an option to enable it? | 21:06 |
| *** gshippey has quit IRC | 21:10 | |
| *** rfolco has quit IRC | 21:10 | |
| *** SecOpsNinja has left #openstack-ansible | 21:52 | |
| *** trident has quit IRC | 21:56 | |
| *** trident has joined #openstack-ansible | 22:00 | |
| *** spatel has quit IRC | 22:11 | |
| *** maharg101 has joined #openstack-ansible | 22:20 | |
| *** cshen has quit IRC | 22:33 | |
| *** maharg101 has quit IRC | 22:54 | |
| *** macz_ has quit IRC | 23:08 | |
| *** macz_ has joined #openstack-ansible | 23:09 | |
| *** MickyMan77 has joined #openstack-ansible | 23:17 | |
| *** yann-kaelig has joined #openstack-ansible | 23:23 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!