| goldyfruit | Seems very quiet here! | 00:13 |
|---|---|---|
| *** stackedsax has quit IRC | 00:25 | |
| *** stackedsax has joined #openstack-kolla | 00:26 | |
| *** Tony31 has joined #openstack-kolla | 01:53 | |
| kevinz | mgoddard: +1 | 02:06 |
| *** zhanglong has joined #openstack-kolla | 02:15 | |
| openstackgerrit | Merged openstack/kolla-ansible master: Docs: Document EXTRA_OPTS https://review.opendev.org/707358 | 02:24 |
| *** zhanglong has quit IRC | 02:27 | |
| *** zhanglong has joined #openstack-kolla | 02:31 | |
| *** zhanglong has quit IRC | 02:45 | |
| *** zhanglong has joined #openstack-kolla | 02:47 | |
| *** goldyfruit has quit IRC | 03:05 | |
| *** goldyfruit has joined #openstack-kolla | 03:05 | |
| *** zhanglong has quit IRC | 03:17 | |
| *** k_mouza has joined #openstack-kolla | 03:39 | |
| *** dannins has joined #openstack-kolla | 03:39 | |
| *** k_mouza has quit IRC | 03:44 | |
| *** igordc has joined #openstack-kolla | 04:37 | |
| *** igordc has quit IRC | 04:48 | |
| *** igordc has joined #openstack-kolla | 04:48 | |
| *** igordc has quit IRC | 04:48 | |
| *** dking_desktop has quit IRC | 04:56 | |
| *** skramaja has joined #openstack-kolla | 04:59 | |
| *** shyamb has joined #openstack-kolla | 05:25 | |
| *** evrardjp has quit IRC | 05:34 | |
| *** evrardjp has joined #openstack-kolla | 05:34 | |
| *** Tony31 has quit IRC | 05:49 | |
| *** zhanglong has joined #openstack-kolla | 06:12 | |
| *** Tony31 has joined #openstack-kolla | 06:15 | |
| sorin-mihai | is there some way to automatically generate the certificates using certbot or other method to get certificates signed by letsencrypt? | 06:28 |
| Tony31 | Hi sorin-mihai - you could use a proxy like `Traefik` and then forward on to openstack | 06:29 |
| Tony31 | you could use http or https between traefik and openstack, but I have not yet divised a plan on openstack cli access. it's purely for web | 06:30 |
| sorin-mihai | what about generating the whole set of keys/certs with certbot and then moving files to the right location, would that work? | 06:31 |
| sorin-mihai | the advantage being that certbot would be able to keep track and update the certs and do just that, while anything else would get more complicated | 06:33 |
| Tony31 | I don't know certbot, does it provide you with the certificates pem ? | 06:34 |
| *** pbing19 has joined #openstack-kolla | 06:44 | |
| sorin-mihai | i think this would help as a starting point https://certbot.eff.org/lets-encrypt/pip-other | 06:44 |
| sorin-mihai | trying to figure out where it saves the files by default | 06:45 |
| sorin-mihai | i'm thinking about "certbot-auto certonly --standalone" since it spins up a temp httpd to do all the work, so no need for other things on the host that will run it | 06:47 |
| Tony31 | I'm trying to attach the `flat` network to my openstack, so that I can create tennant networks, project router and floating ip's, but I must be doing something wrong as it's not working. I've created a external network as flat and referenced `physnet2`. I've created the subnet which matches the real network on the switch for `physnet2`. Then I | 06:57 |
| Tony31 | attach a router and it gets an external interface IP within the physnet network, but from the switch I am unable to ARP or ping the router. | 06:57 |
| Tony31 | How can I Get sudo access to the neutron container so I can run `ovs show` commands? | 06:58 |
| openstackgerrit | Yongjun Bai proposed openstack/kolla-ansible master: Add support for encrypting nova api https://review.opendev.org/707131 | 07:02 |
| lklimin | Hi, Tony31. Do you consider replace haproxy by traefik (or caddy) in order to have letsencrypt from kolla deploy? | 07:11 |
| cosmicsound | Good day | 07:11 |
| cosmicsound | Does anyone else using RAID of any kind on the OS disk ? Working on a deployment with SSD on OS level and just thinking if I should keep it on RAID1 or just break the raid and make use of both disks. | 07:11 |
| Tony31 | no I have not | 07:11 |
| *** hu_berlin_kalle has joined #openstack-kolla | 07:12 | |
| osmanlicilegi | morning | 07:14 |
| cosmicsound | greets osmanlicilegi | 07:16 |
| sorin-mihai | cosmicsound, i'm using intel vroc raid1 (basically fake raid, but linux kernel sees it as managed through mdadm) with 2 nvme drives for the OS. the reason is to have a safety net in case one of the drives dies before i'm aware. in the end, it depends how important is the setup or what kind of data you have on the drives. i don't think it really matters | 07:18 |
| sorin-mihai | and i'm also trying to use a raid50, done through a lsi controller, as main disk for ceph, but i didn't get that far with the deployment. lol | 07:19 |
| cosmicsound | I am just thinking about the CEPH journals are better stored on SSD, and kolla has a way of using entire disk / partition, guess will need to make partition for Journals | 07:28 |
| *** shyamb has quit IRC | 07:29 | |
| cosmicsound | sorin-mihai , raid50? :D | 07:29 |
| *** gbatir_ has joined #openstack-kolla | 07:31 | |
| *** sluna has joined #openstack-kolla | 07:34 | |
| sorin-mihai | cosmicsound, raid50 might not be the best solution under ceph, obviously because of the potential hardware overhead, but the BBU of the dedicated controllers is a good way to make sure the data is written to the drives in case of power outage. other than that, i never had any real issues with drives, maybe i was just lucky. now if i would only figure out how to use luks encrypted containers as ceph devices... | 07:38 |
| yoctozepto | mgoddard, mnasiadka, hrw: stable branches also hit by py issues | 07:38 |
| hrw | yoctozepto: virtualenv stuff? | 07:42 |
| yoctozepto | hrw: yeah and family | 07:42 |
| yoctozepto | missing packages, generally tripleo blocks kolla now | 07:43 |
| hrw | tripleo team works on getting their stuff working again | 07:43 |
| yoctozepto | yeah, they do :-) | 07:44 |
| cosmicsound | is https://launchpad.net/kolla-cli maintained? | 07:50 |
| *** cah_link has joined #openstack-kolla | 07:50 | |
| cosmicsound | kolla_cli.commands.exceptions.CommandError: ERROR: Required file (/etc/kolla/kolla-cli/ansible/inventory.json) does not exist. | 07:51 |
| cosmicsound | Please re-install the kollacli to recreate the file. | 07:51 |
| cosmicsound | Ofcourse I reinstalled | 07:51 |
| yoctozepto | cosmicsound: so-so, it had some patches during train, but not much effort | 08:05 |
| Tony31 | Kayobe / Kolla guys - could you help me with networking? I have updated `external_net_names:` and re configured the host and redployed but the ml2_conf.ini is still showing `bridge_mappings = physnet1:eth0-ovs,physnet2:eth5-ovs` | 08:06 |
| Tony31 | before the update I had 2 x networks in `external_net_names:` and now, I removed one. But config applied on the network node (controller) is showing two | 08:07 |
| yoctozepto | Tony31: odd, could be kayobe not updated material for kolla-ansible hmm | 08:08 |
| yoctozepto | I'd grep that config on deployer for the extra interface name | 08:08 |
| yoctozepto | to see if it is somewhere still there | 08:09 |
| Tony31 | I dont think it's a blocker, because I only removed 1 of the networks from the list. But I am unable to get instances network access through eth5 | 08:09 |
| hrw | yoctozepto: can kolla patch X depend on kolla-ansible patch Y which depends on patch X? | 08:09 |
| yoctozepto | hrw: no, cyclic are not allowed in zuul | 08:09 |
| *** bengates has joined #openstack-kolla | 08:09 | |
| hrw | yoctozepto: hm. how to test infra k-a then ;d | 08:09 |
| Tony31 | on openstack I create a external network and create a subnet that matches whats really on eth5 of the controller. Then I attach a router to it, which gets an IP from that network. But I cannot ARP or ping the router. | 08:10 |
| Tony31 | how can I run ovs commands to debug it ? | 08:10 |
| Tony31 | i need sudo access on the container? | 08:10 |
| yoctozepto | hrw: the ugly way is to disable some CI jobs now and reenable later | 08:11 |
| *** tonythomas has joined #openstack-kolla | 08:11 | |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: WIP: Create 'infra' type of images https://review.opendev.org/698753 | 08:11 |
| hrw | yoctozepto: that's how it is now | 08:11 |
| hrw | V213 adds --infra switch | 08:11 |
| yoctozepto | what is V213 | 08:12 |
| hrw | yoctozepto: just random version number | 08:12 |
| mnasiadka | morning | 08:15 |
| mnasiadka | yoctozepto: stable branch also? oh lord | 08:15 |
| yoctozepto | hrw: great | 08:15 |
| yoctozepto | mnasiadka: yeah, indeed | 08:16 |
| hrw | yoctozepto: --infra is enabled by default | 08:16 |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: WIP: Create 'infra' type of images https://review.opendev.org/698753 | 08:17 |
| hrw | and one forgotten tweak | 08:17 |
| yoctozepto | hrw: oh, you can actually just tweak k-a to conditionally use infra | 08:18 |
| yoctozepto | this way we also don't get breakage until they publish | 08:18 |
| yoctozepto | so it's smart to do it conditionally | 08:18 |
| yoctozepto | hrw: biggest issue could be tripleo adapting | 08:20 |
| yoctozepto | no idea how hardcoded they are to binary now | 08:20 |
| yoctozepto | but could just switch the switch | 08:20 |
| yoctozepto | in theory | 08:20 |
| hrw | fk. you just made me realize that adding --infra switch just gave me huge amount of extra work for k-a | 08:21 |
| hrw | have to rethink it | 08:21 |
| hrw | because if --infra switch is present then k-a needs to be able to handle own infra switch. and all infra roles need to have a way to handle 'is it infra or binary/source image this time' | 08:23 |
| mnasiadka | coming to that, why don't we just label images that are infra as Docker LABEL, and when pushing tag them as both source and binary? | 08:23 |
| *** ivve has joined #openstack-kolla | 08:24 | |
| hrw | the idea of infra images was to make a bit less images so our push from CI would be smaller. But does it matter for 61 images? | 08:27 |
| *** ivve has quit IRC | 08:27 | |
| hrw | I removed ceph images from list | 08:27 |
| mnasiadka | hrw: well, you can tag images with two names when pushing, it won't push them twice I think | 08:28 |
| mnasiadka | or it should only push metadata | 08:28 |
| mnasiadka | I think also we wanted to be able to build infra images once a week or something, and the rest as usual | 08:29 |
| hrw | mnasiadka: merge binary+source build to one CI job and you get that nearly that | 08:29 |
| openstackgerrit | Radosław Piliszek proposed openstack/kolla-ansible master: Fix Prometheus mysqld exporter pointing to VIP address https://review.opendev.org/707582 | 08:29 |
| hrw | mnasiadka: if we take BINARY_SOURCE_IMAGES + injecting ENV vars from infra patch then we get same 'infra' images for binary/source without changing tags | 08:30 |
| mnasiadka | hrw: right, so then maybe a separate CI job for only infra images build+publish? | 08:30 |
| hrw | mnasiadka: no. building infra separatelly gives you only extra problems | 08:30 |
| hrw | 1st job builds infra, 2nd job waits for 1st one, pulls all images, do build with --skip-existing? huge time waste | 08:31 |
| hrw | our binary/source images differ at level of 'base' image because KOLLA_INSTALL_TYPE is present | 08:33 |
| hrw | once we get rid of it (and inject where needed) we have 61 images same for binary/source. then we merge CI jobs to build binary/source at same machine == 61 duplicate images with different tags == 61 less to push. all stays as it was | 08:35 |
| yoctozepto | the reason for separation is that we don't block on binary | 08:36 |
| yoctozepto | binary is hardly controlled by us | 08:36 |
| hrw | are docker images reproductible? always generate same result? | 08:41 |
| hrw | if yes then I can extract some stuff from infra patch and we can abandon it. | 08:43 |
| Tony31 | yoctozepto - looks like the config is not getting updated in /etc/kolla | 08:46 |
| mnasiadka | I lived in the ideal world - that we could make infra images and openstack images a bit independent from each other | 08:46 |
| yoctozepto | Tony31: on the deployer? then it's up to kayobe | 08:46 |
| Tony31 | yea | 08:47 |
| yoctozepto | hrw: reproducible? each command may fail randomly :-) | 08:47 |
| Tony31 | yoctozepto `neutron_external_interface: "p-eth0-ovs"` is the external bridge for linking openstack to the real world? This will be why it;s not working, then. I have now set eth5. But ideally I would like to run `ovs` commands to view what the node is doing, but I am unable to do it at the moment because sudo is prompting for a password and I dont | 08:48 |
| Tony31 | know what is the password - any idea? | 08:48 |
| Tony31 | maybe mnasiadka may know? | 08:49 |
| *** rgogunskiy has joined #openstack-kolla | 08:50 | |
| yoctozepto | Tony31: it's actually a list of such, yeah | 08:51 |
| yoctozepto | though usually used with one | 08:51 |
| yoctozepto | could be better refined and documented | 08:51 |
| Tony31 | ah - I had 2 x put in there initially | 08:51 |
| Tony31 | so I have probably broken it from the onset | 08:52 |
| hrw | https://paste.centos.org/view/97d928b9 is more or less what is needed to get same binary/source images for any infra stuff. without any renaming of images | 08:52 |
| yoctozepto | mnasiadka: approved the ceph removal | 08:53 |
| yoctozepto | mnasiadka: just do followups | 08:53 |
| hrw | but will it make it use less space without merging binary/source jobs? not tested | 08:53 |
| hrw | or even easier way: introduce non-infra-base image which just do ENV injects. | 08:54 |
| yoctozepto | hrw: would have to pull existing to not waste space | 08:54 |
| hrw | yoctozepto: have you ever built with --skip-existing? | 08:55 |
| yoctozepto | hrw: it broke last time I used it | 08:55 |
| Tony31 | what I am finding though, is I manually changed /etc/kolla/inventory/overcloud/host_vars/node to update the bridge there, but when running reconfigure it gets reverted | 08:55 |
| mnasiadka | Tony31: I wouldn't override the neutron_external_interface when using kayobe, if I remember correctly it has some logic to create ovs veth pair and connecting it to the bridge just based on the physical interfaces you tell it to configure | 08:55 |
| mnasiadka | yoctozepto: will do | 08:56 |
| hrw | and the whole 'would have to pull existing' exactly same shit we have now. One of goals of doing infra images was to use less network on CI | 08:56 |
| Tony31 | mnasiadka - not overriding, I am setting what I want and it's not being used | 08:56 |
| Tony31 | `external_net_names:` in networks.yml | 08:57 |
| hrw | not to mention that chaining jobs == even longer CI queues | 08:57 |
| yoctozepto | hrw: hmm, but zuul has dependent jobs and promote can use artefacts from gate | 08:59 |
| yoctozepto | hrw: maybe there is solution also for us | 08:59 |
| yoctozepto | hrw: could help speaking in infra or zuul | 09:00 |
| yoctozepto | (or both) | 09:00 |
| yoctozepto | ideally there would be one job for base | 09:00 |
| yoctozepto | one for infra | 09:00 |
| yoctozepto | one for binary | 09:00 |
| yoctozepto | one for source | 09:01 |
| yoctozepto | all 3 depending on base | 09:01 |
| hrw | why I even started that... ;D | 09:01 |
| yoctozepto | hrw: indeed | 09:01 |
| mnasiadka | yeah well, try to finish what you started | 09:01 |
| hrw | https://paste.centos.org/view/d6cb7453 - the easiest infra patch ever | 09:03 |
| hrw | let me send it to gerrit, test and abandon | 09:04 |
| hrw | but first local build | 09:07 |
| hrw | https://paste.centos.org/view/c99a262e is whole git patch | 09:08 |
| *** Tony31 has quit IRC | 09:12 | |
| hrw | non-infra/debian-binary-non-infra-base 9.1.0 275dcb79d6fd 3 minutes ago 269MB | 09:16 |
| *** k_mouza has joined #openstack-kolla | 09:26 | |
| openstackgerrit | Radosław Piliszek proposed openstack/kolla-ansible master: Fix Prometheus mysqld exporter pointing to VIP address https://review.opendev.org/707582 | 09:27 |
| mnasiadka | hrw: so noninfra is only kolla-toolbox? | 09:31 |
| hrw | mnasiadka: openstack-base and kolla-toolbox | 09:34 |
| mnasiadka | hrw: ah-ha, wouldn't it be easier to just put kolla-toolbox under openstack-base? :) | 09:43 |
| hrw | mnasiadka: go for it! | 09:45 |
| mnasiadka | hrw: I see you are tired :) | 09:45 |
| hrw | mnasiadka: no, I looked at it in past. and not only me | 09:45 |
| mnasiadka | I see - https://review.opendev.org/#/c/435023/ | 09:46 |
| patchbot | patch 435023 - kolla - Reparent kolla-toolbox from openstack-base (ABANDONED) - 17 patch sets | 09:46 |
| mnasiadka | once I finish with ceph stuff and rip out kolla parts - I can look into this | 09:47 |
| hrw | mnasiadka: the thing is that kolla-toolbox is source always nowadays. | 09:48 |
| openstackgerrit | Michal Nasiadka proposed openstack/kolla-ansible stable/rocky: WIP: Change /run bind mount for neutron/openvswitch https://review.opendev.org/707598 | 09:56 |
| openstackgerrit | Michal Nasiadka proposed openstack/kolla-ansible stable/rocky: WIP: Change /run bind mount for neutron/openvswitch https://review.opendev.org/707598 | 09:58 |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: WIP: introduce non-infra-base image https://review.opendev.org/707599 | 10:03 |
| hrw | and commented | 10:06 |
| openstackgerrit | Merged openstack/kolla-ansible master: Remove kolla-ceph https://review.opendev.org/704309 | 10:09 |
| mnasiadka | yay ^^ | 10:10 |
| *** gfidente|afk is now known as gfidente | 10:11 | |
| yoctozepto | mnasiadka: yay, follow up :-) | 10:19 |
| mnasiadka | yeah, in a minute - first a change to Kolla to remove that crap | 10:20 |
| openstackgerrit | Michal Nasiadka proposed openstack/kolla master: Remove kolla container images https://review.opendev.org/707603 | 10:20 |
| openstackgerrit | Mark Goddard proposed openstack/kayobe master: WIP: CI: Test SSH connectivity to deployed instances https://review.opendev.org/707604 | 10:22 |
| yoctozepto | mnasiadka: commented :-) | 10:24 |
| openstackgerrit | Radosław Piliszek proposed openstack/kolla-ansible master: CI: Generate ARA HTML https://review.opendev.org/707450 | 10:34 |
| mnasiadka | yoctozepto: I guess it would make sense to remove the ceph images in Docker Hub? :) | 10:37 |
| *** well100 has joined #openstack-kolla | 10:37 | |
| well100 | Good Morning | 10:38 |
| yoctozepto | mnasiadka: yeah, but that is part of general cleanup effort | 10:39 |
| hrw | well100: morning | 10:40 |
| openstackgerrit | Michal Nasiadka proposed openstack/kolla master: Remove kolla container images https://review.opendev.org/707603 | 10:40 |
| hrw | mnasiadka: ready my passive-agressive comment ;d | 10:41 |
| well100 | have any one a image for octavia? | 10:42 |
| hrw | well100: octavia is one of images we build and publish | 10:43 |
| well100 | a amphora image | 10:44 |
| yoctozepto | mnasiadka: mention deprec and we are golden | 10:47 |
| hrw | mnasiadka: and do not forget to remove rest of images too | 10:48 |
| hrw | bbiab | 10:49 |
| yoctozepto | hrw: are there any left? ;O | 10:51 |
| mnasiadka | yoctozepto: they were never deprecated | 10:59 |
| mnasiadka | yoctozepto: we only deprecated K-A kolla-ceph :D | 10:59 |
| mnasiadka | yoctozepto: I spoke with mgoddard - we just remove since k-a was only consumer. | 10:59 |
| yoctozepto | mnasiadka: all right | 11:00 |
| yoctozepto | mgoddard, mnasiadka: so we are golden | 11:00 |
| *** zhanglong has quit IRC | 11:02 | |
| *** zhanglong has joined #openstack-kolla | 11:03 | |
| hrw | ok, as my sarcasm was not clear enough I commented and gave -1 | 11:04 |
| yoctozepto | hrw: hahaha | 11:06 |
| *** well100 has quit IRC | 11:06 | |
| openstackgerrit | Michal Nasiadka proposed openstack/kolla master: Remove kolla Ceph container images https://review.opendev.org/707603 | 11:07 |
| mnasiadka | hrw: thanks for noticing - even picky yoctozepto didn't notice it :D | 11:08 |
| hrw | and now +2 | 11:08 |
| yoctozepto | mnasiadka: yeah, I read "ceph" instead of "container" | 11:08 |
| mnasiadka | good, now I can go with the followups | 11:13 |
| hrw | Total reclaimed space: 105.9GB | 11:21 |
| hrw | just small 'docker image prune -a' | 11:21 |
| yoctozepto | not that small | 11:25 |
| *** hamzy_ is now known as hamzy | 11:26 | |
| *** rlljorge has joined #openstack-kolla | 11:28 | |
| rlljorge | Hello, Guys ! Good morning. | 11:28 |
| rlljorge | Someone using cloudkitty in stein ? I cannot get values for reports .... | 11:29 |
| openstackgerrit | Michal Nasiadka proposed openstack/kolla-ansible master: WIP: CI: Add CentOS 8 ceph-ansible job https://review.opendev.org/706886 | 11:30 |
| openstackgerrit | Radosław Piliszek proposed openstack/kolla-ansible master: Revert "Upgrade virtualenv in pre" https://review.opendev.org/707622 | 11:31 |
| openstackgerrit | Radosław Piliszek proposed openstack/kolla master: Revert "Upgrade virtualenv in pre" https://review.opendev.org/707623 | 11:32 |
| hrw | yay! keystone built for debian/source in stein | 11:32 |
| hrw | just one py2 package needed to be built | 11:32 |
| openstackgerrit | Alfredo Moralejo proposed openstack/kolla master: WIP Use StorageSIG repos for Ceph in CentOS8 https://review.opendev.org/707338 | 11:33 |
| *** Wellie has joined #openstack-kolla | 11:34 | |
| openstackgerrit | Mark Goddard proposed openstack/kayobe master: Fix sphinx8 script used by pep8 https://review.opendev.org/707624 | 11:34 |
| *** ivve has joined #openstack-kolla | 11:39 | |
| *** iniazi has quit IRC | 11:44 | |
| mnasiadka | yoctozepto: they fixed virtualenv finally? :) | 11:46 |
| yoctozepto | mnasiadka: stable looks unblocked, just testing | 11:46 |
| yoctozepto | mnasiadka: honestly no idea how it moves | 11:46 |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla stable/stein: base: add 'py2hacks' repository with python-ldappool https://review.opendev.org/707628 | 11:59 |
| *** kplant has joined #openstack-kolla | 12:09 | |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla master: add Linaro repo with fixed libvirt https://review.opendev.org/706705 | 12:13 |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla stable/train: add Linaro repo with fixed libvirt https://review.opendev.org/707631 | 12:17 |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla stable/stein: add Linaro repo with fixed libvirt https://review.opendev.org/707632 | 12:19 |
| openstackgerrit | Marcin Juszkiewicz proposed openstack/kolla stable/stein: base: add 'py2hacks' repository with python-ldappool https://review.opendev.org/707628 | 12:24 |
| openstackgerrit | Michal Nasiadka proposed openstack/kolla master: Remove kolla Ceph container images https://review.opendev.org/707603 | 12:26 |
| mnasiadka | bloody ADD directive in Dockerfile :) | 12:27 |
| yoctozepto | mnasiadka: we need them both | 12:27 |
| yoctozepto | for swift | 12:27 |
| openstackgerrit | Radosław Piliszek proposed openstack/kolla-ansible master: CI: Generate ARA HTML https://review.opendev.org/707450 | 12:29 |
| hrw | those libvirt/py2hacks patches had to be done that way ;( | 12:30 |
| openstackgerrit | Radosław Piliszek proposed openstack/kolla-ansible master: CI: Generate ARA HTML https://review.opendev.org/707450 | 12:43 |
| openstackgerrit | Mark Goddard proposed openstack/kayobe master: WIP: CI: Test SSH connectivity to deployed instances https://review.opendev.org/707604 | 12:46 |
| yoctozepto | mgoddard: k-a could use that as well :-) | 12:47 |
| yoctozepto | especially to verify e.g. ovn | 12:48 |
| *** skramaja has quit IRC | 12:56 | |
| *** zhanglong has quit IRC | 13:00 | |
| *** zhanglong has joined #openstack-kolla | 13:04 | |
| *** tumble has joined #openstack-kolla | 13:05 | |
| cosmicsound | so for ceph best would be to start from ceph-ansible instead of using anything kolla ceph related | 13:05 |
| *** kplant has quit IRC | 13:06 | |
| yoctozepto | cosmicsound: we don't want to be ceph deployers as ceph is an independent project - ceph-ansible is presumably going to die soon and the future is ceph-orchestrator running off installed mgr | 13:09 |
| *** kplant has joined #openstack-kolla | 13:10 | |
| yoctozepto | cosmicsound: so best just follow official docs on ceph, it's one (or two) orders of magnitude easier to deploy ceph in general than openstack :-) | 13:11 |
| kplant | ^ | 13:12 |
| kplant | i'm about to upgrade from s->t in my largest prod region | 13:12 |
| kplant | wish me luck | 13:12 |
| hrw | kplant: fingers crossed | 13:18 |
| yoctozepto | kplant: /me wishing luck | 13:20 |
| yoctozepto | kplant: my cluster is small, below 20 nodes, but managed not to break it :-) | 13:20 |
| kplant | in my last s->t upgrade i noticed a bunch of dead nova endpoints were left behind - easy cleanup | 13:20 |
| yoctozepto | dead how? | 13:21 |
| kplant | if it does it again i'll raise it as a bug, if not already known | 13:21 |
| kplant | some of the old things that aren't needed anymore like consoleauth | 13:21 |
| yoctozepto | ah, these | 13:21 |
| yoctozepto | we don't clean too well | 13:21 |
| kplant | :D | 13:21 |
| yoctozepto | though they were gone before train | 13:21 |
| kplant | really easy to find with `nova service-list` so it's not a big deal | 13:22 |
| yoctozepto | kplant: how big that cluster? | 13:22 |
| kplant | ~10,000 hypervisors | 13:22 |
| yoctozepto | the one you are planning to destroy | 13:22 |
| yoctozepto | upgrade* | 13:22 |
| kplant | rofl | 13:22 |
| yoctozepto | oh my | 13:22 |
| yoctozepto | quite nice | 13:22 |
| kplant | totally kidding | 13:22 |
| kplant | more like 10.000 hypervisors | 13:23 |
| kplant | :-) | 13:23 |
| hrw | kplant: any scaling issues? | 13:23 |
| *** pbing19 has quit IRC | 13:23 | |
| yoctozepto | kplant: you mean . is a decimal point? ;D | 13:23 |
| kplant | i do | 13:23 |
| kplant | as in ten | 13:23 |
| kplant | not ten thousand | 13:23 |
| yoctozepto | then mine is biggur | 13:23 |
| kplant | >:( | 13:24 |
| kplant | all of the app teams are slow to migrate, so we're slow to buy | 13:24 |
| kplant | we'll get there one decade | 13:24 |
| yoctozepto | I see | 13:24 |
| kplant | the scariest part, mariadb, was successful | 13:29 |
| yoctozepto | kplant: https://review.opendev.org/705414 no doubt | 13:32 |
| patchbot | patch 705414 - kolla-ansible (stable/train) - Fix multiple issues with MariaDB handling (MERGED) - 1 patch set | 13:32 |
| kplant | very nice, i wonder if that's what randomly burned me on a stein deploy a few months ago | 13:35 |
| yoctozepto | kplant: could be | 13:36 |
| cosmicsound | right | 13:36 |
| cosmicsound | thanks yoctozepto it gives a clearer sense. | 13:37 |
| *** sluna has quit IRC | 13:37 | |
| yoctozepto | cosmicsound: yw :-) | 13:37 |
| cosmicsound | I am wondering if I should just use cinder and swift | 13:37 |
| cosmicsound | without ceph | 13:38 |
| kplant | you still need a backend for cinder | 13:38 |
| yoctozepto | swift can only be used as object store | 13:38 |
| cosmicsound | Yes would need it for Magnum | 13:39 |
| cosmicsound | and k8s in future | 13:39 |
| yoctozepto | ah, future :-) | 13:39 |
| cosmicsound | always plan ahead or think ahead | 13:39 |
| cosmicsound | kplant , will go with LVM / SCSI | 13:39 |
| hrw | NFS! | 13:40 |
| * kplant slaps hrw around with a large non-redundant trout | 13:41 | |
| yoctozepto | hahahaha | 13:41 |
| * yoctozepto slaps yoctozepto around a bit with a large trout | 13:41 | |
| kplant | upgrade done | 13:42 |
| kplant | \o/ | 13:42 |
| yoctozepto | I forgot about /slap | 13:42 |
| kplant | the ol' mirc days | 13:42 |
| yoctozepto | now I'm gonna abuse it | 13:42 |
| yoctozepto | :D | 13:42 |
| kplant | it did leave six 'old' nova service endpoints again | 13:42 |
| kplant | do you care? | 13:42 |
| hrw | ~curse kplant for reminding /slap | 13:43 |
| * cosmicsound slaps kplant around a bit with a large trout | 13:43 | |
| cosmicsound | wh000t it works | 13:43 |
| yoctozepto | kplant: yeah, report a bug on it, we might take some action | 13:43 |
| yoctozepto | we planned more cleaning up | 13:43 |
| yoctozepto | but priorities are priorities | 13:43 |
| kplant | yeah, this is totally superficial | 13:44 |
| cosmicsound | can someone explain why the explicit external TLS it still resolves me an SSL from kolla generated | 13:45 |
| *** icarusfactor has quit IRC | 13:49 | |
| kplant | yoctozepto, hrw, mnasiadka, mgoddard, (everyone else involved): sincerely great job, every single stein -> train upgrade has been painless for me | 13:50 |
| mgoddard | kplant: w00t | 13:50 |
| cosmicsound | kplant , congrats! | 13:50 |
| kplant | ty! | 13:52 |
| kplant | https://bugs.launchpad.net/kolla-ansible/+bug/1863107 <--- feel free to deprioritize, purely cosmetic from what i can tell | 13:52 |
| openstack | Launchpad bug 1863107 in kolla-ansible "Stein->Train Upgrade : Old nova services not cleaned" [Undecided,New] | 13:52 |
| hrw | kplant: I just try to not break things too much (or fix if someone managed to break) | 13:54 |
| yoctozepto | kplant: appreciation always appreciated :-) | 14:03 |
| mnasiadka | yoctozepto: "we need them both" - both what? :) | 14:06 |
| * hrw off | 14:07 | |
| yoctozepto | mnasiadka: both files | 14:18 |
| yoctozepto | in image | 14:18 |
| *** zhanglong has quit IRC | 14:19 | |
| mnasiadka | yoctozepto: ah, find_disks is used by swift? oh lol | 14:19 |
| mnasiadka | let me add it back | 14:20 |
| openstackgerrit | Michal Nasiadka proposed openstack/kolla master: Remove kolla Ceph container images https://review.opendev.org/707603 | 14:20 |
| openstackgerrit | Michal Nasiadka proposed openstack/kolla-ansible master: WIP: CI: Add CentOS 8 ceph-ansible job https://review.opendev.org/706886 | 14:27 |
| *** rlljorge has quit IRC | 14:27 | |
| openstackgerrit | Mark Goddard proposed openstack/kayobe master: WIP: CI: Test SSH connectivity to deployed instances https://review.opendev.org/707604 | 14:30 |
| openstackgerrit | Michal Nasiadka proposed openstack/kolla-ansible master: Fix swift entries in group_vars/globals.yml https://review.opendev.org/707652 | 14:33 |
| openstackgerrit | Michal Nasiadka proposed openstack/kolla-ansible master: CI: Remove ceph-base and remove file check from ceph-ansible-base https://review.opendev.org/707655 | 14:35 |
| *** hu_berlin_kalle has quit IRC | 14:35 | |
| openstackgerrit | Mark Goddard proposed openstack/kayobe master: CI: Redirect stderr to ansible logs https://review.opendev.org/707657 | 14:36 |
| openstackgerrit | Michal Nasiadka proposed openstack/kolla-ansible master: OVN Support https://review.opendev.org/696841 | 14:38 |
| *** riuzen has joined #openstack-kolla | 14:38 | |
| riuzen | Hi...i'm trying multinode deployment, how to deal with this error? "FAILED! => {"changed": false, "msg": "Can not parse the inner module output: localhost | SUCCESS => {\n \"changed\": false, \n \"user\": \"haproxy\"\n}\n"}" | 14:39 |
| openstackgerrit | Michal Nasiadka proposed openstack/kolla-ansible master: CI: Remove ceph-base and remove file check from ceph-ansible-base https://review.opendev.org/707655 | 14:42 |
| *** pbing19 has joined #openstack-kolla | 14:49 | |
| *** TrevorV has joined #openstack-kolla | 14:55 | |
| openstackgerrit | Michal Nasiadka proposed openstack/kolla-ansible master: OVN Support https://review.opendev.org/696841 | 15:02 |
| openstackgerrit | Mark Goddard proposed openstack/kayobe master: Docs: fix libvirt connection URI for tenks https://review.opendev.org/707667 | 15:07 |
| Wellie | who checks the code that was submitted for verification? | 15:08 |
| mgoddard | riuzen: that probably means you are using the wrong version of kolla-ansible for your images | 15:16 |
| openstackgerrit | Michal Nasiadka proposed openstack/kolla-ansible master: WIP: CI: Add CentOS 8 ceph-ansible job https://review.opendev.org/706886 | 15:16 |
| *** pbing19 has quit IRC | 15:16 | |
| mgoddard | riuzen: for stein images you need kolla-ansible 8.x.y | 15:16 |
| openstackgerrit | Merged openstack/kayobe master: Fix sphinx8 script used by pep8 https://review.opendev.org/707624 | 15:17 |
| *** primeministerp has joined #openstack-kolla | 15:19 | |
| openstackgerrit | Pierre Riteau proposed openstack/kayobe master: Docs: fix libvirt connection URI for tenks https://review.opendev.org/707667 | 15:21 |
| riuzen | @mgoddrd ahh..i see, thank you. | 15:22 |
| mgoddard | Wellie: anyone can register and submit a review. There is a team of core reviewers who have permission (+2). Once there are two +2 votes, a core reviewer can vote +W to submit the change for a gate check and merge (if successful) | 15:24 |
| Wellie | mgoddard: thx for information :-) | 15:24 |
| mgoddard | np | 15:25 |
| mgoddard | Wellie: core reviewers can be seen on https://wiki.openstack.org/wiki/Kolla | 15:25 |
| hrw | folks: which version of kibana we use? | 15:25 |
| mgoddard | hrw: 5.x | 15:26 |
| hrw | uf. | 15:26 |
| mgoddard | hrw: need to upgrade to 6.x for centos 8 | 15:26 |
| hrw | have to build it for stein/train images | 15:26 |
| Wellie | mgoddard: thx for infomation :-) | 15:26 |
| mgoddard | hrw: debian? | 15:27 |
| hrw | yes | 15:27 |
| hrw | mgoddard: arm64 debian | 15:27 |
| hrw | [ 109s] dpkg-deb: building package 'kibana' in '../kibana_5.6.9-0linaro1_all.deb'. | 15:29 |
| hrw | 5.6.16 is current... | 15:29 |
| *** pbing19 has joined #openstack-kolla | 15:30 | |
| openstackgerrit | Mark Goddard proposed openstack/kayobe master: Fix error logging of dev/functions https://review.opendev.org/705180 | 15:32 |
| *** cah_link has quit IRC | 15:35 | |
| hrw | mgoddard: thanks for reminding about overrides ;D | 15:46 |
| mgoddard | hrw: might be a better solution in this case | 15:46 |
| hrw | for now I will reuse patches from gerrit. will work on overrides after vacations | 15:47 |
| *** primeministerp has quit IRC | 15:51 | |
| *** priteau has joined #openstack-kolla | 15:52 | |
| openstackgerrit | Michal Nasiadka proposed openstack/kolla-ansible master: WIP: CI: Add CentOS 8 ceph-ansible job https://review.opendev.org/706886 | 15:55 |
| openstackgerrit | Michal Nasiadka proposed openstack/kolla-ansible master: WIP: CI: Add CentOS 8 ceph-ansible job https://review.opendev.org/706886 | 15:55 |
| *** Trevor_V has joined #openstack-kolla | 15:55 | |
| *** TrevorV has quit IRC | 15:58 | |
| *** rgogunskiy has quit IRC | 16:03 | |
| openstackgerrit | Mark Goddard proposed openstack/kayobe master: Docs: Configure firewall to allow testing of baremetal https://review.opendev.org/707678 | 16:03 |
| * hrw off | 16:08 | |
| mnasiadka | yoctozepto: https://review.opendev.org/#/c/707652/ | 16:12 |
| patchbot | patch 707652 - kolla-ansible - Fix swift entries in group_vars/globals.yml - 1 patch set | 16:12 |
| *** bersace has quit IRC | 16:14 | |
| *** bersace has joined #openstack-kolla | 16:15 | |
| Wellie | RITICAL designate.manage.pool [designate-manage - - - - -] Bad ID Supplied for pool. pool_id have any one idea? | 16:33 |
| *** bengates has quit IRC | 16:33 | |
| openstackgerrit | Mark Goddard proposed openstack/kayobe master: WIP: CI: Test SSH connectivity to deployed instances https://review.opendev.org/707604 | 16:37 |
| openstackgerrit | Merged openstack/kolla-ansible master: CI: Ignore zuul.d https://review.opendev.org/706836 | 16:40 |
| openstackgerrit | Merged openstack/kolla-ansible master: CI: Fix deploy guide jobs https://review.opendev.org/706834 | 16:43 |
| openstackgerrit | Merged openstack/kolla-ansible stable/train: Ironic: fix documentation https://review.opendev.org/707090 | 16:43 |
| admin0 | if i have to use something like this: compute01 neutron_external_interface=eth0 api_interface=em1 storage_interface=em1 tunnel_interface=em1 , does the interface need to have an ip ? | 16:44 |
| admin0 | or does kolla assign ip automatically ? | 16:45 |
| admin0 | neutron_external_interface -- this does not need an IP right ? | 16:45 |
| mgoddard | admin0: all need an IP except neutron_external_interface | 16:50 |
| admin0 | ok | 16:50 |
| k_mouza | hello all! I have the mountpoint issue that's mentioned in this bug: https://bugs.launchpad.net/kolla/+bug/1783978 , using kolla-ansible stable/stein with centos-binary images. I'm seeing this behaviour when using cinder backends with shared bind mounts (tried with NFS and Quobyte) | 16:54 |
| openstack | Launchpad bug 1783978 in kolla "nova_compute mount nfs backend every time when vm was launched " [Undecided,Expired] | 16:54 |
| openstackgerrit | Mark Goddard proposed openstack/kayobe master: Switch from shade to openstacksdk https://review.opendev.org/707689 | 16:54 |
| openstackgerrit | Mark Goddard proposed openstack/kayobe master: WIP: CentOS 8 https://review.opendev.org/707690 | 16:54 |
| k_mouza | Anyone else seen this and might have an idea for a solution? Thanks a lot ! | 16:54 |
| yoctozepto | k_mouza: you can open the ticket there | 16:56 |
| yoctozepto | k_mouza: anything interesting in logs? | 16:57 |
| openstackgerrit | Mark Goddard proposed openstack/kayobe master: CI: Redirect stderr to ansible logs https://review.opendev.org/707657 | 16:57 |
| openstackgerrit | Mark Goddard proposed openstack/kayobe master: WIP: CentOS 8 https://review.opendev.org/707690 | 16:58 |
| k_mouza | yoctozepto : nothing out of the ordinary in the logs! I'm pretty sure it has to do with the /var/lib/nova/mnt bind mount : http://paste.openstack.org/show/789538/ because this is the only differentiator with the other backends that work fine (ceph for example) | 17:02 |
| yoctozepto | k_mouza: hmm, it's shared for the very reason to not lose mounts | 17:05 |
| sorin-mihai | fatal: [localhost]: FAILED! => {"action": "uri", "attempts": 20, "changed": false, "msg": "Failed to validate the SSL certificate for 10.2.1.253:9200. Make sure your managed systems have a valid CA certificate installed. You can use validate_certs=False if you do not need to confirm the servers identity but this is unsafe and not recommended. Paths checked for this platform: /etc/ssl/certs, /etc/pki/ca-trust/extracted/pem, | 17:10 |
| sorin-mihai | /etc/pki/tls/certs, /usr/share/ca-certificates/cacert.org, /etc/ansible. The exception msg was: hostname '10.2.1.253' doesn't match 'internal.stack'.", "status": -1, "url": "https://10.2.1.253:9200/.kibana"} | 17:10 |
| sorin-mihai | i get that while trying to deploy. the certs have been generated, they're in /etc/kolla/certificates, why would this one be requested from anywhere else but the /etc/kolla/certificates? | 17:14 |
| k_mouza | yoctozepto: yeah, not sure why this happens to be honest... I even tried stracing the docker restart command to see if I can find anything useful in there! I've also commented on the https://bugs.launchpad.net/kolla/+bug/1783978 bug | 17:16 |
| openstack | Launchpad bug 1783978 in kolla "nova_compute mount nfs backend every time when vm was launched " [Undecided,Expired] | 17:16 |
| *** paddyone1ll has joined #openstack-kolla | 17:19 | |
| *** paddyone1ll is now known as paddyoniell | 17:19 | |
| sorin-mihai | where should i "use" validate_certs=False, that's really unclear | 17:21 |
| openstackgerrit | Mark Goddard proposed openstack/kayobe master: WIP: CentOS 8 https://review.opendev.org/707690 | 17:26 |
| *** rlljorge has joined #openstack-kolla | 17:31 | |
| rlljorge | Hello there someone using cloudkitty ? | 17:31 |
| *** negronjl has quit IRC | 17:32 | |
| *** evrardjp has quit IRC | 17:34 | |
| *** evrardjp has joined #openstack-kolla | 17:34 | |
| *** negronjl has joined #openstack-kolla | 17:34 | |
| *** k_mouza has quit IRC | 17:36 | |
| openstackgerrit | Merged openstack/kolla-ansible stable/train: Python 3: Use distro_python_version for dev mode https://review.opendev.org/705928 | 17:39 |
| admin0 | "Kolla can function with or without a local registry, however for a multinode deployment some type of registry is mandatory" -- if i am doing multinode, i need to have a registry ? | 17:41 |
| admin0 | and the link to the pokey registry takes me to https://github.com/moby/moby/issues/14018 -- which is like ??? | 17:42 |
| admin0 | Proposal: push performance improvements #14018 | 17:42 |
| openstackgerrit | Mark Goddard proposed openstack/kayobe master: WIP: Allow OVS bridges to connect directly to interface https://review.opendev.org/705037 | 17:42 |
| admin0 | how is this relevant to me setting up some sort of registry for getting multinode done | 17:42 |
| yoctozepto | admin0: "mandatory" is an exaggeration | 17:47 |
| yoctozepto | i'd say "strongly recommended" ;-) | 17:48 |
| admin0 | yes .. but where is the link to set it up :) the link there points to some proposal without saying "hey operator .. welcome .. in this page, you will setup a registry strongly recommended to run kolla multinode install" | 17:48 |
| admin0 | rather, it gives me a bunch of stuff which i have no idea what i am looking for | 17:49 |
| admin0 | https://docs.openstack.org/kolla-ansible/latest/user/multinode.html --it has a line For more information reference pokey registry. which goes to some moby/moby on gitlab and that issue | 17:50 |
| *** vmixor has quit IRC | 17:50 | |
| admin0 | also on the controller, compute etc, do i need to install/configure docker beforehand or the bootstrap process will take care of it ? | 17:50 |
| yoctozepto | bootstrap | 17:51 |
| admin0 | ok | 17:51 |
| admin0 | then the only thing i am missing is how to setup my own registry | 17:51 |
| admin0 | i have a gitlab .. does that count .. ? | 17:51 |
| admin0 | https://docs.openstack.org/kolla-ansible/latest/user/multinode.html line "cd kolla | 17:52 |
| admin0 | ; tools/start-registry" -- where is this ? | 17:52 |
| admin0 | its not in the kolla venv | 17:52 |
| admin0 | all hosts have python3, but when doing the multinode all -m ping i get "[DEPRECATION WARNING]: Distribution Ubuntu 18.04 on host c16 should use /usr/bin/python3, but is using /usr/bin/python for backward | 17:59 |
| admin0 | compatibility with prior Ansible releases" -- how is this solved ? | 17:59 |
| *** gfidente is now known as gfidente|afk | 17:59 | |
| admin0 | help me with the container thing .. the documentation does not help | 18:02 |
| *** gfidente|afk is now known as gfidente | 18:03 | |
| openstackgerrit | Mark Goddard proposed openstack/kayobe master: WIP: CentOS 8 https://review.opendev.org/707690 | 18:04 |
| openstackgerrit | Dincer Celik proposed openstack/kolla-ansible master: Follow up on Ansible version checking https://review.opendev.org/705171 | 18:06 |
| openstackgerrit | Michal Nasiadka proposed openstack/kolla-ansible master: WIP: CI: Add CentOS 8 ceph-ansible job https://review.opendev.org/706886 | 18:12 |
| openstackgerrit | Michal Nasiadka proposed openstack/kolla-ansible master: CI: Add CentOS 8 ceph-ansible job https://review.opendev.org/706886 | 18:13 |
| *** gfidente is now known as gfidente|afk | 18:14 | |
| *** dking_desktop has joined #openstack-kolla | 18:21 | |
| openstackgerrit | Merged openstack/kolla-ansible master: Fix swift entries in group_vars/globals.yml https://review.opendev.org/707652 | 18:26 |
| *** riuzen has quit IRC | 18:28 | |
| *** igordc has joined #openstack-kolla | 18:28 | |
| *** igordc has quit IRC | 18:28 | |
| *** priteau has quit IRC | 18:29 | |
| *** tonythomas has quit IRC | 18:30 | |
| *** k_mouza has joined #openstack-kolla | 18:34 | |
| *** k_mouza has quit IRC | 18:39 | |
| openstackgerrit | Dincer Celik proposed openstack/kolla master: Upgrade RabbitMQ to 3.8 https://review.opendev.org/694799 | 18:44 |
| openstackgerrit | Dincer Celik proposed openstack/kolla master: WIP: Upgrade Elasticsearch and Kibana to 6.x https://review.opendev.org/647748 | 18:49 |
| *** negronjl has quit IRC | 18:56 | |
| admin0 | registry is also done | 18:58 |
| admin0 | ready for deployment i guess | 18:58 |
| *** negronjl has joined #openstack-kolla | 19:01 | |
| admin0 | what is a network_address_family ? | 19:02 |
| admin0 | oh .. nvm | 19:02 |
| openstackgerrit | Dincer Celik proposed openstack/kolla-ansible master: Introduce /etc/timezone to containers https://review.opendev.org/704258 | 19:03 |
| osmanlicilegi | mgoddard: I have been able to read the meeting notes right now. +2 for kolla-sig, it's a great idea. | 19:10 |
| *** pbing19 has quit IRC | 19:16 | |
| *** pbing19 has joined #openstack-kolla | 19:17 | |
| admin0 | how to tell kolla-ansible to target only 1 host ? | 19:17 |
| admin0 | -l does not seem to work | 19:17 |
| admin0 | found it . | 19:18 |
| admin0 | hmm.. so it tries to pull from the local registry .. but what process inserts the image to the local registry ? | 19:24 |
| openstackgerrit | Merged openstack/kayobe master: Fix error logging of dev/functions https://review.opendev.org/705180 | 19:27 |
| yoctozepto | mnasiadka: I approved ceph change in zuul but needs you to check c-a jobs because they mostly failed | 19:28 |
| yoctozepto | admin0: you use kolla to build images and push to this local registry | 19:29 |
| admin0 | is there a document for this ? | 19:29 |
| mnasiadka | yoctozepto: virtualenv fault, fix is in the centos8 ceph-ansible change | 19:29 |
| admin0 | right now i # the local registry and just did a pull .. so its pulling from somewhere | 19:30 |
| openstackgerrit | Zhuo Zhen proposed openstack/kolla-ansible master: Configurable logrotate rotation interval and count https://review.opendev.org/706616 | 19:30 |
| yoctozepto | mnasiadka: ubuntu upgrade too? ood | 19:30 |
| yoctozepto | odd* | 19:30 |
| admin0 | yoctozepto, is this the one: https://docs.openstack.org/kolla/latest/admin/image-building.html | 19:30 |
| Wellie | yes | 19:30 |
| yoctozepto | admin0: if you set it up to act as a proxy, it could | 19:30 |
| yoctozepto | admin0: yeah, this | 19:31 |
| Wellie | rlljorge: yes | 19:31 |
| admin0 | yoctozepto, i did "docker run -d -p 5000:5000 --name registry registry:2 | 19:31 |
| admin0 | " to setup a registry in the deployment host | 19:31 |
| admin0 | how to make it act as a proxy ? | 19:31 |
| mnasiadka | yoctozepto: odd, will look into it later :) | 19:31 |
| yoctozepto | admin0: https://docs.docker.com/registry/recipes/mirror/ | 19:32 |
| yoctozepto | mnasiadka: thanks | 19:32 |
| admin0 | failed here: TASK [swift : Fail if ring files don't exist] | 19:37 |
| admin0 | doesn't kolla configure swift also ? | 19:37 |
| *** xinliang has quit IRC | 19:37 | |
| *** k_mouza has joined #openstack-kolla | 19:37 | |
| admin0 | or do i need to do extra stuff for swift and its not as easy as all in one where i just point it to the controller ? | 19:37 |
| admin0 | https://docs.openstack.org/kolla-ansible/queens/reference/swift-guide.html --is this valid for train release as well ? i could not find train one | 19:38 |
| openstackgerrit | Zhuo Zhen proposed openstack/kolla-ansible master: Configurable logrotate rotation interval and count https://review.opendev.org/706616 | 19:39 |
| *** k_mouza has quit IRC | 19:42 | |
| mnasiadka | yoctozepto: https://zuul.opendev.org/t/openstack/build/b2c052b2fb17496c815ba255d63f9c67/log/primary/logs/kolla/all-CRITICAL.txt#2 | 19:44 |
| yoctozepto | admin0: https://docs.openstack.org/kolla-ansible/train/reference/storage/swift-guide.html | 19:44 |
| mnasiadka | yoctozepto: ubuntu fails on wsrep :) | 19:44 |
| yoctozepto | mnasiadka: https://771eb0e01a0d24a7571a-6177775aca2ef9bf86cf10df13a88f01.ssl.cf5.rackcdn.com/707655/2/check/kolla-ansible-ubuntu-source-upgrade-ceph-ansible/b2c052b/primary/logs/kolla/mariadb/mariadb.txt | 19:45 |
| yoctozepto | they lost one node indeed | 19:45 |
| yoctozepto | but it's good | 19:45 |
| admin0 | yoctozepto, what is storage node ? is it swift node or cinder node ? | 19:45 |
| yoctozepto | we stay with mostly real issues with networking | 19:45 |
| yoctozepto | memcached | 19:45 |
| yoctozepto | rabbitmq | 19:45 |
| yoctozepto | and mariadb | 19:46 |
| yoctozepto | only keepalived still funky | 19:46 |
| yoctozepto | (with haproxy) | 19:46 |
| yoctozepto | admin0: storage is storage :-) obviously swift in here | 19:46 |
| yoctozepto | the term "swift" is not used because it has components on controllers as do most services | 19:47 |
| admin0 | so if i start with 1 storage node ( for swift) .. and when i need to add 2 more later, do I have to redo the object ring creation ? | 19:49 |
| admin0 | won't it affect existing data ? | 19:49 |
| openstackgerrit | Andreas Jaeger proposed openstack/kayobe master: Cleanup tox.ini deps https://review.opendev.org/705173 | 19:51 |
| yoctozepto | admin0: you have to grow the rings, it's a painful process to scale swift in general (well, compared to ceph) | 19:52 |
| yoctozepto | admin0: any swift integration improvements are welcome | 19:52 |
| admin0 | https://docs.openstack.org/kolla-ansible/train/reference/storage/swift-guide.html == this needs to be done in the storage node right ? | 19:53 |
| admin0 | the step | 19:53 |
| openstackgerrit | Zhuo Zhen proposed openstack/kolla-ansible master: Configurable logrotate rotation interval and count https://review.opendev.org/706616 | 19:59 |
| yoctozepto | admin0: the setup of disks, sure | 19:59 |
| yoctozepto | the setup of rings - irrelevant | 19:59 |
| admin0 | precheck fails saying ring file not found in c18 | 20:01 |
| admin0 | i am going to skip this for the time being | 20:01 |
| admin0 | i have enable cinder = yes and cinder_volume_group = cloud, which is a vg in the storage node .. but precheck still says ""changed": false, "msg": "Please enable at least one backend when enabling Cinder" | 20:03 |
| admin0 | so what am i missing here | 20:03 |
| yoctozepto | backend | 20:03 |
| yoctozepto | e.g. enable_cinder_backend_lvm | 20:03 |
| admin0 | enable_cinder_backend_lvm = yes ? | 20:03 |
| yoctozepto | yeah | 20:04 |
| Wellie | can I set multiple dns records in kolla? (Designate) | 20:08 |
| openstackgerrit | Merged openstack/kolla-ansible master: CI: Remove ceph-base and remove file check from ceph-ansible-base https://review.opendev.org/707655 | 20:12 |
| yoctozepto | Wellie: not sure what you mean, in general we don't limit what deployed projects can do :-) | 20:13 |
| Wellie | yoctozepto: for example designate_ns_record: "ns1.openstack.org, ns2.openstack.org" | 20:15 |
| yoctozepto | Wellie: no, I believe this will fail because it's used in places where only one is expected | 20:17 |
| Wellie | so that I have 2 ns entries in the recordset | 20:17 |
| yoctozepto | Wellie: if this is something designate should handle, please raise a bug report :-) | 20:17 |
| openstackgerrit | Zhuo Zhen proposed openstack/kolla-ansible master: Configurable logrotate rotation interval and count https://review.opendev.org/706616 | 20:18 |
| *** factor has joined #openstack-kolla | 20:18 | |
| admin0 | yoctozepto, there is nothing in globals or docs regarding this step TASK [octavia : Checking certificate files exist for octavia] .. it fails .. but does not tell me what path to put the certs in | 20:21 |
| Wellie | yoctozepto: we can build a for each in kolla-ansible? | 20:23 |
| admin0 | octavia also disabled for now | 20:25 |
| yoctozepto | admin0: octavia needs some love | 20:26 |
| yoctozepto | Wellie: I think you can provide your own template for bind9, but not too sure - not used designate yet | 20:27 |
| admin0 | finally precheck passed .. without octavia and swift that i planned for | 20:27 |
| yoctozepto | admin0: swift works for sure, tested regularly, I implemented its CI scenario based on the docs | 20:27 |
| admin0 | we should also have a list of things that work 100% out of the box and things that "need some more love and care" | 20:27 |
| yoctozepto | admin0: so they are worky | 20:27 |
| yoctozepto | admin0: octavia not really tested | 20:27 |
| yoctozepto | admin0: but some *did* deploy it successfully... | 20:28 |
| admin0 | where can i see this list of things thats in the CI and which are tested to be 100% working ? | 20:28 |
| yoctozepto | admin0: yeah, it's on the list of things to do | 20:28 |
| yoctozepto | admin0: docs wip -> https://etherpad.openstack.org/p/KollaAnsibleScenarios (discussion) | 20:28 |
| *** kplant has quit IRC | 20:29 | |
| yoctozepto | we want to stabilize the scenarios and then document it | 20:29 |
| johnsom | Let the Octavia team know if there are questions we can help answer. Most of us are not kolla users, so can't address that part directly, but we are happy to help with the Octavia side. | 20:29 |
| johnsom | FYI, for certificates, we have a detailed guide: https://docs.openstack.org/octavia/latest/admin/guides/certificates.html but this won't answer where/how kolla handles them. | 20:30 |
| admin0 | johnsom, the issue is take me as an operator .. i am not looking to install octavia as seperate/standalone project .. but its either with openstack-ansible or kolla-ansible .. so no matter how good the project is by itself, it also needs to be something that can be deployed .. i know its not a part of the core team job | 20:30 |
| *** patchbot has quit IRC | 20:31 | |
| admin0 | but operators more look into systems like ansible, kolla etc to make sure we get all the components working without doing manual task | 20:31 |
| johnsom | Yeah, there are somewhere around 16 deployment projects that support Octavia now. We just can't keep up with them all. | 20:31 |
| admin0 | :D | 20:31 |
| admin0 | \o/ | 20:31 |
| yoctozepto | johnsom: please keep up with kolla :-) | 20:31 |
| *** pbing19 has quit IRC | 20:31 | |
| johnsom | Personally I have helped with OpenStack Ansible and Tripleo as those are/were related to my day job. | 20:33 |
| johnsom | That said, I am happy to help the kolla team with questions, etc. | 20:34 |
| yoctozepto | then at least kolla images have worky octavia :-) (due to tripleo there) | 20:34 |
| johnsom | Is there a kolla bug/story I can look at for you? | 20:35 |
| yoctozepto | johnsom: thanks for the offer, we'll reach out when we hit something establishing those scenarios | 20:36 |
| yoctozepto | admin0: ^^ is there? | 20:36 |
| * yoctozepto waving good night (see you tomorrow) | 20:36 | |
| admin0 | its not related to octavia but kolla .. kolla precheck failed on octavia saying no cert is found .. but no documentation or anything where to put that cert | 20:36 |
| johnsom | I think you are using kolla-ansible, so a good first step would be to open a bug https://bugs.launchpad.net/kolla-ansible for the kolla team. | 20:39 |
| *** patchbot has joined #openstack-kolla | 20:41 | |
| *** rlljorge has quit IRC | 20:45 | |
| dking_desktop | For ironic baremetal, how should I be setting up the cleaning_network/provisioning_network? | 21:55 |
| *** Trevor_V has quit IRC | 21:55 | |
| dking_desktop | Is that just a random internal network and subnet? | 21:55 |
| sorin-mihai | admin0, regarding the certs for octavia i was told to "generate certs as per Octavia's documentation and put them in /etc/kolla/config/octavia, with those filenames: https://github.com/openstack/kolla-ansible/blob/3269b13ddde39128c246e1af18bace12e9926bab/ansible/roles/octavia/tasks/config.yml#L105-L107 " | 22:26 |
| admin0 | sorin-mihai, thanks | 22:28 |
| sorin-mihai | what are the steps to add something on top of a already deployed kolla-ansible? i guess it'l be something like pull, prechecks, deploy, check, post-deploy. is that correct? should do the same after a git pull in kolla-ansible cloned repo, or is there something i forget? | 22:33 |
| *** rlljorge has joined #openstack-kolla | 22:42 | |
| rlljorge | someone have the cloudkitty installed with stein release ? I cannot receive value and I cannot see errors in logs ... | 22:45 |
| *** rlljorge has quit IRC | 22:49 | |
| *** negronjl has quit IRC | 22:58 | |
| *** negronjl has joined #openstack-kolla | 23:02 | |
| *** ivve has quit IRC | 23:11 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!