The red means it’s not cooked all the way. Better be safe…
The red means it’s not cooked all the way. Better be safe…
I like my Kubernetes setup at work. It runs Nextcloud, Mattermost, GitLab, company website, several embedded firmware OTA update sites, a few internal apps. Nextcloud was pretty easy to install on it with Helm, just a single command line and a yaml file to specify domain, settings, etc. I had some teething issues in my early setup where the database would get wiped inexplicably, but it’s been running smooth for years now. (Yes, I know, bad juju running databases on Kubernetes…I’m used to it and it mostly works)
Well, then, they ought to drop enough dirt for me to patch it. You think I carry stacks of dirt around with me all the time?
A little slower by today’s standards, but if your needs are light, it’ll do the job. Keep in mind it only has a gigglebyte of RAM, so its capacity for running things may be limited, especially as docker applications go (since they bring a copy of each dependency). You won’t be able to run something as large as GitLab or Nextcloud, but a smattering of small apps should be within its capabilities
The thing with using the “latest” tag is you might get lucky and nothing bad happens (the apps are pretty stable, fault tolerant, and/or backward compatible), but you also might get unlucky and a container update does break something (think a 1.x going to 2.x one day). Without pinning the container to a specific version, you might have an outage suddenly due to that container becoming incompatible with one of your other applications. I’ve seen this happen a number of times. One example is a frontend (UI) container that updates to no longer be compatible with older versions of the backend and crashes as a result.
If all your apps are pretty much standalone and you trust them to update properly every time a new version of the container is downloaded, then you may never run into the problems that make people say “never use latest”. But just keep an eye out for something like that to happen at some point. You’ll save yourself some time if you have records of what versions are running when everything’s working, and take regular backups of all their data.
I had to close my bank account to cancel mine. I moved and didn’t want to head all the way back to go in person to cancel. They wouldn’t accept a cancel request online or over the phone. Why is it always gym memberships that want to be next to impossible to cancel?
The problem child for me right now is a game built in node.js that I’m trying to host/fix. It’s lagging at random with very little reason, crashing in new and interesting ways every day, and resisting almost all attempts at instrumentation & debugging. To the point most things in DevTools just lock it up full stop. And it’s not compatible with most APMs because most of the traffic occurs over websockets. (I had Datadog working, but all it was saying was most of the CPU time is being spent on garbage collection at the time things go wonky–couldn’t get it narrowed down, and I’ve tried many different GC settings that ultimately didn’t help)
I haven’t had any major problems with Nextcloud lately, despite the fragile way in which I’ve installed it at work (Nextcloud and MariaDB both in Kubernetes). It occasionally gets stuck in maintenance mode after an update, because I’m not giving it enough time to run the update and it restarts the container and I haven’t given enough thought to what it’d take to increase that time. That’s about it. Early on I did have a little trouble maintaining it because of some problems with the storage, or the database container deciding to start over and wipe the volume, but nothing my backups couldn’t handle.
I have a hell of a time getting the email to stay working, but that’s not necessarily a Nextcloud problem, that’s a Microsoft being weird about email problem (according to them it is time to let go of ancient apps that cannot handle oauth2–Nextcloud emailer doesn’t support this, same with several other applications we’re running, so we have to do some weird email proxy stuff)
I am not surprised to hear some of the stories in this thread, though. Nextcloud’s doing a lot of stuff. Lots of failure points.
Even worse is taking over the back button / back gesture to redirect you to “more to read before you go”
Pressing F to pay respects. R.I.P. in pieces
Depending on how mission critical your data is…Set up delayed replicas and backups (and test that your backups can actually be restored from). Get a second pair of eyeballs on your query. Set up test environments and run it there before running it in production. The more automated testing you put into your pipeline, the better. Every edit should be committed and tested. (Kubernetes and GitLab Auto DevOps makes this kind of thing a cinch, every branch has a new test environment set up automatically)
Don't beat yourself up too much though. It happens even to seasoned pros.
I just want a picture of a goddang hotdog
I lol'd out loud irl
16 on the machine I use the most at work. (MacBook Air M1)
Actually, 128 gigglebytes on my home PC, though. I upgraded so I could play pretty Minecraft.
ITT: some people are mad the web became the application platform of choice, in part due to handy dandy cross platform app tools like Electron and accessible languages like JavaScript.
There is no perfect answer. Qt isn’t using the platform’s native capabilities to the fullest extent either. Qt requires a “wrapper” too–all those libraries your app depends on, to name a few (unless you got a commercial license and are compiling statically, you rich devil).
Let’s celebrate the onslaught of apps that work with Linux instead of trying to scare off developers any more than Linux already did. Make love not war. <3
In my experience, Electron and other “web wrapper” apps run just fine and I have enough CPU and RAM to run a dozen of them alongside my 50 browser tabs. Slack, Discord, VSCode, Teams, IRCCloud, it all works fine. Hardware is cheap compared to my time.
I’m sad it doesn’t work with Quake 2 RTX, I still find that to be prettier.
Teach me your ways
I get mad sometimes
It’s lighter than a VM but a bit heavier than aiming to run an application natively (and all the dependency & configuration hell that entails).
Basically a convenient way to package and run applications with all their dependencies, without regard for what libraries & configurations exist in the host OS and other containers.
If your application only works with up to version 42 of the Whatchamacallit library, you ship it with that version of Whatchamacallit, the underlying OS doesn’t need to install it. Other containers running on the system that depend on that library don’t get broken since they’re packaged with version 69 which works fine for them.
Meme answer:
This is my little setup at work
Kubernetes cluster (created by kubespray)
I really should check my notifications more often.
I can’t really say from experience, as the only game I’ve played on Linux that isn’t Linux native is Starcraft, and that’s not exactly a demanding game these days. Linux is on my main workstation, but I still have a Windows PC as my main gaming PC, so I haven’t had a reason to check out how well the virtualization & compatibility stuff works.
My intuition says you’ve still got enough CPU oomph to muscle through any such virtualization overhead though.
Yes. I literally had to close a bank account to get Planet Fitness to stop charging me.