infrastructure November 2022

infrastructure@lists.fedoraproject.org

19 participants
14 discussions

planet
by Kevin Fenzi 19 Apr '23

19 Apr '23

Hey folks. I thought I would open a discussion about fedoraplanet and possibly some plans for it. Right now: fedoraplanet.org runs on people02.fedoraproject.org (aka fedorapeople). To add a blog/rss feed you have to login there and edit your .planet file, then scripting pulls all those .planet files and tries to fetch all the feeds and then serves them up at http://fedoraplanet.org. It uses a app called 'venus' to do this. venus is written in very old python2 and very very dead upstream. We run into the following problems with it: * Sometimes it gets stuck and just stops processing until it's killed. * It's serving on a http site, which causes people to ask us to make it https, but that would just change the errors because many feeds it pulls are still http since they were added back before letsencrypt existed. * We have a handy 'website' field in our new account system, but aren't using it at all. * The .planet parsing is poor, any number of things can cause it to break. We have two open tickets on it: - https://pagure.io/fedora-infrastructure/issue/10383 (upgrade to pluto, a ruby based, but maintained thing) - https://pagure.io/fedora-infrastructure/issue/10490 ( planet not served via ssl) Which I am just going to close now. So, I can think of a number of options and would love everyone who has thoughts on it to chime in: 1. Do nothing. Venus "works" and .planet files are cool and retro. 2. Switch to pluto and use account system 'website' fields of contributors. We could likely shove it in openshift and serve it directly from there to avoid fedorapeople entirely. (This would likely break anyone who has multiple feeds in there) 3. Switch to something better/bigger. I would think (although I don't know) that there might be something that would not only aggregate rss feeds for contributors, but perhaps mastodon/twitter/whatever also. 4. Planets are old and tired, just drop the entire thing. People can maintain their own rss lists. 5. Planets are old and tired, just drop the entire thing. But also, get our social media people to maintain contributor / interesting lists. ie, the fedoraproject twitter account could maintain a list of 'fedora contributors' and 'fedora packagers' or whatever. 6. Switch to pluto as in 2, but also setup some curators. Have a 'firehose' of all feeds, but the main fedora planet would be just curated things that are known to be related to fedora and not off topic or unrelated. 6. Get someones (not it!) to take in all the twitter/facebook/mastodon/blog posts/rss feeds and post some kind of curated round up every week or something. 7. Your brilliant idea here! So, thoughts? this is not at all urgent, but we should end up doing something with it sometime. :) kevin

13 44

Rethinking fedora websites deployment
by Francois Andrieu 04 Dec '22

04 Dec '22

Hi everyone! The Websites & Apps team is currently working on rewriting all major Fedora websites (such as getfedora, spins & alt) and I believe this is a good opportunity to revisit the current deployment workflow and try to make it simpler. Currently, websites are being built in Openshift, with a cronjob running every hour that fetches code from git, builds it, then saves it to an NFS share. That same NFS share is also mounted on sundries01, which exposes it through rsync for proxies to sync it, then serve it to the world. I have a few solutions in mind to replace that, and I would like your input on them. A) Full Openshift We can build, and deploy the websites directly in Openshift, and serve it from here. Just like silverblue.fp-o. While this is probably the most straightforward solution, it has one major downside: if Openshift is unavailable for any reason, our major websites become also unavailable. I believe this is why we are still using our proxies to host them, as such a scenario is unlikely to happen on every single one of them at the same time. B) Same as before, with a twist We build on Openshift, but instead of going through NFS and sundries with rsync, we store the websites on S3 storage provided by Openshift, then we sync the proxies using `s3cmd sync`. C) Same as B, but with an external builder We already build the new websites on Gitlab CI, and since the S3 gateway is accessible from the outside, we could just push the build artifacts to s3 directly from GitLab CI. Then sync the proxies from it. D) Keep using the Openshift->Sundries->proxies workflow E) Your solution here. We could also improve B and C by adding fedora-messaging to the mix to trigger a proxy resync as soon as a new build is available instead of doing so every hour. What do you all think? -darknao

6 8

Shared Redis instance
by Aurelien Bompard 29 Nov '22

29 Nov '22

Hey folks! The new version of FMN will run in OpenShift and will use Redis as a cache backends (we chose it over memcached because it can do native "is-this-string-in-this-set" operations). I can deploy redis inside my openshift project easily enough , but I was wondering if it would be worthwhile to have a shared Redis instance, like we have a shared PostgreSQL instance. It's not just for ease of use, but I expect to store quite a bit of data in our Redis instance, and since we don't attach persistent storage to OpenShift that means that it will live in the pod's memory. So I'm being conscious of the memory hog it can become. Unless I'm mistaken there can be several databases in the same Redis instance, so we could share it between projects without stepping on each other's toes. What do you think?

3 6

Re: Rethinking fedora websites deployment
by darknao 25 Nov '22

25 Nov '22

On 2022-11-24 18:58, Jan K wrote: > What is likelyhood of Openshift going down? A would be best > solution if stable enough. > > copperi > Openshift is quite stable. But everything around it may not. Having a network outage, internal VPN issue, datacenter incident, or just an openshift migration that goes wrong can still happen.

2 1

What is Catalog's UV process?
by Fanny Zeng 23 Nov '22

23 Nov '22

As a must-see field in the catalogue printing market, album printing has shown a good development prospect for many years. It can not only use as an art display but also can use as a means of corporate publicity. Compared with a single pattern or text, the perfect combination of these two elements has an incomparable advantage. We often hear about the catalog's cover needing to do partial UV. So, what is the UV procedure? In fact, UV means ultraviolet. In the printing industry, this is a special ink treatment procedure. It is also referred to as the UV glazing process. The catalog cover has undergone partial UV treatment. Will clarify the parts to highlight. Such as the logo. After a partial UV treatment. The logo part will have a swelling sensation and a special brightness after the polish. The text and images after UV processing are more textured. More three-dimensional, and more artistic, giving the final touch. Depending on the UV zone, it is divide into full-screen UV and partial UV. The UV effect of the entire page is usually not very evident. It is like the super polish on the surface, but it is lighter than the super polish. It does not protect the surface from the printed material. Some protection, like the prevention of streaks. The smell on the nose, the smell is quite unpleasant, there is a smell of plastic dough. There are also many kinds of UV inks. These inks tend to have an unusual shine and texture. They have mirror ink, frosted ink, foam ink, wrinkle ink, and hammer ink. As well as color sand ink, snow ink, ice ink, pearl ink, crystal ink, laser ink and so on. The most common is spectacular UV. The equipment used in UV calls a UV light curing machine. And the principle is to cure UV ink by ultraviolet light irradiation. We can apply partial UV either after lamination or on the printed matter. But to highlight the effect of local UV. They carry it out after the lamination of the printed matter, and it is sub-filmed. It accounts for about 80% of local UV products. The printed matter after UV is difficult to recycle. chinaprinting4u is a China printing company with 20 years of printing experience. If you need more professional catalog printing services, please contact us. Article source:https://www.chinaprinting4u.com/news326.htm

1 0

CPE Weekly Update – Week 46 2022
by Lenka Segura 18 Nov '22

18 Nov '22

*Hi everyone,This is a weekly report from the CPE (Community Platform Engineering) Team.The report could be found at https://communityblog.fedoraproject.org/cpe-weekly-update-week-46-2022 <https://communityblog.fedoraproject.org/cpe-weekly-update-week-46-2022>/.If you want to receive weekly reports by emails in the future, please subscribe to either https://communityblog.fedoraproject.org/ <https://communityblog.fedoraproject.org/> or https://discussion.fedoraproject.org/c/news/commblog/61 <https://discussion.fedoraproject.org/c/news/commblog/61>. We will stop sending them in 2023.Regards,CPE Team*

1 0

Fedora 37 Final freeze now in effect!
by Kevin Fenzi 16 Nov '22

16 Nov '22

Greetings. we are now in the infrastructure freeze leading up to the Fedora 37 Final release. This is a final release freeze. We do this to ensure that our infrastructure is stable and ready to release Fedora 37 when it's available. You can see a list of hosts that do not freeze by checking out the ansible repo and running the freezelist script: git clone https://infrastructure.fedoraproject.org/infra/ansible.git ansible/scripts/freezelist -i inventory Any hosts listed as freezes is frozen until 2022-10-18 (or later if release slips). Frozen hosts should have no changes made to them without a sign-off on the change from at least 2 sysadmin-main or rel-eng members, along with (in most cases) a patch of the exact change to be made to this list. Thanks, kevin

2 2

I've problem with gdm on wayland
by Setve 16 Nov '22

16 Nov '22

Hello, I executed the update command "sudo dnf update --refresh", and immediately after that without rebooting, I executed the upgrade command "sudo dnf system-upgrade download --releaseservr=37" in the boot normal mode, the boot process takes place until "Started gdm.services" reached, and here it freezes. I cannot use "ctrl-alt-f1 to f7",using a liveusb i uncomment "waylandenable=false" at /etc/gdm/custom.conf. It worked normally and you logged in X11, i've intel gpu. I used it fpaste And these outputs https://paste.centos.org/view/2f797796 In addition to an attached file to journalctl -b -1 -r. thank you.

2 3

Freeze break request: update/reboot armv7 builders
by Kevin Fenzi 15 Nov '22

15 Nov '22

Hey folks. I'd like to update/reboot our buildvm-a32* builders. I'm hoping that the 6.0.x kernel will make them happier building python packages. See https://pagure.io/releng/issue/11095 So, this means applying updates on them and rebooting them all. I don't see how this could affect the release tomorrow. f37 doesn't actually use armv7 builders for anything, it's already done anyhow if it did. :) +1s ? kevin

3 3

fate/plans for fedmsg-irc
by Kevin Fenzi 15 Nov '22

15 Nov '22

Greetings everyone. I thought I would open a discussion about fedmsg-irc and what we want to do with it moving forward. First some background on what it is. ;) fedmg-irc is part of fedmsg (Our old zmq message bus). It's a small irc client that listens to the fedmsg bus and sends messages to a irc channel based on it's configuration. It used to send to a bunch of channels, but we dropped most of them off when we setup matrix bridges. The only two it has left are: 1. #fedora-fedmsg - This channel gets a constant stream of (most) everything from the message bus. (and #fedora-fedmsg-stg for staging) 2. #fedora-releng - This channel gets reports about releng related events (failed composes, compose starts, syncs, tickets, etc). Personally I have found 1 useful because my irc client logs it and I can use handy things like grep to look for messages. Also, if something is wrong and messages aren't flowing by there I can notice and fix things. Also, it's sometimes useful because you can use a IRC hilight and see messages of interest go by. All that said, there are probibly better ways to do all those things: I could make a fedora-messaging listener to just log everything, or get datagrepper better so I didn't want use grep or have alerts for lack of messages, etc. 2 is there because releng folks liked the notices about things and it allowed releng to react faster to things like failed composes, etc. So, likely we don't want to keep running the fedmsg-irc instances forver, but of course we can keep them going for a while. Options then would be: A) Just stop the service and let people figure out their own alternatives. B) Add something to do this to the FMN re-write (ie, have a way to configure a channel/room instead of a user)? (Ideally a matrix thing not a IRC thing) C) Retire the service, but write something based on fedora-messaging that interested folks could use to self service do the same thing. (Also ideally a matrix thing) D) Your idea here. Thoughts everyone? kevin

3 2

Problems with my @fedoraproject.org Mail adress
by Simon Bachenberg 07 Nov '22

07 Nov '22

Hi, Today I wanted to become a package maintainer and started following the official guide: https://docs.fedoraproject.org/en-US/package-maintainers/Joining_the_Packag… While i was joining the Mail lists I notices my "approved" mail address bachenberg(a)fedoraproject.org. But i dont know how to receive Mails from that address. So i tried sending myself an e-mail, but I got "550 5.1.1 <bachenberg(a)fedoraproject.org>: Recipient address rejected: User unknown in local recipient table" How do i use my @fedoraproject.org mail address and how did I approve it in the past ? Greetings, Simon

2 1

FMN / notifications status
by Kevin Fenzi 07 Nov '22

07 Nov '22

So, yesterday we managed to get the python3 version of fmn working in staging seemingly well, so I stupidly decided to try and migrate production to it. This has been fraught with peril. ;( First, only the notifs-backend01 ip is allowed to reach out to IRC, so after saving off the notifs-backend01 (python2) fmn disk and vm xml I tried to get notifs-backend02 to use 01's ip. That works, but made things like IPA mad, since the hostname/ip was mismatched. So, I just redeployed a new notifs-backend01 with the python3 version. Sadly, the IRC part which we hadn't been able to test in staging, wasn't working right at all. So, I went back to the old notifs-backend01. Unfortunately, that vm is now not processing correctly either. (Even though it's the same exact VM that was there before). :( With help from pingou, we cleaned up some errors, but it's still not seeming to be sending out things. It may be that it's just loading caches and such and will start working here soon. So, here's my plan: * Leave the current python2 one alone for a while and hope it starts processing things again. * Later in the day, I am going to save off the current python2 backend01 (again) and deploy the python3 one (again). Hopefully that gets at least emails flowing again. * pingou is going to look at it when he gets in in the morning tomorrow and fix issues. Hopefully that will result in a working python3 FMN. ;( Since we are going into freeze this morning, I'm going to treat this work as "it's an outage, fix it" and not worry about the freeze for it. So, sorry for the lack of notifications last night and today, I wish things had gone smoother. kevin

5 19

Freeze Break: openssl update
by Kevin Fenzi 01 Nov '22

01 Nov '22

Greetings. I don't think we are particularly vulnerable to the openssl issue(s) that came out today, but just to be carefull I would like to apply the update to all our Fedora machines and RHEL9 machines (when that update is available). Also, I would want to restart at least httpd on the proxies and perhaps some more services there. +1s? kevin

5 4

Freeze break request: Drop getfedora.org/cloud redirect
by darknao 01 Nov '22

01 Nov '22

Hi everyone, As part of F37 release, we are restoring Fedora Cloud as an official edition on getfedora.org [1]. There is an old redirect to https://alt.fedoraproject.org/en/cloud/ that will prevent the new page on getfedora to show up. We need to get rid of it before the release. See PR [2]. Any +1s? [1]: https://fedoraproject.org/wiki/Changes/RestoreCloudEdition [2]: https://pagure.io/fedora-infra/ansible/pull-request/1240 Thanks, -darknao

3 2

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

infrastructure November 2022