| *** ivve has quit IRC | 00:32 | |
| *** markvoelker has joined #openstack-ansible | 00:39 | |
| *** gyee has quit IRC | 00:43 | |
| *** spsurya has joined #openstack-ansible | 01:05 | |
| *** altlogbot_2 has quit IRC | 02:44 | |
| *** altlogbot_3 has joined #openstack-ansible | 02:44 | |
| *** rholloway has quit IRC | 03:10 | |
| *** markvoelker has quit IRC | 03:18 | |
| *** redrobot has quit IRC | 03:22 | |
| *** shyamb has joined #openstack-ansible | 04:23 | |
| *** shyamb has quit IRC | 04:37 | |
| *** raukadah is now known as chkumar|ruck | 04:38 | |
| *** markvoelker has joined #openstack-ansible | 04:48 | |
| *** markvoelker has quit IRC | 04:52 | |
| *** udesale has joined #openstack-ansible | 05:33 | |
| *** dave-mccowan has joined #openstack-ansible | 05:33 | |
| *** shyamb has joined #openstack-ansible | 05:33 | |
| *** shyamb has quit IRC | 05:59 | |
| *** shyamb has joined #openstack-ansible | 06:04 | |
| *** shyamb has quit IRC | 06:17 | |
| *** mcarden has quit IRC | 06:26 | |
| *** shyamb has joined #openstack-ansible | 06:33 | |
| *** shyam89 has joined #openstack-ansible | 06:38 | |
| *** Adri2000 has quit IRC | 06:39 | |
| *** shyamb has quit IRC | 06:42 | |
| *** rpittau|afk is now known as rpittau | 06:52 | |
| *** shyam89 has quit IRC | 06:56 | |
| *** kopecmartin|off is now known as kopecmartin | 07:04 | |
| *** shyamb has joined #openstack-ansible | 07:07 | |
| *** tosky has joined #openstack-ansible | 07:23 | |
| *** ppetit has joined #openstack-ansible | 07:35 | |
| *** shyamb has quit IRC | 07:45 | |
| zbr | ca anyone help with https://review.opendev.org/#/c/673859/ ? | 07:47 |
|---|---|---|
| *** ivve has joined #openstack-ansible | 08:15 | |
| *** gchamoul has joined #openstack-ansible | 08:27 | |
| *** vincelevi has joined #openstack-ansible | 08:33 | |
| vincelevi | hello, I'm still struggling with my first OSA deployment, getting stuck at "TASK [lxc_container_create : Create container (cow)]" even after deleting all containers and relaunching all back from setup-hosts.yml. The error message is "Error: container centos-7-amd64 is not defined" | 08:33 |
| *** miloa has joined #openstack-ansible | 09:02 | |
| miloa | Morning | 09:02 |
| jrosser | vincelevi: i see that you are using a COW filesystem and centos | 09:04 |
| vincelevi | yep, at least I'm trying to... | 09:04 |
| *** shyamb has joined #openstack-ansible | 09:04 | |
| jrosser | it might be worth review how our CI tests are setup for lxc_container_create, like here as an example https://review.opendev.org/#/c/672616/ | 09:04 |
| vincelevi | is the 3.10 kernel too old for that ? | 09:04 |
| jrosser | probably best to stick with an OS / backing store combination that show to be working in test | 09:05 |
| jrosser | which at first sight looks to be either dir or machinectl for centos | 09:06 |
| vincelevi | Yep, I was about to say that I don't see lvm there... OK I'll disable lvm for lxc backing store then, and retry, thanks a lot for the hint | 09:07 |
| vincelevi | BTW is there a document that explain how to cleanup your hosts between deployment retries ? I was just lxc-stopping & lxc-destroying containers at first, then saw someone speaking about "lxc-container-destroy.yml" playbook on an irc log, so started to use that... what's the official way ? | 09:09 |
| jrosser | that is the official way to delete containers and their associated data, yes | 09:10 |
| vincelevi | OK, is this mentionned somewhere in the doc (that I missed) or not ? I think this should be in (probably at the beginning) | 09:11 |
| *** markvoelker has joined #openstack-ansible | 09:18 | |
| *** markvoelker has quit IRC | 09:23 | |
| jrosser | Ideally if you are in a lab setup you have some way of reprovisioning the hosts or snapshot them back to a clean state | 09:32 |
| jrosser | But that’s up to you how you want to do that | 09:32 |
| vincelevi | Yes, I will redo from scratch (OS reinstall) when I have found the config that is properly working, but in the mean time I don't want to reinstall the OS, and I was not aware of hat to do to retry the deployment from the start, so I tried to guess, which the doc should have helped me to avoid... I'll try to get some time to PR a NOTE in the OSA doc | 09:35 |
| vincelevi | (after holidays) | 09:35 |
| vincelevi | thank a lot for your help though ! | 09:36 |
| admin0 | good morning ..checkig if anyone has lxd+ceph on osa | 09:45 |
| *** noonedeadpunk has joined #openstack-ansible | 09:46 | |
| noonedeadpunk | mornings | 09:46 |
| admin0 | mornings | 09:46 |
| noonedeadpunk | jrosser: can you check https://review.opendev.org/#/c/675429/ to get gates finally unblocked? | 09:47 |
| *** shyamb has quit IRC | 10:01 | |
| *** aedc has quit IRC | 10:06 | |
| *** aedc has joined #openstack-ansible | 10:06 | |
| jrosser | noonedeadpunk: won’t get to look properly till this afternoon but I’m still worried about it https://github.com/ansible/ansible/issues/44261 | 10:13 |
| noonedeadpunk | to be honest, in that exact case it doesn't really matter, since the only place this variable is used https://opendev.org/openstack/openstack-ansible-os_nova/src/branch/master/defaults/main.yml#L397 | 10:15 |
| *** shyamb has joined #openstack-ansible | 10:17 | |
| noonedeadpunk | and in terms of that it doesn't matter if it's None or empty string. But I fully understand your concern and I still like approach with simplifying condition here https://opendev.org/openstack/openstack-ansible/src/branch/master/playbooks/common-playbooks/nova.yml#L39 just to gather fact against first nova node more | 10:18 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_manila master: Add integrated coverage https://review.opendev.org/660333 | 10:44 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_manila master: Add integrated coverage https://review.opendev.org/660333 | 10:45 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_nova master: Do not set upgrade_levels for aio builds https://review.opendev.org/675587 | 11:08 |
| *** kplant has joined #openstack-ansible | 11:11 | |
| *** mkuf has joined #openstack-ansible | 11:22 | |
| *** mkuf_ has joined #openstack-ansible | 11:28 | |
| *** mkuf has quit IRC | 11:32 | |
| *** ansmith has quit IRC | 11:35 | |
| *** rohit02 has joined #openstack-ansible | 11:40 | |
| *** mkuf_ has quit IRC | 11:47 | |
| zbr | does anyone have any insights regarding purpose of maturity_info in meta/openstack-ansible.yml ? | 11:48 |
| zbr | somehow that file confuses linter due to its location inside the role. | 11:49 |
| noonedeadpunk | zbr: I'd say it might be related to https://docs.openstack.org/openstack-ansible/latest/contributor/testing.html#role-development-maturity | 11:50 |
| noonedeadpunk | But don't now if it's actually used.... | 11:50 |
| zbr | i bet is used to build that table. nobody would want to maintain such table manually. | 11:50 |
| zbr | sadly, ansible-lint complains about format | 11:51 |
| *** ppetit has quit IRC | 11:52 | |
| *** markvoelker has joined #openstack-ansible | 11:59 | |
| *** rohit02 has quit IRC | 12:04 | |
| *** shyamb has quit IRC | 12:07 | |
| *** ansmith has joined #openstack-ansible | 12:13 | |
| *** vincelevi has quit IRC | 12:14 | |
| *** jamesdenton has quit IRC | 12:17 | |
| openstackgerrit | Sorin Sbarnea proposed openstack/openstack-ansible-os_tempest master: POC: Implement role functional testing via molecule https://review.opendev.org/675615 | 12:40 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_manila master: Add integrated coverage https://review.opendev.org/660333 | 12:41 |
| openstackgerrit | Sorin Sbarnea proposed openstack/openstack-ansible-os_tempest master: POC: Implement role functional testing via molecule https://review.opendev.org/675615 | 12:43 |
| *** udesale has quit IRC | 12:56 | |
| *** udesale has joined #openstack-ansible | 12:57 | |
| openstackgerrit | Sorin Sbarnea proposed openstack/openstack-ansible-os_tempest master: Fixed bugs identified by newer ansible-lint https://review.opendev.org/675617 | 12:57 |
| *** Guest1857 has joined #openstack-ansible | 12:59 | |
| logan- | jrosser: http://paste.openstack.org/raw/755691/ | 13:03 |
| noonedeadpunk | logan-: I guess he ment, that omit supposed to pass None type (or undefined variable), while it defined and is str() | 13:05 |
| logan- | gotcha | 13:05 |
| noonedeadpunk | But yes, in case of condition check both str() and None will result in False... | 13:06 |
| noonedeadpunk | So it doesn't matter here I guess... | 13:06 |
| *** Adri2000 has joined #openstack-ansible | 13:07 | |
| openstackgerrit | Logan V proposed openstack/openstack-ansible master: Omit nova_cinder_rbd_inuse when it is not set https://review.opendev.org/675429 | 13:08 |
| logan- | that should address the concern ^ | 13:08 |
| noonedeadpunk | logan-: change commit msg please then:) | 13:08 |
| jrosser | havent we had it evaluate to some bit random omit_<great-big-string> before | 13:08 |
| *** rholloway has joined #openstack-ansible | 13:09 | |
| openstackgerrit | Logan V proposed openstack/openstack-ansible master: Default nova_cinder_rbd_inuse when it is not set https://review.opendev.org/675429 | 13:09 |
| noonedeadpunk | yeah, excactly what I've offered 2 days ago :P | 13:10 |
| mgariepy | anyone here knows if it's doable to migrate ephemral storage (local ssds/hdds ) to ephemral ceph ? | 13:10 |
| noonedeadpunk | ¯\_(ツ)_/¯ | 13:11 |
| mgariepy | that's what i think also ;) | 13:12 |
| logan- | jrosser: tested, and yes it does evaluate to the omit placeholder. in our use of the var where we are passing it through the bool filter it evaluates to false. so you're spot on with the concern about that. | 13:16 |
| *** jamesdenton has joined #openstack-ansible | 13:21 | |
| *** jamesdenton has quit IRC | 13:22 | |
| *** BjoernT has joined #openstack-ansible | 13:28 | |
| openstackgerrit | Sorin Sbarnea proposed openstack/openstack-ansible-os_tempest master: Improve loading of operating system specific vars https://review.opendev.org/675624 | 13:32 |
| *** ivve has quit IRC | 13:33 | |
| *** jamesdenton has joined #openstack-ansible | 13:36 | |
| *** lbragstad has joined #openstack-ansible | 13:46 | |
| *** spsurya has quit IRC | 13:54 | |
| *** jeh has left #openstack-ansible | 13:56 | |
| *** chkumar|ruck is now known as raukadah | 13:57 | |
| *** lbragstad has quit IRC | 13:58 | |
| *** lbragstad has joined #openstack-ansible | 14:04 | |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-tests master: sync: add service_setup.yml https://review.opendev.org/675632 | 14:10 |
| *** lbragstad has quit IRC | 14:14 | |
| *** ivve has joined #openstack-ansible | 14:39 | |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible master: Add image for manila test https://review.opendev.org/674653 | 14:47 |
| grabes | Anyone here do any MS product licensing for your VMs specifically for hosting? | 14:48 |
| noonedeadpunk | grabes: me previous employee did, but they used to have DC edition, so was paying per cores on compute nodes instead VMs | 14:49 |
| noonedeadpunk | but I'm not a big MS licensing expert though | 14:49 |
| grabes | Yeah, I think its SPLA licensing, but its hard to get straight forward answers from companies. They keep thinking I want virtualize VMs inside Windows. | 14:52 |
| kplant | we do, the licensing model just plain sucks | 14:58 |
| kplant | you pay for the hardware on the hypervisor, even if that's 1 vm with 1 core | 14:59 |
| *** ansmith_ has joined #openstack-ansible | 14:59 | |
| noonedeadpunk | kplant: it depends on the amount of windows VMs on the node. | 14:59 |
| noonedeadpunk | And since you place only windows vms on the licensed host, and have enough vms to cover node costs it make sense | 15:00 |
| *** ansmith has quit IRC | 15:01 | |
| kplant | you sure about that? we have to pay according to the physical hardware | 15:01 |
| kplant | whether there's 1 core of windows or 84 cores of windows | 15:02 |
| kplant | it's the same price | 15:02 |
| grabes | kplant: Thats strange, so if the VM has 1 core, and the hypervisor has 40 cores, you are paying for 40 cores? I could see if I was using MS to virtualize, but if your not this doesnt make sense to use OS | 15:02 |
| noonedeadpunk | kplant: so I was saying about the same... Just misunderstanding:) | 15:03 |
| kplant | noonedeadpunk: oh, my bad :] | 15:03 |
| noonedeadpunk | grabes: you're paying for cpu physical cores * sockets | 15:03 |
| kplant | ^ | 15:03 |
| noonedeadpunk | So if you have on node 2 CPU, 24 cores each to you'll have to pay for 48 cores. And you may place whatever amount of vms on it | 15:04 |
| grabes | noonedeadpunk: How would they know that given the VM doesn’t even see the true HV cpus? | 15:04 |
| noonedeadpunk | they kinda trust you:)))) | 15:04 |
| grabes | noonedeadpunk: I got what you are saying now, so basically you need to group your MS vms together and have them only use certain HV group | 15:05 |
| Carcer | you'll probably find that you agree they can come and audit you somewhere in the terms and services | 15:05 |
| noonedeadpunk | ++ | 15:05 |
| noonedeadpunk | Yeah, so they trust but verify:) | 15:05 |
| Carcer | my shop has some windows server virtualisation and it's not my specific domain but the per-core licensing stuff sounds very familiar from overheard/read conversations | 15:06 |
| kplant | RH does the same iirc | 15:09 |
| kplant | it's per socket pair | 15:09 |
| kplant | and you just cram as many vms on there as you want | 15:09 |
| *** shyamb has joined #openstack-ansible | 15:15 | |
| *** shyamb has quit IRC | 15:20 | |
| *** shyam89 has joined #openstack-ansible | 15:20 | |
| *** shyam89 has quit IRC | 15:20 | |
| *** spatel has joined #openstack-ansible | 15:23 | |
| spatel | How do you guys quickly find instance-1 running on compute-X ? we can do using GUI but any quick command line way? | 15:24 |
| noonedeadpunk | you can both from horizon or cli | 15:24 |
| noonedeadpunk | you just should have andmin proviliges on project | 15:25 |
| spatel | what command do you use for cli ? | 15:25 |
| noonedeadpunk | openstack server show uuid | 15:25 |
| evrardjp | o/ | 15:25 |
| spatel | I want to find any host starting with *www* running on XXX compute nodes | 15:25 |
| noonedeadpunk | openstack server list --host XXX --name www | 15:27 |
| noonedeadpunk | you may do openstack server list --help for more options | 15:27 |
| spatel | we have 800 instance running on cloud and any openstack command api calls take hell of time :( if i want to find using some short of regex hostname | 15:28 |
| noonedeadpunk | mysql request?:) | 15:29 |
| spatel | wish openstack has some short of command which directly fetch data from mysql DB, instead of api calls | 15:29 |
| noonedeadpunk | it's pretty easy with python to write model to db and issue a request to it, especially since you need just one table | 15:31 |
| noonedeadpunk | I guess you can do this with bash and without model as well | 15:32 |
| noonedeadpunk | And of course you shouldn't do any changes that way | 15:32 |
| spatel | Damnn!! i found it.. http://paste.openstack.org/show/755702/ | 15:33 |
| spatel | mysql -h 172.28.15.79 -u nova --password=xxxxxxxxxx nova -e "select hostname,vm_state,node,availability_zone,vcpus,memory_mb,root_gb from instances WHERE (vm_state='active' OR vm_state='stopped');" | 15:33 |
| spatel | i will put that in script :) | 15:34 |
| *** AR2018 has joined #openstack-ansible | 15:45 | |
| *** gyee has joined #openstack-ansible | 15:50 | |
| miloa | spatel: openstack server list --all-project --status ACTIVE --name www --host compute-xxxxx | 16:00 |
| *** AR2018 has quit IRC | 16:00 | |
| spatel | miloa: beautiful :) thanks | 16:02 |
| miloa | :) | 16:02 |
| *** udesale has quit IRC | 16:02 | |
| spatel | i still think mysql db would be quick when we have 800 instance to look and find patterns :) | 16:02 |
| miloa | yes, perhaps but with mysql you are using mysql nova account, it could be a problem depending on your security needs, with openstack cli you are using your user credentials. | 16:04 |
| *** Guest1857 is now known as redrobot | 16:17 | |
| *** aedc has quit IRC | 16:20 | |
| openstackgerrit | Sorin Sbarnea proposed openstack/openstack-ansible-os_tempest master: Improve loading of operating system specific vars https://review.opendev.org/675624 | 16:24 |
| *** spatel has quit IRC | 16:25 | |
| rholloway | regarding all that MS licensing stuff does anyone actually have a ballpark around how much that runs for something like 48 cores? We've historically been all linux but do have more windows needs coming up, thought I'd ask here in case someone knows although I'm sure it may vary a lot based on agreements/bulk discounts/etc | 16:27 |
| rholloway | I'm a tech guy and just dread going through the dance to get rough estimate of what to expect | 16:28 |
| *** kopecmartin is now known as kopecmartin|off | 16:29 | |
| rholloway | also I'm assuming that's just limited to a specific license like win server 2016 or something - not a mix? ie if we want to run windows clients as well, would that mean licensing 48 cores for win server and 48 cores for win clients (if they even have an option for clients) for the same physical compute node? | 16:31 |
| noonedeadpunk | Sorry, never knew license prices - it all was behind the scenes | 16:33 |
| *** ansmith_ has quit IRC | 16:33 | |
| jrosser | zbr: is your vars loading change affected by hash_behaviour setting? | 16:35 |
| *** rpittau is now known as rpittau|afk | 16:35 | |
| zbr | jroll: it should not, but I can always make a local test | 16:37 |
| zbr | which i am going to to anyway, just to be sure. | 16:38 |
| *** markvoelker has quit IRC | 16:41 | |
| jroll | I think you mean jrosser :P | 16:42 |
| zbr | indeed | 16:42 |
| *** markvoelker has joined #openstack-ansible | 16:44 | |
| *** csmart has quit IRC | 16:45 | |
| *** miloa has quit IRC | 16:46 | |
| *** irclogbot_2 has quit IRC | 16:47 | |
| *** irclogbot_2 has joined #openstack-ansible | 16:47 | |
| *** irclogbot_2 has quit IRC | 16:49 | |
| *** irclogbot_0 has joined #openstack-ansible | 16:49 | |
| zbr | jrosser: tests done, hash_behaviour affects loading in the expected way. | 16:50 |
| zbr | but I would personally not rely on that because is a setting external to the role | 16:50 |
| jrosser | So the setting in ansibke.cfg determines if the vars are merged or replaced then? | 16:50 |
| *** ansmith_ has joined #openstack-ansible | 16:50 | |
| zbr | yes | 16:51 |
| jrosser | Does that mean that with your change the behaviour is potentially different depending on the config outside the role? | 16:51 |
| *** csmart has joined #openstack-ansible | 16:52 | |
| zbr | yes but the same could be said about previous code: user could define these vars before calling the role. and with altered hash_behavior, it would get the same kind of result. | 16:53 |
| zbr | if someone is playing with hash_behaviour they are doing it at their own risks, and we cannot really protect from it. | 16:54 |
| zbr | hash_behaviour applies to the entire priority stack (11 levels last time i checked) | 16:55 |
| *** gillesMo has quit IRC | 16:58 | |
| zbr | but if you have ideas for improving the pattern, I am more than happy to hear. | 16:59 |
| zbr | this pattern is needed to almost any role, so it makes sense to polish it very well. | 16:59 |
| zbr | i fixed on suse bug recently (space in distribution name) | 16:59 |
| zbr | hopefully I will make an ansible module at some point, but we will have to wait ages before being able to use it. | 17:01 |
| jrosser | I think here if this is changed it probably would need to be changed across all OSA roles | 17:02 |
| zbr | i could do that, bit of copy/paste work for the moments when I am tired. | 17:03 |
| *** jcosmao has left #openstack-ansible | 17:07 | |
| *** allanb has left #openstack-ansible | 17:32 | |
| *** spatel has joined #openstack-ansible | 18:20 | |
| *** goldenfri has joined #openstack-ansible | 19:06 | |
| *** BjoernT has quit IRC | 19:44 | |
| *** allanb has joined #openstack-ansible | 19:51 | |
| *** kplant has quit IRC | 19:53 | |
| *** spatel has quit IRC | 20:04 | |
| *** ansmith_ has quit IRC | 20:13 | |
| *** ansmith_ has joined #openstack-ansible | 20:53 | |
| *** EmilienM is now known as EmilienM|pto | 21:01 | |
| *** markvoelker has quit IRC | 21:01 | |
| *** markvoelker has joined #openstack-ansible | 21:11 | |
| *** markvoelker has quit IRC | 21:21 | |
| *** allanb has quit IRC | 21:31 | |
| *** markvoelker has joined #openstack-ansible | 22:23 | |
| *** markvoelker has quit IRC | 22:28 | |
| *** allanb has joined #openstack-ansible | 22:38 | |
| openstackgerrit | Merged openstack/openstack-ansible master: Default nova_cinder_rbd_inuse when it is not set https://review.opendev.org/675429 | 22:53 |
| *** weifan has joined #openstack-ansible | 23:19 | |
| *** weifan has quit IRC | 23:22 | |
| *** tosky has quit IRC | 23:32 | |
| *** weshay is now known as weshay_pto | 23:37 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!