When I say two servers I mean two VMs to get the system to work effectively.
From memory, the admin interface doesn't get an SSL certificate issued to it. It perpetually stays HTTP. If you don't set up another server as a reverse proxy, it won't let you log in due to CORS issues. Add another server as a reverse proxy, and it'll come good and let you log in.
Hopefully that makes sense?
I mean, my local Colesworth doesn't really give a shit. I've seen people on the phone buying mass gift cards and had to step in because the staff just allow the purchases. I haven't heard a warning over the PA/in-store radio for years now. YMMV but it's still a problem around NW Sydney it seems.
Anyone else had experience with OneUptime?
As per title.
Wondering if anyone else here has had any experience with the self-hosted version of OneUptime? And specifically, has anyone had any luck setting it up behind Nginx Proxy Manager?
I've managed to set it up, but I'm honestly not 100% happy with how you have to essentially have two servers to host it (one being a reverse proxy for the admin interface, the other being the application core).
Don't get me wrong, it's neat and definitely full featured, but there is still a long way to go with it. For my use case, I wanted a public status page that people can subscribe to for updates. I'd come from UptimeKuma which was fantastic but lacked the subscriber feature. I used to use Cachet back in the day before it became abandonware (the original owner bought the rights back for it and has rebooted development for it, though!).
Gizmo: Just after a nap
She was asleep like this, but as I picked my phone up she woke up... still with tongue out.
Sorry about the downtime...
What happened?
Our SSL certificate expired, which rendered our instance unusable - that is, no content would load.
Why'd it happen?
We're not 100% sure. NGINX Proxy Manager is supposed to renew the certificate automatically, but it didn't. I'm suspecting this has something to do with how I've set it up. I'll need a bit of time to investigate further.
Why'd you take soo long to fix it?
Busy with IRL work, and didn't notice it. I haven't had time to jump onto Lemmy in the last 24 hours.
What steps are you taking to ensure this doesn't happen again?
I'll be adding the Lemmy instance to my UptimeKuma server to monitor the cert expiry, and give me notice before it expires. If I can't figure out why it's not renewing automatically, I'll renew the certificate before it expires again.
Music production. And IT in general.
But specifically the music production; started off as "I'll by FL Studio and muck around with it" to "I need ALL THE VSTs!". I've sunk like $2500 into it in the last two months (which is a hell of a lot of money to me), and I keep buying shit for it.
Am I any good at it? Fuck no. But it's not stopping me from keeping at it and buying shit I probably don't need :P
And the IT stuff consists of rack-mount servers and Pi's. I've sunk around $25k into it all over the last 12 years.
Like most instances, we've purged all images from the last 24 hours.
Unfortunately, some people have decided to post CSAM on remote instances. The moderators of the communities have been very proactive, but because of the federated nature of Lemmy, it means the material had federated.
So like most instances are doing, we've initiated a purge on the pictrs database/filesystem to delete ALL images from the last 24 hours.
Images may appear broken. I'd like to apologise for the inconvenience, but as Lodion said best on his post in [email protected] - "... this is pretty much the nightmare scenario for an instance. I’d rather nuke all images ever than to host such content."
It's pretty disgusting.
For fellow Lemmy admins who are curious, the script to delete is:
sudo find /srv/lemmy/example.com/volumes/pictrs/files -type f -ctime -1 -exec shred {} \;
Be sure to modify the path to point to your pictrs directory!
I hope that in my lifetime I can see ISPs regulated as a public utility.
From an Aussie where our Internet is somewhat considered a "public utility" (NBNCo), it's not the best. I'm paying $130/mo (Aussie bucks) for 250/100 fibre.
Our NTDs are capable of gigabit symmetrical, but thanks to our Lord and Saviour, Rupert Murdoch, it was essentially limited speed wise and the network was built with ridiculous complexity, such as the CVC constraints (Connectivity Virtual Circuit), which means ISPs have to buy additional bandwidth and hope and pray that every user doesn't max out their connections at the same time.
For example, the POI (Point of Interconnect) I'm connected to has a total of 1.5Gbps with the ISP I'm with. Based on their stats which they make public to customers, I'm guesstimating that there's approximately ~50 other households in my POI area connected with this ISP. We all have to share that bandwidth otherwise it slows to a crawl.
ETA: I'm purely talking about the FTTP network here, not the other part of the mess that is NBNCo and FTTN/C/B, Fixed Wireless, Satellite & HFC... the NBN is a complete mess.
We'll be updating to 0.18.4 in an hour!
As per the title, 0.18.4 was released a few days ago and we'll be updating to it in the next hour.
As usual, attached is an overview of the (new) server! Barely a scratch so far. Everything is running optimally.
Thanks for all your support! ❤️
And for context, this VM now has 4 cores (compared to 1), 8GB of RAM to play with (compared to 1GB), 120GB of storage (compared to 19GB) and is no longer contending with other VMs on the host (as I'm managing the host in colocation). RAM and storage is also easily expandable at the drop of a hat.
It should be a lot quicker!
We did have an issue with NPM refusing to start, which is what took the most time to get sorted and why we were offline for more than 5 minutes.
We're back, baby!
Kinda. There are a few small issues to still figure out, but for the most part we're back and should be federating.
I'll be working over the coming hours to finalise things and fix the niggly errors.
If you notice any errors, please message me or post here.
Cheers!
And we're ready for the migration! I sure hope this works 😂
Shutdown will begin at 16:00 AEST (in 13 minutes).
Currently preparing the new VM for the migration (firewall, installing docker, etc)
We'll be migrating to the new hosting system within 24 hours!
Hopefully you won't notice much downtime, because it's going to be a relatively simple migration. But if you do notice, apologies in advance!
I plan on doing the migration within the next 24 hours. Possibly sooner. Can't give an exact time at the moment, but check the comments. I'll post updates here.
Yeah it's an absolutely shitty predicament to be placed in.
My partner was lucky in the sense that she's been given basically guaranteed shifts... but she's off crook with the flu, which she caught from one of her clients. Which means she's going to be up shit creek this fortnight money wise.
perhaps having some sort of “earnings bank”
There already is this feature in place for CLink recipients. It's called "Working Credits". But, you can only earn 48 "credits" (being dollars) per fortnight, up to a maximum bank of 1000 on JobSeeker. Source
Their example of how it works:
Janine is single, has no children and doesn’t earn an income. Over time, Janine has built up 1,000 Working Credits as part of her JobSeeker payments.
Janine starts a full time job earning $1,600 per fortnight. In the first fortnight the 1,000 Working Credits reduce the amount we count as income from $1,600 to $600. This means Janine gets some JobSeeker Payment that fortnight. Janine’s Working Credit balance is zero.
The next fortnight all of Janine’s income will count. This reduces Janine’s JobSeeker Payment to zero.
My partner went through this recently. She got a casual job (which is great), and used up all of her working credit for the fortnight. I think the amount you can "bank" is a bit too small IMO. Along with the amount you can earn before you're cut off completely. With Cost of Living going through the roof, she's finding that she has to cut back on food and other essentials because rent is getting ridiculous, shopping for food is getting ridiculous and yeah. Plus the risk of it being casual work, there's no guarantee of shifts.
¯\_(ツ)_/¯
On a similar note, we will be migrating the Fediverse instances to a colocation server this coming week. This will give us the ability to rapidly expand the resources allocated to each instance without breaking the bank.
So, expect some downtime! I'll post another warning 24 hours before the migration.
We've just updated to 0.18.3
As per the title, we've just upgraded the instance to 0.18.3. Hopefully this'll fix the database taking up enormous amounts of space!
I've also attached some graphs of the instance's status for the nerds out there.
If you notice any issues, please let us know!
Thanks ❤️
Service Instability - 16JUL23
Apologies if you experienced any issues accessing the site over the last 30 minutes or so!
I was attempting to trim down a database table which has been a massive storage hog. The command being run caused a massive load on the server as it attempted to essentially truncate the table with set parameters (i.e, leave the most recent 100k records).
But, we're back. I have to fix the modlog again, but that's a small issue I really don't care about at the present time. I'll get to it eventually.
If you notice any issues, please let me know!
And as a bonus, this is how much space a single table was using (we're only on a 20GB drive right now)
We've just upgraded to 0.18.2
Like other instances, we've just updated our instance to 0.18.2 in order to help mitigate the risks of the XSS vulnerability.
If you notice any issues, please comment below!
Cheers!
I introduce: Tiger, basking in the sun
She has come so far, especially in the last three years. She's gone from being really aggressive, to incredibly placid and loving ❤️
G'day db0 (& Fediseer community)!
Firstly, love the idea of Fediseer! I hope to see it take off.
I'm a small Aussie instance looking for a guarantee. I've had a few small niggling issues with Lemmy (specifically, forgot to set the .well-known directory location in NPM 😖) but I've finally got it all working as intended.
Will of course be paying it forward/guaranteeing other instances!
Instance is lemmy.bleh.au
TIA!
Why we're committed to defederating from big-tech (such as Threads)
What is Threads?
If you've been living under a rock for the last few days, you mightn't have heard about Threads. Basically, Threads is another Twitter-style social media platform, heavily integrated with Instagram (Meta). From what I understand (and I'm happy to be corrected), Meta plans on integrating Threads to the ActivityPub protocol in the future. ActivityPub (colloquially known as Federation) is how services like Lemmy, Kbin, Mastodon, Pixelfed, etc, integrate and can show posts/images from a remote instance on the local instance. In fact, it's probably how most of you are reading this (as we're a small server with less than 100 users)!
Why is Threads a threat to the Fediverse?
The way I see it, is it is similar to a company buyout. The company that's purchasing the smaller company/competitor reassures the clientele of the smaller company that "Nothing will change. We want to embrace how this company operates and make the parent company like this one!". Seems harmless, right? Wrong. I've seen it happen before with companies in Australia (and globally). It never turns out this way. It pretty much always works out for the worst, and ends up in pissed off clientele and a husk of a great company (I'm looking at you, iiNet... thanks, TPG for ruining it).
While not the same to what I believe will happen to the Fediverse should Meta/big-tech be able to Federate, it's similar. Microsoft said it best themselves (as found by the U.S. DoJ); and this is what I believe we'll end up with if we allow federation with Meta/big-tech. EEE (Embrace the technology (so, federate with the existing network), Extend the technology by developing PROPRIETARY extensions to the ActivityPub protocol (so, for example, Threads has a new feature that won't be compatible or made available with the Open Source ActivityPub protocol), and finally Extinguish the Open Source ActivityPub protocol by marginalising existing Federation hosts because the new proprietary features they've added to ActivityPub are either exorbitantly expensive for small instances (like us) to use, or by simply locking the features up as "Trade Secrets" and never releasing them to the rest of the open network, thus killing all other instances because they have the largest share of users, and it has become the de-facto platform for everyone to use.
What is the stance of Bleh.au?
We have preemptively defederated from threads.net in the event they activate the ActivityPub protocol within Threads. Mainly for the reason above, but also because it has grown ridiculously fast overnight, and I don't believe the content on there at this present time is what I want people to see when they come to my Fediverse instances.
I want to be clear; it's not JUST Threads/Meta we should be concerned about. If, for example, Apple created a social media platform and it grew this ridiculously quickly overnight, I'd defederate it, too. Similar could be said with Twitter. If they decided to jump-on-in to the Fediverse, I'd defederate it as well, simply because Musk is nuts and his mentality towards technology mostly screams Embrace, Extend and Extinguish. And I don't like the content on Twitter.
Many other instances aren't speaking up about Threads or what they plan on doing regarding Federating with them. They're staying quiet. I'm not sure why, but if a small server like us can be vocal about it, why can't they?
That's my stance on it. It's not my only issue and sole reason why I'm defederating from Threads.net, but it's the main, major reason. I'm happy to hear your thoughts on it! Let me know what you think.
- iKill101
B365 build here (2020 build).
- i7 9700
- MSI B365M PRO-VDH
- RX 5700 XT 8GB
- 32GB DDR4 @ 2666MHz (Yeah, it's slowwww but it does what I need it to)
- 970 EVO Plus 500GB NVMe (boot drive, with TRIM disabled because otherwise it takes 5 minutes to boot)
- 860 QVO 1TB (Bulk Storage for music samples and work files, again with TRIM disabled due to boot times)
I'm going to miss Hackintoshing and (same as you) dual booting when they go full ARM-only. But I'll probably still rock this build for at least the next 3 years before upgrading to a Mac Studio or a Mac Mini.
macOS on a Hackintosh. It’s stable AF, and I love it. Use it for work and music production. Then I dual boot Windows 10 for games.
Very nice! I'm always impressed by the graphs and stats of larger instances and how things like storage and bandwidth are managed. Thanks for sharing, Lodion!
OTT by Xfer Records
Link to Splice’s page about it, because Xfer’s site is hot garbage. - https://splice.com/plugins/3788-ott-vst-au-by-xfer-records