18:00:58 <nirik> #startmeeting Fedora Infrastructure Ops Daily Standup Meeting
18:00:58 <zodbot> Meeting started Wed Oct 14 18:00:58 2020 UTC.
18:00:58 <zodbot> This meeting is logged and archived in a public location.
18:00:58 <zodbot> The chair is nirik. Information about MeetBot at http://wiki.debian.org/MeetBot.
18:00:58 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
18:00:58 <zodbot> The meeting name has been set to 'fedora_infrastructure_ops_daily_standup_meeting'
18:00:59 <nirik> #chair mboddu nirik smooge pingou mobrien
18:00:59 <zodbot> Current chairs: mboddu mobrien nirik pingou smooge
18:00:59 <nirik> #meetingname fedora_infrastructure_ops_daily_standup_meeting
18:00:59 <zodbot> The meeting name has been set to 'fedora_infrastructure_ops_daily_standup_meeting'
18:00:59 <nirik> #info meeting is 30 minutes MAX. At the end of 30, its stops
18:00:59 <nirik> #info agenda is at https://board.net/p/fedora-infra-daily
18:01:00 <nirik> #topic Tickets needing review
18:01:03 <nirik> #info https://pagure.io/fedora-infrastructure/issues?status=Open&priority=1
18:01:12 <nirik> I had it set, just got distracted.
18:01:18 <smooge> no problem
18:01:46 <mboddu> Oh man, its time already, time is passing away very quickly
18:01:48 <nirik> #info no new infra tickets to triage today
18:01:54 <nirik> any releng ones mboddu ?
18:01:57 <mboddu> Yup
18:02:13 <mboddu> .releng 9808
18:02:15 <zodbot> mboddu: Issue #9808: Remove private branch ddiskit/private-ckalina-1885256 - releng - Pagure.io - https://pagure.io/releng/issue/9808
18:02:23 <mboddu> It should be low,low,ops?
18:02:37 <nirik> yep. +1
18:03:28 <mboddu> .releng 9809
18:03:29 <zodbot> mboddu: Issue #9809: Drop special restictions on fedora-release - releng - Pagure.io - https://pagure.io/releng/issue/9809
18:03:53 <nirik> I'm 100% fine with this, but I can't remember us agreeing to it... might just not be remembering tho.
18:04:47 <mboddu> I dont remember agreeing to it, but we definitely talked about it on non dist-git repos like releng, pungi-fedora and others..
18:05:52 <nirik> we did drop fedora-release from the secure-boot channel in koji...
18:06:00 <nirik> perhaps we agreed to more then... can't recall.
18:06:06 <nirik> anyhow, I'm in favor.
18:06:17 <mboddu> I am in favor and I dont recall as well
18:07:59 <smooge> I had a couple of old ones but most of them are political storms which can wait. 6801 was one which I didn't know where we got to (and where one should run hardlink and on what directories /pub/fedora-secondary and /pub/archive ?
18:08:25 <nirik> .ticket 6801
18:08:25 <zodbot> nirik: Issue #6801: run hardlink on fedora-secondary - fedora-infrastructure - Pagure.io - https://pagure.io/fedora-infrastructure/issue/6801
18:09:19 <nirik> ah yeah. so... our rawhide/branched compose scripts don't hardlink aarch64 with the other arches... because they are done seperately since they don't have everything in the same place
18:09:55 <nirik> again I wish we could put everything in the same tree...
18:10:08 <smooge> I remember you were looking at this recently
18:10:21 <nirik> yeah, I don't know that there is any easy way to fix it.
18:10:41 <nirik> without putting all the stuff in one tree and copying it with one rsync
18:10:44 <smooge> ok for me to put that info on the ticket and move it to high-trouble
18:11:17 <nirik> sure, or we could just close it and say 'sorry, we can't hardlink everything because we split composes up'
18:11:29 <nirik> or...
18:11:40 <nb> could we make a cron to run hardlink every week or something?
18:11:42 <nirik> one thing we could do is run hardlink after we rsync the content.
18:11:58 <nb> yeah
18:12:15 <nirik> nb: 100% useless, unless it's right after composes, because they would just unhardlink everything the next compose
18:12:22 <nb> oh
18:12:28 <nb> so we'd have to do it after every compose
18:12:46 <nirik> yes, and it would make composes take a lot longer and it would hit the storage pretty hard also
18:13:04 <nirik> but we could try it and see how bad it is.
18:14:29 <nirik> I guess I'd say lets try and see how bad, then if it's bad we revert it and close CANTFIX.
18:15:27 <nirik> thoughts?
18:15:34 <fm-admin> pagure.issue.comment.added -- smooge commented on ticket fedora-infrastructure#6801: "run hardlink on fedora-secondary" https://pagure.io/fedora-infrastructure/issue/6801#comment-696211
18:15:41 <smooge> added something
18:15:51 <mboddu> I am not 100% sure, but nothing wrong in trying
18:17:20 <mobrien> I have a quick question regarding #9388
18:17:42 <nirik> I can make a PR. I wonder if we shouldn't wait until after f33 tho... in case it hammers the storage too much for f33 compsoes to finish, etc
18:17:47 <smooge> nirik, where would this change in hardlink etc occur? in admin scripts or releng scripts
18:18:11 <smooge> I am ok with waiting til afterwords.. I just had a window that had me looking at doing a hardlink of /pub/archive
18:18:25 <nirik> smooge: pungi-fedora/nightly.sh script... I guess 3 times... once after syncing to pub/ once for alt/ and once for fedora-secondary
18:18:30 <smooge> and I was like.. do I need to do this still, did nirik do it, what ticket?
18:19:12 <smooge> wait till after F33
18:19:16 <mboddu> nirik: Better do it post f33 release
18:19:19 <smooge> mobrien, what was the question
18:19:31 <nirik> ok, can put in the PR and mark it post-freeze...
18:19:36 <nirik> .ticket 9388
18:19:37 <zodbot> nirik: Issue #9388: AWS: create S3 bucket for Testing Farm - fedora-infrastructure - Pagure.io - https://pagure.io/fedora-infrastructure/issue/9388
18:19:37 <mboddu> +1
18:19:54 <nirik> ah yeah, I was working on that before the meeting... sorry if you already were.
18:20:10 <fm-admin> pagure.issue.comment.added -- smooge commented on ticket fedora-infrastructure#6801: "run hardlink on fedora-secondary" https://pagure.io/fedora-infrastructure/issue/6801#comment-696212
18:20:40 <mobrien> I was going to take it earlier but I wanted to ask first, is it a case generally where we create a new bucket or have 1 main fedora bucket and do subdirs?
18:21:06 <nirik> we have been making new seperate buckets for things.
18:21:24 <nirik> and then roles/users get perms to only access their bucket arn
18:22:20 <nirik> If they need it to be public tho, we have to go to s3 and uncheck the 'deny all public access' thing. I guess they put that in because people were leaking things to the public they didn't intend to...
18:22:53 <mobrien> I read an article before about the things you can find on public S3 buckets
18:23:12 <nirik> yeah, pretty crazy.
18:24:25 <mobrien> We could do permissions based on dirs in a bucket if we need to, I know there is a limit for buckets per account/region
18:25:05 <nirik> ok, yeah, if thats better we could look at it... or if we hit the limit?
18:25:32 <smooge> alright we are hitting meeting limit
18:25:54 <mobrien> Ya, just soemthing to keep in mind for future but I'm getting off topic.
18:26:22 <smooge> #info smooge will be unable to attend meetings regularly until 26th
18:26:33 <smooge> #endmeeting