| *** xinliang has joined #openstack-kolla | 00:26 | |
| *** k_mouza has joined #openstack-kolla | 00:57 | |
| *** k_mouza has quit IRC | 01:05 | |
| *** k_mouza has joined #openstack-kolla | 01:15 | |
| *** k_mouza has quit IRC | 01:19 | |
| *** k_mouza has joined #openstack-kolla | 01:56 | |
| *** k_mouza has quit IRC | 02:00 | |
| *** xinliang has quit IRC | 02:36 | |
| *** xinliang has joined #openstack-kolla | 03:07 | |
| *** Tony31 has joined #openstack-kolla | 03:26 | |
| *** Tony31 has quit IRC | 03:33 | |
| *** xinliang has quit IRC | 03:53 | |
| openstackgerrit | Merged openstack/kolla-ansible stable/train: Fix qemu loading of ceph.conf (permission error) https://review.opendev.org/705329 | 04:13 |
|---|---|---|
| *** mixor has joined #openstack-kolla | 04:31 | |
| *** negronjl has quit IRC | 04:42 | |
| *** negronjl has joined #openstack-kolla | 04:43 | |
| openstackgerrit | Dincer Celik proposed openstack/kolla master: WIP: Upgrade Elasticsearch and Kibana to 6.x https://review.opendev.org/647748 | 05:07 |
| *** hongbin has joined #openstack-kolla | 05:23 | |
| *** mixor has quit IRC | 05:32 | |
| *** evrardjp has quit IRC | 05:33 | |
| *** evrardjp has joined #openstack-kolla | 05:34 | |
| *** icarusfactor has joined #openstack-kolla | 05:36 | |
| *** factor has quit IRC | 05:38 | |
| *** skramaja has joined #openstack-kolla | 05:49 | |
| *** icarusfactor has quit IRC | 06:02 | |
| *** hongbin has quit IRC | 06:16 | |
| *** cah_link has joined #openstack-kolla | 06:18 | |
| *** ricolin has quit IRC | 06:24 | |
| osmanlicilegi | morning | 06:33 |
| cosmicsound | Good day osmanlicilegi | 06:35 |
| osmanlicilegi | cosmicsound: thx, you too :] | 06:40 |
| *** Luzi has joined #openstack-kolla | 06:48 | |
| *** Tony31 has joined #openstack-kolla | 06:58 | |
| Tony31 | hi guys - could anyone help me get my head around the "all in one" kayobe approach? | 06:59 |
| Tony31 | My problems are around the networking - what ever I try doesnt seem to work, or rather gets past the problem but then I find another | 06:59 |
| Tony31 | my latest problem seems to be that the aio hypervisor/host can't route to the seed VM's IP because the host has real network and bridge network in the same subnet | 07:00 |
| Tony31 | For example, my host VM has IP 192.168.1.100. So I have `seed_hv_ip=192.168.1.100`. If I then use 192.168.1.0/24 for the network allocations, this passes all the validations but I'm unable to access the seed VM once it's deployed | 07:04 |
| *** mixor has joined #openstack-kolla | 07:08 | |
| *** numans has joined #openstack-kolla | 07:20 | |
| *** cah_link has quit IRC | 07:30 | |
| *** mixor has quit IRC | 07:33 | |
| *** jbadiapa has joined #openstack-kolla | 07:39 | |
| *** pcaruana has joined #openstack-kolla | 07:43 | |
| *** gfidente has joined #openstack-kolla | 07:52 | |
| yoctozepto | Tony31: yeah, they seem to have to be different, could be prechecks are not tight enough there @ mgoddard, mnasiadka | 07:57 |
| Tony31 | hi Yoctozepto - i started again and did not change any config files this time... I have managed to progress beyond the last problem I had. I am not sure if this will be functional (yet) but I'll see :) (learning all the way...) | 07:58 |
| Tony31 | thanks for your reply :) | 07:58 |
| yoctozepto | Tony31: yeah, getting openstack done properly is a great sysops adventure :-) | 07:59 |
| Tony31 | :) I felt like Alice in Wonderland last time :) | 08:00 |
| yoctozepto | haha | 08:03 |
| mnasiadka | yoctozepto: it's trial by fire, not adventure :) | 08:03 |
| yoctozepto | mnasiadka: shhh, don't scare users :-) | 08:04 |
| *** rpittau|afk is now known as rpittau | 08:05 | |
| *** tonythomas has joined #openstack-kolla | 08:07 | |
| *** dswebb has quit IRC | 08:08 | |
| *** xinliang has joined #openstack-kolla | 08:09 | |
| *** luksky has joined #openstack-kolla | 08:09 | |
| *** xinliang has quit IRC | 08:34 | |
| *** xinliang has joined #openstack-kolla | 08:41 | |
| mgoddard | morning | 08:46 |
| mgoddard | I'm out today and tomorrow | 08:46 |
| yoctozepto | as a matter of fact, I am out tomorrow :-) (local conf) | 08:51 |
| *** dougsz has joined #openstack-kolla | 08:53 | |
| yoctozepto | mgoddard: if you have just one moment, I would be glad to get your blessing on mariadb https://review.opendev.org/704507 | 08:55 |
| patchbot | patch 704507 - kolla-ansible - Followup on MariaDB handling fixes - 3 patch sets | 08:55 |
| mgoddard | yoctozepto: let's see how train wifi goes | 08:56 |
| yoctozepto | mgoddard: crossing fingers | 08:56 |
| Tony31 | kayobe guys - I have a couple of machines I built for machine learning, with multiple nvidia GPUs etc. . I am reading up the stackHPC website about ironic and managing hardware. I dont have the right nvidia GPUs to be able to virtualise them (example, run the bare metal as a host and run VMs on top, with pci pass-through). But I wanted to ask - | 08:58 |
| Tony31 | could I manage this hardware from openstack and allow people to deploy and access the real hardware, then destroy when theyr'e done? | 08:58 |
| Tony31 | basically, use the real hardware and gpu's like a VM? | 08:59 |
| Tony31 | I guess I'd need to make a custom image with the drivers | 08:59 |
| mgoddard | hi Tony31. Are you sure you can't use PCI passthrough? No sharing, just pass through the entire GPU to a VM | 09:01 |
| openstackgerrit | Radosław Piliszek proposed openstack/kolla-ansible master: DNM: keepalived testing https://review.opendev.org/690920 | 09:01 |
| mgoddard | otherwise, you could use ironic to make the bare metal available to users. In that case it's treated like a black box - the user can do (almost) what they like with the machine | 09:02 |
| yoctozepto | Tony31: like mgoddard says, pci passtrough requires only vt-d in case of intel to work, no nvidia specific stuff nor sr-iov | 09:02 |
| yoctozepto | (mentioning because some assume otherwise) | 09:03 |
| yoctozepto | mgoddard: thx, I'll be backporting this to stable branches, at least train is nice as it affects upgrade jobs stability | 09:04 |
| Tony31 | I have 3 x nvidia titan XP in each system - I was told that they wont work for virtualisation o_0 - but it would be better using them in Openstack I think anyway | 09:04 |
| yoctozepto | Tony31: it depends on platform/mobo, not gpu | 09:05 |
| mgoddard | yoctozepto: ok. Let's try to land the original patch and follow up together | 09:05 |
| yoctozepto | any pci device can go passthrough | 09:05 |
| yoctozepto | mgoddard: yes, indeed | 09:05 |
| Tony31 | the system can be virtualised. I'll need to have a play with it - thanks for the info :) | 09:05 |
| Tony31 | it's dual xeon, tons of RAM, trying to find info on the motherboard | 09:07 |
| yoctozepto | Tony31: look up vt-d, dmar and iommu in google | 09:07 |
| yoctozepto | it generally needs to be enabled in bios setup and then linux kernel | 09:07 |
| Tony31 | thank you for that info | 09:09 |
| *** bersace has quit IRC | 09:09 | |
| openstackgerrit | Michal Nasiadka proposed openstack/kolla-ansible master: WIP: Ceph-Ansible upgrade jobs https://review.opendev.org/704984 | 09:12 |
| Tony31 | ok - i managed to get the seed deployed in the all-in-one.. the "network" is hidden behind the VM. Heading off now and will deploy the control plane tomorrow. Thanks for the help again today :) | 09:13 |
| Tony31 | have a good one | 09:13 |
| *** k_mouza has joined #openstack-kolla | 09:15 | |
| *** xinliang has quit IRC | 09:16 | |
| *** xinliang has joined #openstack-kolla | 09:17 | |
| *** Tony31 has quit IRC | 09:18 | |
| *** dougsz has quit IRC | 09:18 | |
| *** dougsz has joined #openstack-kolla | 09:21 | |
| *** bersace has joined #openstack-kolla | 09:30 | |
| openstackgerrit | Radosław Piliszek proposed openstack/kolla-ansible master: DNM: haproxy timeout - raise value for server https://review.opendev.org/690702 | 09:36 |
| *** xinliang has quit IRC | 09:41 | |
| *** Tengu has quit IRC | 09:46 | |
| *** hjensas has joined #openstack-kolla | 09:47 | |
| *** Tengu has joined #openstack-kolla | 09:48 | |
| *** Tengu has quit IRC | 09:54 | |
| *** Tengu has joined #openstack-kolla | 09:55 | |
| openstackgerrit | Dincer Celik proposed openstack/kolla master: WIP: Upgrade Elasticsearch and Kibana to 6.x https://review.opendev.org/647748 | 10:01 |
| *** Tengu has quit IRC | 10:03 | |
| *** skramaja has quit IRC | 10:03 | |
| *** Tengu has joined #openstack-kolla | 10:03 | |
| openstackgerrit | Merged openstack/kolla-ansible master: Followup on MariaDB handling fixes https://review.opendev.org/704507 | 10:06 |
| *** Tengu has quit IRC | 10:08 | |
| *** Tengu has joined #openstack-kolla | 10:13 | |
| *** jopdorp has joined #openstack-kolla | 10:14 | |
| *** Tengu_ has joined #openstack-kolla | 10:14 | |
| *** dougsz has quit IRC | 10:19 | |
| *** dougsz has joined #openstack-kolla | 10:27 | |
| *** jopdorp has quit IRC | 10:41 | |
| openstackgerrit | Michal Nasiadka proposed openstack/kolla-ansible master: WIP: Ceph-Ansible upgrade jobs https://review.opendev.org/704984 | 10:47 |
| openstackgerrit | Michal Nasiadka proposed openstack/kolla-ansible master: Ceph-Ansible upgrade jobs https://review.opendev.org/704984 | 10:47 |
| *** jbadiapa has quit IRC | 10:49 | |
| openstackgerrit | Radosław Piliszek proposed openstack/kolla-ansible stable/train: Fix multiple issues with MariaDB handling https://review.opendev.org/705414 | 10:55 |
| *** reph has joined #openstack-kolla | 11:08 | |
| reph | hi, deploy using kolla 9.0.1 yields "pull access denied for kolla/ubuntu-binary-cloudkitty-api, repository does not exist or may require \\'docker login\\" | 11:09 |
| reph | kolla/ubuntu-binary-cloudkitty-api is in fact not available on dockerhub | 11:11 |
| *** reph has quit IRC | 11:21 | |
| *** reph has joined #openstack-kolla | 11:23 | |
| mnasiadka | reph: cloudkitty ubuntu binary is marked as unbuildable | 11:27 |
| mnasiadka | try ubuntu-source | 11:27 |
| reph | mnasiadka thanks, didn't stumble upon that piece of information :/ | 11:39 |
| reph | cloudkitty will also fail when using source, unfortunately, at least that's what it did on friday | 11:41 |
| reph | it said "cloudkitty "AttributeError: 'NoneType' object has no attribute 'auth_url'"" | 11:42 |
| *** rpittau is now known as rpittau|bbl | 11:42 | |
| reph | in cloudkitty/tenant_fetcher/keystone.py somewhere around line 50, wheree it tries to read from self.auth.auth_url | 11:44 |
| mnasiadka | reph: then it might be something wrong with configuration done by k-a | 11:50 |
| mnasiadka | reph: never used that, so can't tell | 11:50 |
| *** luksky has quit IRC | 11:52 | |
| yoctozepto | reph: https://docs.openstack.org/kolla/train/support_matrix.html | 11:55 |
| yoctozepto | the other issue might be a bug in k-a | 11:55 |
| yoctozepto | please report | 11:56 |
| yoctozepto | link: https://bugs.launchpad.net/kolla-ansible | 11:56 |
| *** jbadiapa has joined #openstack-kolla | 12:03 | |
| *** luksky has joined #openstack-kolla | 12:17 | |
| *** cah_link has joined #openstack-kolla | 12:19 | |
| *** kplant has joined #openstack-kolla | 12:21 | |
| openstackgerrit | Michal Nasiadka proposed openstack/kolla-ansible master: Fix keystone fernet bootstrap https://review.opendev.org/698710 | 12:41 |
| mnasiadka | yoctozepto: if you have a moment - I would be grateful for looking into this: https://review.opendev.org/#/c/704984 | 12:51 |
| patchbot | patch 704984 - kolla-ansible - Ceph-Ansible upgrade jobs - 12 patch sets | 12:51 |
| *** hjensas has quit IRC | 13:16 | |
| *** hjensas has joined #openstack-kolla | 13:27 | |
| *** theZorro266 has joined #openstack-kolla | 13:30 | |
| *** rpittau|bbl is now known as rpittau | 13:34 | |
| theZorro266 | Hello, couldn't find a solution with some searches: I have been following https://docs.openstack.org/kolla-ansible/train/reference/storage/external-ceph-guide.html after using ceph-ansible as I understand ceph deployment by kolla-ansible is deprecated. I am running into the error "MissingRequiredSource: /var/lib/kolla/config_files/ceph.* file is not found" although the files exist in | 13:40 |
| theZorro266 | "/etc/kolla/config/glance/". Is this a user error? The guide tells me nothing about permissions, so I assumed it should be fine with the same permissions as "/etc/kolla/glance-api/". | 13:40 |
| kplant | what services are you using ceph for? | 13:43 |
| kplant | and did you include both ceph.conf and the ceph keyrings? | 13:43 |
| theZorro266 | I would like to use ceph for glance, cinder, nova, gnocchi and manila. I am generating keys for all of these services and the keys are in the appropiate directories as written in the linked guide | 13:44 |
| *** rlljorge has joined #openstack-kolla | 13:44 | |
| theZorro266 | And yeah, both ceph.client.glance.keyring and ceph.conf are in /etc/kolla/config/glance/ | 13:44 |
| theZorro266 | but not in /etc/kolla/glance-api/ | 13:45 |
| yoctozepto | mnasiadka: I will | 13:45 |
| theZorro266 | and I understood that they should be merged into /etc/kolla/glance-api/ | 13:45 |
| kplant | this is what mine looks like: https://pastebin.com/hSDusB7v | 13:46 |
| theZorro266 | this is what it looks like over here: https://pastebin.com/iyghGxnM | 13:49 |
| theZorro266 | Which version of kolla-ansible are you using? I see things have changed regarding glance external ceph setup between 9.0.1 and master | 13:51 |
| kplant | stable/stein and stable/train | 13:51 |
| kplant | so | 13:52 |
| kplant | your permissions are 0660 | 13:52 |
| kplant | what user are you running ansible as? | 13:52 |
| kplant | if it's not root or in the root group it won't read them | 13:53 |
| theZorro266 | I see. I run ansible as a user, of course | 13:53 |
| theZorro266 | That is what I was asking above. I assumed /etc/kolla/glance-api/ should have the same permissions as in /etc/kolla/config/glance | 13:54 |
| kplant | if you o+r on ceph.* you should be good | 13:55 |
| theZorro266 | Yeah, not forgetting the directories, thank you! :) I'll try and see if that works | 13:57 |
| *** reph has quit IRC | 14:07 | |
| theZorro266 | It seems to work now, thanks again kplant :) This seems related to the "with_fileglob" in roles/glance/tasks/external_ceph.yml, because the list of files is obviously gathered as the user running ansible. This would mean only the folders in /etc/kolla/config should be world-readable/listable and not the keyrings/configs. Shouldn't this be in the docs? I could make a PR, but I'd need a hint | 14:14 |
| theZorro266 | where to insert it to. | 14:14 |
| theZorro266 | strike that. It copies the files, so it needs read access to the files as well | 14:15 |
| kplant | ^ | 14:15 |
| theZorro266 | but still, it would be better to have the files user-owned by the user running Ansible in that case | 14:16 |
| kplant | i think that's in the quickstart | 14:16 |
| theZorro266 | thats only the /etc/kolla directory | 14:17 |
| theZorro266 | and all files within are being created and owned by root:root with zero permissions for world | 14:18 |
| theZorro266 | that's where it misled me | 14:18 |
| *** Luzi has quit IRC | 14:22 | |
| theZorro266 | the external ceph guide is the first time I came into contact with the /etc/kolla/config directory, so any other place with that info should at least be linked from there in my understanding | 14:23 |
| *** TrevorV has joined #openstack-kolla | 14:33 | |
| *** dougsz has quit IRC | 14:33 | |
| mnasiadka | theZorro266: it's in the docs - https://docs.openstack.org/kolla-ansible/latest/admin/advanced-configuration.html#openstack-service-configuration-in-kolla - just read it :) | 14:42 |
| rlljorge | HI folks, some ideia ? | 14:50 |
| rlljorge | fatal: [network02]: FAILED! => { | 14:50 |
| rlljorge | exact syntax problem.\n\nThe offending line appears to be:\n\n\n- name: Set fluentd facts\n ^ here\n" | 14:50 |
| theZorro266 | mnasiadka: thanks :) What do you mean with "in the docs" though? When I search for "/etc/kolla/config" on my search engine (not the documentation's search), only then I will find that page | 14:50 |
| theZorro266 | My point is that the user should be made aware of ownership of directories and files in /etc/kolla/config as it is different from the rest of the directories and files in /etc/kolla | 14:52 |
| *** dougsz has joined #openstack-kolla | 14:52 | |
| mnasiadka | theZorro266: how different? | 14:57 |
| mnasiadka | rlljorge: ansible version? kolla_ansible package version? | 14:58 |
| *** reph has joined #openstack-kolla | 14:59 | |
| *** jbadiapa has quit IRC | 15:02 | |
| theZorro266 | mnasiadka: well, kolla-ansible creates /etc/kolla/glance-api with root:root and 0770, now in /etc/kolla/config/glance my user running kolla-ansible must have at least read access | 15:02 |
| theZorro266 | I am not saying the later does not make sense, it just didn't strike my mind soon enough | 15:04 |
| rlljorge | mnasiadka: ansible 2.9.2 stable/stein tag 8.1.0 , I deploy train release with success because a bug in cinder/nova I will train stein release | 15:04 |
| rlljorge | deployed* | 15:05 |
| *** vmixor has joined #openstack-kolla | 15:08 | |
| admin0 | ok guys .. i want to test/document a multinode kolla deployment .. i am using osa(openstack ansible) all the way, but now want to give kolla a try .. i need a bit of help with the mappings | 15:17 |
| admin0 | i want to move database ( mysql) and rabbitmq to their own hosts ( not the controllers) .. how do I do this ? | 15:17 |
| kplant | pretty sure there's already host groups in the example inventory files for that | 15:19 |
| admin0 | test rig: | 15:19 |
| admin0 | 3x controllers -- c1 c2 c3 ; 3x data servers - d1 d2 d3 ( for mysql and rabbitmq) ; 3x nova computes - n1 n2 n3 (kvm based) ; 3x lxd computes - x1 x2 x3 ( for use with lxd) ; 3x ceph osd nodes - s1 s2 s3 ;3x swift nodes - os1 os2 os3 ; 2x swift proxy node - op1 and op2 | 15:19 |
| kplant | yep, both mariadb and rabbitmq already have groups defined | 15:19 |
| admin0 | awesome | 15:19 |
| admin0 | https://docs.openstack.org/kolla-ansible/latest/user/multinode.html -- does not say much | 15:20 |
| kplant | "For more advanced roles, the operator can edit which services will be associated in with each group. Keep in mind that some services have to be grouped together and changing these around can break your deployment" | 15:21 |
| kplant | i think that's the closest to what you're looking for | 15:21 |
| kplant | but if you take a look at multinode they're already there | 15:21 |
| *** mloza has joined #openstack-kolla | 15:25 | |
| mloza | Is there plan to backport this to stein https://review.opendev.org/#/c/686369/ ? | 15:26 |
| patchbot | patch 686369 - kolla-ansible - Allow passing arguments to RabbitMQ server (MERGED) - 5 patch sets | 15:26 |
| kplant | ^ that is a pretty nice one to fix the cpu usage | 15:35 |
| kplant | mnasiadka yoctozepto mgoddard what do you guys think about that? | 15:48 |
| kplant | https://bugs.launchpad.net/kolla-ansible/+bug/1846467 impacts stein and that patch gives ops the ability to correct it | 15:48 |
| openstack | Launchpad bug 1846467 in kolla-ansible train "RabbitMQ container - high CPU load on multicore systems" [Medium,Fix released] | 15:48 |
| noxoid | yea my rabbitmq container is using 120-190% cpu mostly idle on a stein deployment | 15:58 |
| kplant | mine does the same, on stein | 15:59 |
| admin0 | kplant, just to reconfirm .. if i want separate api and storage but on the same interface, i need to say for example create 2 vlans, create 2 bridges with br_ (underscore) name and then pass it like api_interface: br_api right ? | 16:01 |
| kplant | they don't have to be bridges | 16:03 |
| kplant | but yes | 16:03 |
| *** jbadiapa has joined #openstack-kolla | 16:03 | |
| kplant | can be a raw interface or bond also | 16:03 |
| admin0 | ok .. so ovs is used all the way and not linuxbridge | 16:04 |
| kplant | em0.11,em0.12 bond0.11,bond0.12 etc. | 16:04 |
| kplant | depends, you can configure k-a to use linubridge and not ovs | 16:04 |
| kplant | heading to lunch, i'll be around after | 16:05 |
| admin0 | where do I map some hosts to be using kvm and some using lxd | 16:05 |
| *** luksky has quit IRC | 16:32 | |
| *** rpittau is now known as rpittau|afk | 16:33 | |
| oncall-pokemon | With kolla, in the inventory file you can setup multiple servers under [mariadb]. What does it mean if you add two or more? will replication start or something? | 16:52 |
| kplant | admin0: you can override variables per host just as you do with ansible | 16:59 |
| kplant | in the inventory or host_vars | 16:59 |
| rlljorge | oncall-pokemon: I think is need 3 os more because is a galera cluster | 17:00 |
| openstackgerrit | Doug Szumski proposed openstack/kolla-ansible master: Extract useful information from HAProxy logs https://review.opendev.org/705484 | 17:02 |
| dougsz | ^ For people who enjoy reading regex's | 17:02 |
| oncall-pokemon | rlljorge: not sure what that means. mind expanding a bit. I'm a bit of a noobie | 17:06 |
| openstackgerrit | Doug Szumski proposed openstack/kolla-ansible master: Extract useful information from HAProxy logs https://review.opendev.org/705484 | 17:09 |
| yankcrime | oncall-pokemon: yeah, if you specify more than one node it'll create a cluster for you | 17:12 |
| yankcrime | data is replicated across all members of the cluster | 17:12 |
| yankcrime | (it's mariadb galera if you want to read more) | 17:12 |
| oncall-pokemon | awesome thanks | 17:18 |
| oncall-pokemon | Can I add more to the list without destruction to the one on the controller now? | 17:18 |
| oncall-pokemon | we already have some data | 17:18 |
| rlljorge | oncall-pokemon: Yes | 17:19 |
| rlljorge | If the answer is 1 node, then you need 3 nodes: when 1 node fails, the two remaining nodes have quorum. | 17:20 |
| kplant | i think the answer is growing from a 1 node mariadb to a 3 node galera | 17:20 |
| kplant | question** | 17:20 |
| kplant | i don't know if that's supported | 17:20 |
| admin0 | kplant, is something like graylog supported ? to make a central logging server ? | 17:27 |
| kplant | kibana is | 17:28 |
| kplant | there's fluentd, collectd, .... i think there's more | 17:28 |
| admin0 | kolla_internal_tls_cert . is this a pem with the crt, ca cert and keyfile all inside those block ? | 17:28 |
| admin0 | i have say domain.crt and domain.key .. what goes where | 17:29 |
| kplant | i just feed it a pem with the entire chain and key | 17:29 |
| admin0 | ok | 17:33 |
| *** evrardjp has quit IRC | 17:33 | |
| *** evrardjp has joined #openstack-kolla | 17:34 | |
| *** k_mouza has quit IRC | 17:39 | |
| *** jbadiapa has quit IRC | 17:40 | |
| dougsz | admin0: there is Fluent + Kibana + ES (enable_central_logging) or Monasca for centralised logging | 17:52 |
| *** gfidente is now known as gfidente|afk | 17:59 | |
| *** dougsz has quit IRC | 18:03 | |
| *** k_mouza has joined #openstack-kolla | 18:11 | |
| *** k_mouza has quit IRC | 18:15 | |
| *** k_mouza has joined #openstack-kolla | 18:15 | |
| *** tonythomas has quit IRC | 18:19 | |
| *** weshay is now known as weshay|ruck | 18:30 | |
| *** reph has quit IRC | 18:34 | |
| *** luksky has joined #openstack-kolla | 18:37 | |
| *** cah_link has quit IRC | 18:40 | |
| rlljorge | Hello @kplant in the last week we talk about the bug https://bugs.launchpad.net/kolla-ansible/+bug/1860739 Do you can share your kolla_base_distro with me ? | 18:41 |
| openstack | Launchpad bug 1860739 in OpenStack Compute (nova) "Cinder backup failed for restore volume snapshot (for instance snapshot)" [Undecided,Confirmed] | 18:41 |
| yoctozepto | oncall-pokemon, yankcrime, rlljorge, kplant: just to clarify - scaling mariadb galera cluster up *is* supported | 18:46 |
| yoctozepto | scaling down is kinda too, it just involves a manual step of stopping extra mariadb when cluster is healthy with all the other member | 18:47 |
| yoctozepto | s alive | 18:47 |
| yoctozepto | (premature enter) | 18:47 |
| rlljorge | yoctozepto nice ! thanks | 18:48 |
| *** rlljorge has quit IRC | 19:36 | |
| yoctozepto | mnasiadka: is that keystone fix ready to merge or you really expected a failure? | 19:50 |
| mnasiadka | yoctozepto: well, it seems to work - was really just sarcastic :) | 19:53 |
| mnasiadka | yoctozepto: I tested it also on scale out | 19:54 |
| yoctozepto | mnasiadka: very well! | 19:54 |
| mnasiadka | Which brings an observation - wouldn’t we want to have a periodic scale out job? | 19:54 |
| mnasiadka | It is probably something our users sometimes do... or not | 19:55 |
| *** gfidente|afk is now known as gfidente | 19:56 | |
| *** jonaspaulo has quit IRC | 20:02 | |
| *** irclogbot_1 has quit IRC | 20:05 | |
| *** irclogbot_1 has joined #openstack-kolla | 20:06 | |
| yoctozepto | mnasiadka: makes sense, add it to scenarios | 20:17 |
| yoctozepto | 1->3 scale obviously :-) | 20:17 |
| *** jonaspaulo has joined #openstack-kolla | 20:17 | |
| mnasiadka | yoctozepto: yeah, 3 to 5 is rather an overkill :) | 20:20 |
| yoctozepto | mnasiadka: well, it tests if an existing multi does not get killed | 20:24 |
| yoctozepto | we *might* want to have it running | 20:24 |
| yoctozepto | but daily is an overkill | 20:24 |
| yoctozepto | especially if you overdo the enabled services | 20:24 |
| mnasiadka | yoctozepto: yeah, maybe weekly makes sense - 1->3->5 - it will probably fail multiple times :) I can start with 1->3 and lets see how that works | 20:27 |
| yoctozepto | mnasiadka: it must not fail, we fixed a lot | 20:28 |
| yoctozepto | mnasiadka: only haproxy/keepalived remains | 20:28 |
| *** gfidente has quit IRC | 20:28 | |
| yoctozepto | mnasiadka: I restored my machinery to play with it | 20:28 |
| yoctozepto | mnasiadka: hoping to get it fixed too | 20:28 |
| yoctozepto | mnasiadka: then we are golden like them kings 8-) | 20:29 |
| mnasiadka | yoctozepto: it would be good to understand why it fails, and how to make it better | 20:29 |
| yoctozepto | mnasiadka: that's the goal | 20:29 |
| mnasiadka | yoctozepto: if we are in keepalived dept - would be good to document a procedure of maintenance failover :) | 20:29 |
| * yoctozepto not listening | 20:30 | |
| mnasiadka | Haha | 20:30 |
| *** k_mouza has quit IRC | 20:34 | |
| *** kplant has quit IRC | 20:58 | |
| *** bjolo has quit IRC | 21:47 | |
| *** vmixor has quit IRC | 22:11 | |
| *** dciabrin has quit IRC | 22:38 | |
| *** dciabrin has joined #openstack-kolla | 22:43 | |
| *** TrevorV has quit IRC | 22:53 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!