Thursday, 2020-09-17

*** MickyMan77 has joined #openstack-ansible00:52
*** MickyMan77 has quit IRC01:02
*** MickyMan77 has joined #openstack-ansible01:28
*** MickyMan77 has quit IRC01:36
*** spatel has joined #openstack-ansible01:39
*** cshen has joined #openstack-ansible01:52
*** cshen has quit IRC01:57
*** spatel has quit IRC02:13
*** spatel has joined #openstack-ansible02:22
*** idlemind has quit IRC02:24
*** idlemind has joined #openstack-ansible02:24
*** spatel has quit IRC02:52
*** MickyMan77 has joined #openstack-ansible03:01
*** MickyMan77 has quit IRC03:11
*** spatel has joined #openstack-ansible03:20
*** spatel has quit IRC03:24
*** MickyMan77 has joined #openstack-ansible03:52
*** MickyMan77 has quit IRC04:01
*** evrardjp has quit IRC04:33
*** evrardjp has joined #openstack-ansible04:33
*** watersj has quit IRC04:47
*** dave-mccowan has quit IRC04:51
*** MickyMan77 has joined #openstack-ansible05:15
*** gyee has quit IRC05:38
*** miloa has joined #openstack-ansible05:54
*** kukacz has joined #openstack-ansible06:18
*** MickyMan77 has quit IRC06:19
*** shyamb has joined #openstack-ansible06:43
*** BlackFX has quit IRC06:57
*** mensis has quit IRC06:59
*** BlackFX has joined #openstack-ansible07:02
*** MickyMan77 has joined #openstack-ansible07:13
*** andrewbonney has joined #openstack-ansible07:22
*** jawad_axd has joined #openstack-ansible07:32
*** yasemind has joined #openstack-ansible07:52
*** maharg101 has joined #openstack-ansible08:05
*** cshen has joined #openstack-ansible08:36
openstackgerritOpenStack Proposal Bot proposed openstack/openstack-ansible master: Imported Translations from Zanata  https://review.opendev.org/75239708:45
*** shyamb has quit IRC08:52
*** shyamb has joined #openstack-ansible09:03
*** shyam89 has joined #openstack-ansible09:26
*** shyamb has quit IRC09:27
*** shyam89 has quit IRC09:31
*** SecOpsNinja has joined #openstack-ansible09:39
*** admin0 has joined #openstack-ansible10:12
*** tosky has joined #openstack-ansible10:19
*** shyamb has joined #openstack-ansible10:24
*** cshen has quit IRC10:31
*** cshen has joined #openstack-ansible10:44
openstackgerritMerged openstack/openstack-ansible master: Imported Translations from Zanata  https://review.opendev.org/75239710:45
*** shyamb has quit IRC11:08
*** watersj has joined #openstack-ansible11:37
*** shyamb has joined #openstack-ansible12:07
*** shyam89 has joined #openstack-ansible12:14
*** shyam89 has quit IRC12:14
*** dave-mccowan has joined #openstack-ansible12:15
*** shyamb has quit IRC12:17
SecOpsNinjafrom what im seeing the magnum creates k8s vms in /var/lib/nova/instance using local storage from compute node (in my case my compute nodes only have a few ssd raid).  is there any way to indicate nova to always use boot-from-volume using cinder? i can't find any way to override magnum/heat vm volumes.  is there any way to do that in openstack-ansible? maybe i need to raise the compute local12:22
SecOpsNinja storage but i was trying to always use the space in cinder (storage nodes). is this possible?12:22
*** cshen has quit IRC12:53
noonedeadpunkI think you should always pass argument while creating template for that12:57
noonedeadpunk*magnum template12:58
noonedeadpunkie `–volume-driver cinder`13:00
noonedeadpunkhttps://docs.openstack.org/magnum/latest/user/#ephemeral-storage13:01
*** cshen has joined #openstack-ansible13:14
SecOpsNinjayep i have taht13:16
SecOpsNinjabut that is only aplied for doker images13:16
SecOpsNinjaim trying to force flavor at 0 disk size and see what happens in k8s cluster creation13:16
SecOpsNinjathe other solution who be to mont /var/lib/nova/instances as a nfs client in os_nova or try to month a ceph cluster (that i don't have enoght expriencie yet)13:17
*** pcaruana has quit IRC13:24
*** mensis has joined #openstack-ansible13:50
SecOpsNinjaok the problem of the disk was resolved but now i can't find a way to resolve the HEAT stack resource problems like created_failed. is there any way to retry that resource that failed or i need to delete and recreate the k8s cluster  again?14:26
*** cshen has quit IRC14:50
jrosserSecOpsNinja: you need to log into the cluster VM and look at the cloud-init and heat agent logs to see why it failed really14:57
SecOpsNinjajrosser,  yep i was able to check in stack resource list the cause of the problem (it was saing the docker volume was already attached to  k8s node) but even after detache the disk and force update with 2openstack stack update avr-5g-no-keystone-i3pxmrukjxle --existing" it way update_complete but k8s clsuter is still in created_failed state...  is there any way to force cluster to check if i15:01
SecOpsNinjats state is already in correct state (i supossed that correcting heat stacks deployed by magnum would resolve this)?15:01
*** cshen has joined #openstack-ansible15:02
*** cp- has quit IRC15:04
*** cp- has joined #openstack-ansible15:09
*** gyee has joined #openstack-ansible15:15
*** spatel has joined #openstack-ansible15:21
*** spatel has quit IRC15:25
*** spatel has joined #openstack-ansible15:44
spatelhurray!! senlin role got merge to master :)15:44
jrossernoonedeadpunk: what do people do for galera backups? noticing that we don't have anything in the role to take a backup for a backup server to then come in and copy out16:01
jrosserlike cron job or something for mysqldump or mariadb-backup16:02
*** gshippey has joined #openstack-ansible16:03
noonedeadpunkmariadb-backup would be my choice16:06
noonedeadpunkbut tbh not sure we should take care of backups16:07
noonedeadpunkas actually somebody may use some 3rd party solutions for taking incremental backups16:08
noonedeadpunkbut I think I had a cron job in one place for takin maria incremental backups when it was innobackupex16:09
noonedeadpunk(it was for 10.1 maria or smth)16:09
jrosseryes was kind of guessing everyone had their own solution here16:11
noonedeadpunklike this http://paste.openstack.org/show/798022/16:11
noonedeadpunkbut it was befor migration to openstack anyway... and for openstack I guess we do backups before upgrades only...16:12
noonedeadpunkanyway, xtrabackup syntax can be pretty simple adopted to work with mariadb-backup (as mariadb-backup has xtrabackup base anyway)16:13
noonedeadpunkseveral of regions are backuped with r1soft at all, so...16:13
jrosseryeah, so we use the galera role for other things too and were looking at making a simple optional cron to use mariadb-backup16:13
noonedeadpunkwe can do that tbh16:14
jrosserbut then something else deployer specific would come along and rsync/scp/whatever the data out16:14
noonedeadpunkbut let's use systemd-timer) it's part of systemd-role anyway16:14
noonedeadpunkjrosser: or, you can do mount nfs/cephfs/manila or use swift, ec2, etc lol16:15
jrosseryeah, so if we define a timer to backup and a destination path that should be everything covered16:15
jrosserand make it opt-in16:16
noonedeadpunkyeah, and we can take that script as a base, as it has been working for years pretty good, but it used 2 crons - like weekly for full backup and daily for incremental or whatever16:17
noonedeadpunk(and supports postgres as well lol)16:17
noonedeadpunkcould be written better, but yeah16:18
noonedeadpunkworth revising anyway16:18
*** mensis has quit IRC16:19
*** MickyMan77 has quit IRC16:22
*** MickyMan77 has joined #openstack-ansible16:23
noonedeadpunkjrosser: but not so sure about ow relevant backups are with osa manifesto https://docs.openstack.org/openstack-ansible/latest/reference/architecture/manifesto.html#project-scope16:25
noonedeadpunk`However, this project solely focuses on the deployment of OpenStack and its requirements.`16:25
noonedeadpunkbackups is kind of requirement I guess...16:25
jrosserright - thats why i ask that it's fairly obviously missing16:27
noonedeadpunkas then I think we can create bridges for deployers as well... that is pretty anoying preparation step for each compute node tbh16:27
*** cshen has quit IRC16:30
*** cshen has joined #openstack-ansible16:30
*** MickyMan77 has quit IRC16:31
noonedeadpunkbut yeah, I think we can add it, it not so big work and we almost have everything needed, but really profitable16:33
*** cshen has quit IRC16:34
*** cshen has joined #openstack-ansible16:51
CeeMacjrosser: was there not a backup utility in osa-ops?16:58
jrosserCeeMac: oh quite possibly, need to take a look there16:58
CeeMacthink i tried that once, couldn't work out how to do a restore from that though :/16:59
jrosserhttps://github.com/winggundamth/openstack-ansible-backup/blob/master/tasks/backup_galera.yml16:59
*** tosky has quit IRC17:00
jrosseri was kind of thinking its easy to re-deploy galera then forget to come along with $random-other-thing that tells it to do backups17:00
jrosserwhich i think points to some minimal capability in the galera role being a good thing17:01
CeeMacmakes sense17:05
logan-personally feels like scope creep in the deployment role, but id be surprised if theres nothing like that in the ops repo. this is a backup playbook i use often https://gist.github.com/logan2211/05e1b65041b92631ab8ab51f807c813417:10
*** pcaruana has joined #openstack-ansible17:18
*** andrewbonney has quit IRC17:19
*** spatel has quit IRC17:20
jrosserlogan-: yes it is kind of scope creep. we use the galera role for OSA and other things and i was trying to avoid solving the same problem twice/differently17:36
*** spatel has joined #openstack-ansible17:36
jrosserif there was a way to hook in a custom role/playbook to OSA entirely from data in /etc/openstack_deploy that would be another route to do this17:36
*** spatel has quit IRC17:41
*** SecOpsNinja has left #openstack-ansible17:41
*** mathlin has quit IRC17:53
*** irclogbot_0 has quit IRC17:58
*** mathlin has joined #openstack-ansible17:58
*** cshen has quit IRC18:00
*** irclogbot_1 has joined #openstack-ansible18:00
*** jawad_axd has quit IRC18:05
mnasernoonedeadpunk: do you think its possible for you to organize the upcoming ptg stuff possibly?18:21
mnaseror maybe jrosser ?  sorry -- i am really swamped with a lot of things lately18:24
mnaserhttps://ethercalc.openstack.org/7xp2pcbh1ncb18:24
*** cshen has joined #openstack-ansible18:34
*** cshen has quit IRC18:39
*** cshen has joined #openstack-ansible18:46
*** cshen has quit IRC18:58
*** openstackgerrit has quit IRC19:21
*** kukacz has quit IRC20:02
noonedeadpunkmnaser: yep, sure, will pick it up20:05
* noonedeadpunk wondering how he missed PTG registration ML20:09
rh-jelabarreI've been looking to see if Ansible allows comments *after* you have set a variable in the defaults/main.yaml file.  Such as something like:   "Color: green       # no, BLUE!" as opposed to having to put it on a line above20:12
noonedeadpunkhm, eventually I still can't find any email to all regarding vptg....20:13
*** watersj has quit IRC20:13
noonedeadpunkok, should add PTG to filters20:13
*** admin0 has quit IRC20:17
noonedeadpunkok, so as for now I filled in wednesday 4UTC-8UTC and thursday 21UTC-23UTC. I gouess that should be enought considering we didn't have much progress since last PTG in feauters implementation20:21
noonedeadpunkwe can discuss time closely during the upcoming meeting20:22
*** rgogunskiy has joined #openstack-ansible20:29
*** openstackgerrit has joined #openstack-ansible20:31
openstackgerritDmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible master: Add upgrade jobs template  https://review.opendev.org/75256720:31
*** tosky has joined #openstack-ansible20:34
*** cshen has joined #openstack-ansible20:54
*** jbadiapa has quit IRC20:57
*** cshen has quit IRC20:58
openstackgerritRafael Folco proposed openstack/openstack-ansible-os_tempest master: Fix tempest workspace initialization  https://review.opendev.org/75257121:18
*** miloa has quit IRC21:25
*** gshippey has quit IRC21:42
*** rgogunskiy has quit IRC21:53
*** jhesketh has quit IRC21:55
*** tosky has quit IRC22:19
*** mgagne has joined #openstack-ansible22:37
*** cshen has joined #openstack-ansible22:54
*** cshen has quit IRC22:58
*** jhesketh has joined #openstack-ansible23:31

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