15:00:26 #startmeeting Infrastructure (2020-09-17) 15:00:26 Meeting started Thu Sep 17 15:00:26 2020 UTC. 15:00:26 This meeting is logged and archived in a public location. 15:00:26 The chair is siddharthvipul. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:26 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:00:26 The meeting name has been set to 'infrastructure_(2020-09-17)' 15:00:26 #meetingname infrastructure 15:00:26 The meeting name has been set to 'infrastructure' 15:00:26 #chair nirik pingou smooge cverna mizdebsk mkonecny abompard siddharthvipul mobrien 15:00:26 Current chairs: abompard cverna mizdebsk mkonecny mobrien nirik pingou siddharthvipul smooge 15:00:26 #info Agenda is at: https://board.net/p/fedora-infra 15:00:26 #info About our team: https://docs.fedoraproject.org/en-US/cpe/ 15:00:26 #topic aloha 15:00:51 morning. aloha 15:00:54 .hello2 15:00:55 creaked: creaked 'Will Chellman' 15:00:58 .hello siddharthvipul1 15:00:59 siddharthvipul: siddharthvipul1 'Vipul Siddharth' 15:01:32 .hello2 15:01:33 mobrien: mobrien 'Mark O'Brien' 15:01:33 ó/ 15:01:41 welcome you all :) 15:02:08 .hello mohanboddu 15:02:09 mboddu: mohanboddu 'Mohan Boddu' 15:03:27 #topic New folks introductions 15:03:27 #info This is a place where people who are interested in Fedora Infrastructure can introduce themselves 15:03:27 #info Getting Started Guide: https://fedoraproject.org/wiki/Infrastructure/GettingStarted 15:03:35 do we have anyone new here? :D 15:04:18 its been awhile so I'm new here again :) 15:04:46 creaked, welcome again ;) 15:04:51 welcome back creaked 15:05:18 .hello2 15:05:19 darknao: darknao 'Francois Andrieu' 15:05:24 .hello2 15:05:25 TomAPC: Sorry, but you don't exist 15:05:35 ouch :( 15:05:39 .hello2 TomAPC 15:05:40 TomAPC: Sorry, but you don't exist 15:05:48 .hello TomAPC 15:05:49 TomAPC: Sorry, but you don't exist 15:06:45 TomAPC: if your irc nick isn't the same as your fas account name, use .hello fasaccountname 15:06:54 .hello tomapc 15:06:55 TomAPC: tomapc 'Tom Collins' 15:06:59 Ah, case sensitive 15:07:17 yeah, fas accounts are all lower cased. 15:07:38 .hello zlopez 15:07:39 mkonecny: zlopez 'Michal Konečný' 15:08:00 we have a lot of people today :) 15:08:04 let's move to agenda 15:08:12 #topic Next chair 15:08:19 #info 2020-09-24 - mkonecny 15:08:30 we need a volunteer for 2020-10-01 15:08:33 do we have anyone? :) 15:09:52 * nirik can if no one else can... 15:10:00 I'll do it if no volunteers 15:10:08 I was about to say the same thing 15:10:12 great minds think alike :) 15:10:17 sooo.. who is it XD 15:10:25 fools seldom differ :D 15:10:26 mix all three and do the next 3? 15:10:33 mobrien, or that :P 15:10:39 nirik, works 15:10:51 #info 2020-10-01 - nirik 15:11:04 I cannot chair the meeting since sometimes I have conflicts :( 15:11:12 #info 2020-10-08 - mobrien 15:11:37 #info 2020-10-15 - siddharthvipul 15:11:44 works for me. 15:11:48 mboddu, no worries at all, we are good for a while :) 15:12:05 #topic announcements and information 15:12:05 #info CPE Sustaining EU-hours team has a Monday through Friday 30 minute meeting going through tickets at 0830 UTC in #centos-meeting 15:12:05 #info CPE Sustaining NA-hours team has a Monday through Friday 30 minute meeting going through tickets at 1800 UTC in #fedora-admin 15:12:05 #info Datacenter move is over, but some items still need to be done: see https://fedoraproject.org/wiki/Infrastructure/2020-post-datacenter-move-known-issues 15:12:07 #info F33 Beta freeze is in effect 15:12:28 #info F33 Beta is still NO GO 15:12:51 mkonecny, thank you :) 15:13:00 do we have anything else? 15:13:19 #info staging is about ~80% installed 15:14:07 Yay, we are almost there 15:14:40 although thats just installed. Still need to sync databases and fix all the broken things. ;) 15:15:52 let's move ahead if we have no announcement :) 15:16:07 #topic Oncall 15:16:07 #info https://fedoraproject.org/wiki/Infrastructure/Oncall 15:16:17 #info pingou is oncall for 2020-09-10 to 2020-09-17 15:16:17 #info mobrien is oncall for 2020-09-17 to 2020-09-24 15:16:28 before we transfer the oncall duty 15:16:42 who will volunteer for 2020-09-24 to 2020-10-01 15:17:14 it's a very easy and very helpful thing (if there is someone new here) :) 15:17:20 * nirik can if no one else. ;) 15:17:32 nirik, sure thing :) thank you 15:17:46 #info nirik is oncall for 2020-09-24 to 2020-10-01 15:17:53 .oncalltakeeu 15:17:53 mobrien: Kneel before zod! 15:18:01 mobrien, oh you are fast :D 15:18:08 #info Summary of last week: (from current oncall ) 15:18:12 pingou, all yours :) 15:19:19 I think he's not around right now. :( 15:19:32 nirik, let's get back to it later then 15:19:34 #undo 15:19:34 Removing item from minutes: INFO by siddharthvipul at 15:18:08 : Summary of last week: (from current oncall ) 15:19:49 #topic Monitoring discussion [nirik] 15:19:49 #info https://nagios.fedoraproject.org/nagios 15:20:07 so, not much differnt that last week. 15:20:15 some mustangs still down. 15:20:21 some virthosts with low swap 15:20:34 some proxies that mobrien is installing 15:21:08 After beta is out we should make a push to clean things up... 15:21:17 thats all, move on? 15:21:21 awesome 15:21:42 sooo... today the plan is to learn about fedora-infrastructure things :) 15:21:46 and the topic for today is 15:21:55 drumroll.... 15:22:10 #topic let's learn openshift setup by nirik 15:22:22 tadaaa 15:22:23 cool. 15:22:50 so, we have 2 openshift clusters currently. Both of them are OpenShift 3.11... 15:23:32 mostly because we installed them back before 4.x existed, and 4.x really wants to have dedicated hardware. Our openshift 3.11 clusters are running on vm's. 15:23:49 There's a production cluster and a staging one. Both are setup pretty much the same way. 15:24:42 for each cluster we have a 'os-control01' vm. This is just a small vm that checks out openshift-ansible git repo, adds our inventory and templates and runs ansible to deploy the rest of the cluster. 15:25:30 It's done this way instead of directly from our ansible host because this allows us to checkout/use a specific version of openshift-ansible and ansible without messing with our main ansible control host. 15:25:54 That playbook deploys: 3 master hosts and 5 compute nodes. 15:26:33 Access to the web side of the cluster is via our proxies. So user -> proxy -> openshift cluster 15:27:12 the proxies handle the ssl side of things, so for example we can use our *.fedoraproject.org cert for 'bodhi.fedoraproject.org' which is a openshift cluster app 15:27:37 All applications/projects are installed from our main ansible repo. 15:27:55 * pingou in another meeting indeed, I don't think I've had a single ping last week 15:27:57 You can see under roles/openshift-apps/ and playbooks/openshift-apps/ the config 15:28:33 was about to ask about the applications running, ^ solves that :) 15:28:35 so a playbook runs and copies templates/config to (by default) os-master01 /etc/openshift-apps/appname/ and then uses oc to load it into the cluster. 15:29:16 so this means you have to make any app changes via the playbooks/ansible git. Permissions are locked down to prevent changing things via web inteface. 15:29:45 this allows us to easily redeploy the cluster and all apps... just run the cluster playbook, then each apps and everything is done. 15:30:17 External storage is nfs and not dynamic. Ie, if we need storage we have to make a netapp volume, export it and make a volume in openshift. 15:30:28 Database is external postgres server 15:30:28 ansible really is a lifesaver 15:30:36 yeah, it's awesome. :) 15:31:00 any plans to migrate to 4.x? 15:31:03 So, thats the basics... any questions on the setup? 15:31:17 that's awesome nirik 15:31:27 what is it deployed on? stood up a openshift 4.2 cluster a few months back 15:31:45 creaked: yes, we would like to. We have put in for some hardware for next year for a cluster. :) also... we have a few machines we may be able to install a cluster on late this year, but they will be older/slower boxes we didn't need for anything else. 15:31:48 IMO the change in playbooks will be the biggest blocker (since everything is so nicely setup) 15:31:53 thanks nirik! interesting to learn as a newbie 15:32:15 creaked: right now, all the 3.11 ones are vm's... 15:32:27 nirik, Fabian installed 4.5 in some vms last week, might be worth a spike.. though I guess we have enough things to deal with now :) 15:32:41 gotcha gotcha cool stuff 15:32:50 siddharthvipul: shouldn't need much change... the app side hasn't changed much betwee 3.11 and 4.x... it's adding operators and cluster management that are completely different 15:33:10 yeah, I know it's possible... but it's... less supported I think? 15:33:23 but that may be an option indeed. 15:33:48 nirik, ah, I understand 15:34:01 If he has notes from that I'd love to see them. :) 15:34:10 nirik, will ask him tomorrow :) 15:34:32 cool. 15:34:36 anyone has any questions for nirik? 15:34:42 * nirik tries to think of anything else to mention. 15:34:42 let's wait a minute 15:35:36 The operators are making the life easier in OpenShift 4 15:35:50 mkonecny, welll, I guess you can say that 15:35:51 mkonecny: yeah, it's a big reason to move to 4 for sure. 15:36:16 It will be nice to have deployment of some apps as operators, so everyone could easily deploy them :-) 15:36:21 4 also handles upgrading much nicer... you just tell it to upgrade and it does it. :) 15:36:38 yeah one click upgrades are nice :) 15:36:54 Is there an eol for 3? 15:37:00 upgrades on 3.x you have to change some variables and re-run the cluster deployment... it takes a long time and sometimes fails for anoying reasons 15:37:15 nirik, can confirm that.. we have been upgrading it almost every week (minor version roll out pretty fast) 15:37:30 june 2022 15:37:33 https://access.redhat.com/support/policy/updates/openshift_noncurrent 15:37:47 yeah. 15:38:28 nirik, should we move ahead? 15:38:39 thanks a lot for taking out time to do this.. we really appreciate it :D 15:38:43 we did have a 4.x cluster (communishift) but it's down and unknown when it will be back. Hopefully someday 15:38:59 siddharthvipul: no problem. Happy to share what I know... great to get questions! 15:40:21 #topic summary from last week oncall 15:40:31 #info nothing to report 15:40:39 #topic Open Floor 15:41:18 * nirik has nothing 15:43:10 any packaging work for this group that I could take on? 15:45:17 creaked, are you in packagers group? 15:45:29 no 15:45:52 creaked: might be the folks working on the mailman3 stack could use some help... but I am not sure where they are at... 15:46:32 https://pagure.io/fedora-infrastructure/issue/8455 15:46:45 ok cool will take a look 15:47:35 creaked, meanwhile, you may also want to look at reviewing packages (to understand guidelines better), packaging some new/retired and needed packages etc. These will help you get into packagers group 15:49:16 ok thnx 15:50:35 let's end the meeting in 2 minutes if there is nothing else 15:53:15 ..3 15:53:18 ..2 15:53:32 #info thank you all for joining (or reading the logs) 15:53:36 #endmeeting