Somehow I miss those days. Now you need weeks of training to understand the black magic behind all the build/deployment stuff in whatever cloud provider your company decided to use…
Naa, once you figure out one the rest click usually.
We got our own platform based on kubernetes and cncf stuff and we don’t have to care anymore about the metal underneath. AWS? OTC? Azure? Thats just a target parameter, platform does the rest. It’s great.
How often do you switch cloud providers that this is even a real rather than a hypothetical benefit? (Compared to the cost of dealing with a much more complicated stack.)
It’s not about switching, it’s about hosting our services on different platforms at the same time.
I manage a stack like this, we have dedicated hardware running a steady state of backend processing, but scale into AWS if there’s a surge in realtime processing needed and we don’t have the hardware. We also had an outage in our on prem datacenter once which was expensive for us (I assume an insurance claim was made), but scaling to AWS was almost automatic, and the impact was minimal for a full datacenter outage.
If we wanted to optimize even more, I’m sure we could scale into Azure depending on server costs when spot pricing is higher in AWS. The moral of the story is to not get too locked into any one provider and utilize some of the abstraction layers so that AWS, Azure, etc are just targets that you can shop around for by default, without having to scramble.
Oh please, you didn’t even have to turn the cassette or floppy disc over. You and your luxuries.
Some of us still do 🙃
If it ain’t broke, don’t fix it.
One could argue the requirements have changed because the security and compliance part of the world finally caught up to modern software delivery concepts. Even the most dinosaur apps at compliant orgs are being dragged kicking and screaming into new CI/CD tools where applying governance and custody chains and permissions and approvals are all self documented automated hooks.
Anybody that actually professionally deals with this kind of thing understands just how wrong you are.
Ah yes, war files
Don’t forget ear files. Oh, and don’t forget the abomination that is the executable war file when you’re using Spring Boot but your company hasn’t fully embraced it yet.
This is how I deployed an app less than 5 years ago (healthcare).
It’s sad
I know a place where they still do this. They’ve got an 8-digit user count, 7 digit monthly profits, all running on one server that costs something like $20 a month. They’ve downsized a few years ago to single-digit employee number and just sit there and collect profits. And this is why I’m now working for a company that casually dropped a few grand for a glorified CPU usage meter and a few grand on top of that for deployment tool that does the same thing that the old guy at a former place was doing with his trusty FTP client.
This is how I deploy my personal website today. The holster doesn’t give ash access.
I remember this. I also remember using
scpinstead. Andftp, if I go back far enough.rsyncis still my friend thoughzfshas mostly replaced it now.How has zfs replaced rsync for you? One is a filesystem, and the other is a filesyncing tool. Does zfs do something im not aware of lol?
I used to use
rsyncto copy data from my storage array on one machine to an external and an off site backup. Since a lot of it was code, it always took forever to scan all the small files, and I had to script unlocking remote partitions.With encrypted ZFS, I can just
zfs snapthenzfs send, and it does the same thing at the block level, raw, so way faster, less data transfer, and no need to send a key or passphrase unless I need to mount it at the destination (meaning a cloud provider could never know the data, for instance).ZFS is also recursive, so if I have
s/storageand/storage/stuffdefined, I can snap and send either level, which makes it as versatile as rsync.Oh interesting, i am not super familar with zfs’ tools, so thats pretty cool! Ill have to look at that for my storage array.
This is how I deployed web servers in school like 3 years ago.
My school had nothing about react, mode, angular, angularJS, SaaS, etc. back in 2015.
We learned Perl, PHP, LAMP stack, SOAP based APIs and other “antiquated” things. Provided a solid foundation of fundamentals that I’ve built a nice career on.
It might have been by design to get a feel for the fundamentals. Or maybe it’s just because the people teaching it have probably left the industry and are teaching how they did it.
My department head was in his 70s and my professors all trended on the older side.
Yeah it’s not all that uncommon in school, just increasingly uncommon in industry.
I mean, a lot of docker files out there with
COPY . .True, but building the image is not the same as deploying to production.
Fair point
A lot are still doing that and haven’t moved up
(Please at least use SFTP!)
FTP Explorer all the way! Preferred that to filezilla… I mean it didn’t support sftp but I liked it.
Cries in REST API
People don’t use FileZilla for server management anymore? I feel like I’ve missed that memo.
They have bundled malware from the main downloads on their own site multiple times over the years, and even denied it and tried gaslighting people that AVs were giving false positives because AV companies are paid off by other corporations. And the admin will even try to delete the threads about this stuff but web archive to the rescue…
You know what? I didn’t believe you, since I’m using it for a long time on Linux and never had any issues with it. Today, when I helped a friend (on Windows) with some SFTP transfer and recommended FileZilla was the first time I realised the official Downloads page provides Adware. The executable even gets flagged by Microsoft Defender and VirusTotal. That’s actually REALLY bad. Isn’t FileZilla operated by Mozilla? Should I stop using it, even though the Linux versions don’t have sketchy stuff? It definitely leaves a really bad taste.
Yeah, it’s bad. Surprised they’re still serving that crap in their own bundle but i guess some things don’t change.
Filezilla is no relation to mozilla. But yeah i moved away from it years ago. The general recommendation I’ve seen is “anything but filezilla”. Personally i use winscp for windows, and will have to figure out what to use when i switch my daily driver to Linux.
I suppose in the days of ‘Cloud Hosting’ a lot of people (hopefully) don’t just randomly upload new files (manually) on a server anymore.
Even if you still just use normal servers that behave like this, a better practice would be to have a build server that creates builds, like whenever you check code into the Main branch, it’ll create a deploy for the server, and you deploy it from there - instead of compiling locally, opening filezilla and doing an upload.
If you’re using ‘Cloud Hosting’ - for example AWS - If you use VMs or bare metal - you’d maybe create Elastic Beanstalk images and upload a new Application or Machine Image as a new version, and deploy that in a more managed way. Or if you’re using Docker, you just upload a new Docker image into a Docker registry and deploy those.
For some of my sites, I still build on my PC and rsync the build directory across. I’ve been meaning to set up Gitlab or something similar and configure automated deployments.
This is what I do because my sites aren’t complicated enough to warrant a build system. Personally I think most websites out there are over-engineered. Example: a Discord friend made a React site that displays stats from a gaming server. It looks nice, but you literally can’t hyperlink to any of the data, it can only be loaded dynamically and only looks coherent on a phone in portrait mode. There are a lot of people following trends (some good trends) but without really thinking about why.
I’m starting to like the htmx model a lot. Server-rendered app that uses HTML attributes to configure the dynamic bits (e.g. which URL to hit and which DOM element to insert the response into). Don’t have to write much JS (or any in some cases).
you literally can’t hyperlink to any of the data
I thought most React-powered frameworks use a URL router out-of-the-box these days? The developer does need to have a rough idea what they’re doing, though.
I remember joining the industry and switching our company over to full Continuous Integration and Deployment. Instead of uploading DLL’s directly to prod via FTP, we could verify each build, deploy to each environment, run some service tests to see if pages were loading, all the way up to prod - with rollback. I showed my manager, and he shrugged. He didn’t see the benefit of this happening when, in his eyes, all he needed to do was drag and drop, and load the page to make sure all is fine.
Unsurprisingly, I found out that this is how he builds websites to this day…
This is from before my times, but… Deploying an app by uploading a pre built bundle? If it’s a fully self-contained package, that seems good to me, perhaps better than many websites today…
That’s one nice thing about Java. You can bundle the entire app in one .jar or .war file (a .war is essentially the same as a .jar but it’s designed to run within a Servlet container like Tomcat).
PHP also became popular in the PHP 4.x era because it had a large standard library (you could easily create a PHP site with no third-party libraries), and deployment was simply copying the files to the server. No build step needed. Classic ASP was popular before it, and also had no build step. but it had a very small standard library and relied heavily on COM components which had to be manually installed on the server.
PHP is mostly the same today, but these days it’s JIT compiled so it’s faster than the PHP of the past, which was interpreted.
okay, but why did you use a password when the ssh/sftp key is right next to the files
