Hi everyone, during upgrade preparations in the staging environment, we inadvertently killed our production system. It is an issue we encountered in the past and noted down how to work around it, but human error ...
We're restoring service as soon as possible. No data is lost, just a performance issue that needs fixing.
See you soon.
bird.makeup/users/nurderk/stat…
#Pilze #natur #photography
"Zwei schwarze RHIBs der Garde Nationale🇹🇳 , vermutl. von DEUTSCHLAND … gespendet … nähern sich … Eins … beginnt, 3x Kreise im Wasser um das Boot zu ziehen, um Wellen zu erzeugen … nimmt Fahrt auf und rammt das Heck des Bootes."
Am 5. April 2024 sinkt vor #Sfax ein fragiles Metallboot mit 42 Menschen aus #Gambia, #Guinea & #Sierra Leone. Die tunesische Garde Nationale wird von den Überlebenden beschuldigt, den Schiffbruch verursacht zu haben. Mind. 15 Menschen ertranken.
irpimedia.irpi.eu/en-the-respo…
The responsibilities of Tunisia’s Garde Nationale in the shipwreck of April 5th 2024 - IrpiMedia
Testimonies, satellite images and videos from social media show how at least 15 migrants died. The Tunisian coastguards have been recently put in charge of a rescue zone at seaBeatrice Cambarau (IrpiMedia)
Es braucht einen kulturellen Wandel. Wir müssen uns fragen, ob wir in einer Hackordnung oder in einer Gemeinschaft leben wollen.
Pro #BGE!
Die Crew der #SEAEYE4 wurde Augenzeugin von neuen Grausamkeiten im Mittelmeer. Maskierte zwangen 22 Menschen, in internationalen Gewässern von Bord eines fahrenden Bootes zu springen und ließen sie dann schutzlos zurück. Die Crew konnte glücklicherweise alle Menschen retten. Wenn die Crew nicht so schnell gewesen wäre, hätten Menschen ertrinken können.
#LeaveNoOneToDie #BorderViolence #SeaEye
#Seebrueccke
1/x
Eigentlich braucht es mehr als Seenotrettung.
Das Übel muß bei der Wurzel gepackt werden. Ich stelle mir eine europäische Mission vor, die dem Treiben der Schlepper ein Ende setzt. Und deren Boote ggfs auf Grund.
Bauch, Beine, Politik
Politiker auf #TikTok sind oft spröde und latent uncool. Caren Lay von der Linken rappt mit 51 über harte politische Themen. Und macht vielleicht gerade einiges richtig.
#Linke #AfD
zeit.de/kultur/2024-10/caren-l…
Caren Lay auf TikTok: Bauch, Beine, Politik
Politiker auf TikTok sind oft spröde und latent uncool. Caren Lay von der Linken rappt mit 51 über harte politische Themen. Und macht vielleicht gerade einiges richtig.Kirsten Küppers (ZEIT ONLINE)
Today is the last day to apply for the fellowship for maintainers, our new program that will pay #FOSS maintainers for a year to work on the important work of critical software infrastructure.
For the pilot year, we're looking for up to 5 fellows, with freelance contract and employment options. Submit your application by the end of the day: Sunday, 20 Oct 2024!
Boosts appreciated!
Details: sovereigntechfund.de/news/appl… #GetFediHired
Forgejo v9.0, which we'll (hopefully!) start using in a few hours, ships with a new quota feature for public and shared Forgejo instances!
We explained what this change will mean for you, the user (as well as some background information), in a new announcement titled "More power for you - what a storage quota will bring".
blog.codeberg.org/more-power-f…
More power for you – what a storage quota will bring — Codeberg News
tl;dr? To make abuse prevention easier, we are applying some default storage...blog.codeberg.org
Is there somewhere you can look up how much storage you are currently using ?
(I have a few docker images in the packages storage used in CI jobs, that get updated a lot on dependency updates. I think I configured the auto-deletion of old versions correctly, but with those muti-arch images it's hard to know if it's working 100% correctly)
@JonathanTreffler When viewing a repository, there is a bar that shows you the amount of commits, releases, etc...
The rightmost element shows you the storage that a specific repository of yours is taking up! ~n
It's time for BEE FACTS
For every donation to this link, I will post one (1) fact about bees!
Ich musste leider schon los. Vielen Dank für die netten Gespräche und die vielen alten und neuen Stimmen!
Für die Abschlussrunde habe ich noch eine unverbindliche bescheidene Frage an @timpritlove: WANN NÄCHSTE?! #Subscribe11 ist vor #subscribe12
Ich hab gar nicht mitbekommen dass der Domino Day wieder zurück ist 😍
rnd.de/medien/domino-day-rtl-b…
Domino Day: RTL bringt TV-Event 2020 zurück ins Fernsehen
Mit bis zu 13 Millionen Zuschauern war der „Domino Day“ ein absoluter Quotenhit bei RTL. Jetzt bringt der Sender das Event zurück auf den Bildschirm – was bisher bekannt ist.RedaktionsNetzwerk Deutschland
Die #ePA ist grundsätzlich eine sinnvolle Idee, aber die Umsetzung lässt zu wünschen übrig. Es drängt sich der Eindruck auf, dass das #Patientenwohl zu einer leeren Worthülse verkommt - im Mittelpunkt stehen weniger die Bedürfnisse der Menschen als ihre Daten.
Ein Kommentar von @lacrosse zur elektronische Patientenakte erscheint heute Abend um 19:30 Uhr auf dem Kuketz-Blog. Stay tuned! 🍿
Codeberg has two downtimes scheduled in the next two days, upgrading #Forgejo to the latest version as well as operation on our owned hardware.
See the notification at status.codeberg.eu/status/code… for more and up-to-date information.
Thank you for your understanding. We'll try to minimize service disruption.
Thank you.
How is the fight against the proprietary wave going? Have you found a rather non-manual way to ensure compliance?
If they're not licensed, they're technically a subset of the category you're talking about. We have explained how we deal with such cases manually over the past few days as well. Is there anything specific that I can help with here? ~n
EDIT: social.anoxinon.de/@Codeberg/1…
@girlintech I'm not responsible for handling these situations, but, from what I've gathered, it'd involve a warning, a popup with a reminder (if you use lots of space), or a personal ask to fix this (sent using your email address). Accidents happen and people can forget things sometimes.This isn't a rule applying to all situations, however. But please don't panic. 😀 ~n
EDIT: A further clarification from ~f: social.anoxinon.de/@Codeberg/1…
Reminder: We're soon on our way to the datacenter, starting maintenance in about two hours, which involves about 30 minutes of downtime for Codeberg.
We know that our availability was limited recently due to various factors, and we're sorry to add on it today, but it's the way to make the future setup more reliable.
We'll keep you updated as time permits.
*click* - a new server was just installed physically. Now cabling and testing begins.
A hardware reboot of the existing machine is scheduled later.
We're now installing new network hardware, requiring to take down the machine.
We are going down for maintenance now.
👋 Say hi to "Aburayama", our second server named after the mountain that inspired the first (now retired) Codeberg logo: en.wikipedia.org/wiki/Mount_Ab…
Just in case people want to know what's running inside it, we placed some stickers prominently on front.
Thanks to everyone involved for their work, and for your patience during the setup.
requests
Request what you need for your project: CI access, more repositories, storage, etc.Codeberg.org
build-deploy-forgejo/etc/gitea/public/img/codeberg.png at 42d20b19e73b792b8cb10676be8be6d799d70f17
build-deploy-forgejo - Build and deploy user-facing frontend (Forgejo)Codeberg.org
@frigidcode See codeberg.org/Codeberg-Infrastr…
Not the latest stuff, but since we do not need the computing power and just the redundancy, the hardware does the job.
meta/hardware/aburayama.md at main
meta - Organizational repo for Codeberg's Infrastructure: Documentation, Organizing, Planning.Codeberg.org
Herzlich willkommen im Fediverse, @EH_Ludwigsburg !
Schön, zu sehen, dass die akademische Welt hier weiter wächst.
Fabian N. T. 🦆
Als Antwort auf Codeberg.org • • •spooky Dragonas
Als Antwort auf Codeberg.org • • •F. Maury ⏚
Als Antwort auf Codeberg.org • • •Codeberg.org
Als Antwort auf F. Maury ⏚ • • •@x_cli We have a "staging" / "migration" environment on our production machine. It allows us to launch an airgapped snapshot of the production data (including database + a readonly snapshot of Ceph holding repositories, attachments etc).
It is very useful to perform extensive testing of the software, and it's easy to update it.
However, there is one caveat: The snapshot is readonly by default, and we use some overlay filesystems to make it writeable. There is a script, but it's not ...
Codeberg.org
Als Antwort auf Codeberg.org • • •@x_cli ... not launched automatically. It needs to be done after restoring fresh snapshots, which is usually not a big thing.
When the test machine tries to write to the snapshot without the overlay, we see a kernel bug and Ceph crashes completely. It is probably fixed in a newer version, but upgrades are currently held back by certain other constraints in our cluster, which we are going to remove by installing new hardware tomorrow and in the next weeks.
So far, we knew how to work around.
Codeberg.org
Als Antwort auf Codeberg.org • • •@x_cli The "mistake" this time was that we wanted to test something by quickly restarting the staging container, and we inadvertently launched the actual test before manually enabling the overlay filesystem.
It's a tiny detail that was not fully automated, and it bit us badly. We'll do better.
Δж3
Als Antwort auf Codeberg.org • • •Codeberg.org
Als Antwort auf Δж3 • • •We're using LXC containers on a single host. The containers can be seen as a VM replacement, not like Kubernetes / Docker (although some services run using containers in containers, too).
@x_cli