Chatting on Slack with a junior dev:

[Junior Dev] How do I get that file from the server ?
[Me] ssh into it and then use scp

I see Junior across the room, literally saying "shshh..." at the computer.

Packed my stuff and quit that day.

Fucking 20 hour days. Third one this week.

Been at work since 6am, it is now midnight. Spent the morning fixing bush league code mistakes from "expert" onshore developers, and explaining how-to-wipe-your-ass level concepts to some rude cunt who is absolutely going to take credit for my work after I leave.

Now I'm just waiting on this slow boat scp to finish because the invalids the customer hired to manage their infra can't figure out the 3 minute exercise that is standing up a registry, so the container deployment process is fucking export multiple 500mb Redhat images as a tar and ship it across the cripplenet they call a datacenter. And of course the same badmins don't understand rsync and can't manage to get network throughput in a datacenter with a $300M annual budget over 128kbps. I guess that's fast for whatever jugaad horseshit network they're used to.

I've said it before, but it bears repeating. Fuck IBM. They're a cancer and at this point I question the moral compass of anyone who works for them.

Often saying fuck it when I've got a project and I kind of really need to use git but I'm not good with it so I just go back to good old scp again 😅

Still not learning git properly. I still find it very hard to grasp how this works and although I'm starting to use it more and more, I just can't wrap my hear around it so I often reside to scp for upload and no version control at all.

:'(

I just can't understand what will lead an so called Software Company, that provides for my local government by the way, to use an cloud sever (AWS ec2 instance) like it were an bare metal machine.

They have it working, non-stop, for over 4 years or so. Just one instance. Running MySQL, PostgreSQL, Apache, PHP and an f* Tomcat server with no less than 10 HUGE apps deployed. I just can't believe this instance is still up.

By the way, they don't do backups, most of the data is on the ephemeral storage, they use just one private key for every dev, no CI, no testing. Deployment are nightmares using scp to upload the .war...

But still, they are running several several apps for things like registering citizen complaints that comes in by hot lines. The system is incredibly slow as they use just hibernate without query optimizations to lookup and search things (n+1 query problems).

They didn't even bother to get a proper domain. They use an IP address and expose the port for tomcat directly. No reverse proxy here! (No ssl too)

I've been out of this company for two years now, it was my first work as a developer, but they needed help for an app that I worked on during my time there. I was really surprised to see that everything still the same. Even the old private key that they emailed me (?!?!?!?!) back then still worked. All the passwords still the same too.

I have some good rants from the time I was there, and about the general level of the developers in my region. But I'll leave them for later!

Is it just me or this whole shit is crazy af?

Uploading something to a server through SCP and copying the composer libs along instead of installing them server side.

Please don't kill me 😅

My coworker doesn't know how to use a terminal. He talked himself into his position and instead of taking the time to learn about the basic commands he keeps asking someone else (including the teammanager, who's actually a software engineer) to do things for him.
For reference; we need the terminal to tail log files, keep track of processes, cron jobs, manipulate file structures, use scp (I use sshfs) to move things between other workstations and servers etc. Being able to use a terminal is one of the basic requirements for our job.

What.
Why.
How.

Why do people do this?

So a coworker made a backup script that asynchronously copies the backup to the remote backup space and then deletes the backup:

scp backup.tgz remote: &
rm backup.tgz

Spot the bullshit.

I just love how liferay keeps finding ways to surprise me...

Customer: need to fix this security issue
me: ok
me: fixed. Testing locally. Works 100%.
Me: testing on dev server. Works 100%
qa: testing on dev server. Works 100%
me: all good. Deploying to preprod
customer: it doesn't work
me: testing preprod - it doesn't work.
Me: scp whole app to local machine. App works 100%.
Me: preview loaded liferay properties in preprod via liferay adm panel. All props loaded ok.
Me: attach jdb to preprod's liferay to see what props are loaded. Only defaults are used [custom props not loaded according to jdb]

me: is there some quantum mechanics involved..? Liferay managed to both load and not to load props at the same time and the state only changes as it's observed...

So im making an app that tracks you real time and sends spooky pictures of your frequented locations (scp 1471, if anyone is familiar with scps), and totally forgot about that it self triggers daily, so when i wanted to check the time, i nearly shat myself. The google logo makes it infinitely more scary.

Fuck you scp. I was uploading 6GB file to an EC2 server. Well, needless to say, "no space left on device" after all 6GB transferred was the biggest FUCK YOU moment. Seriously. Send the file size and check before you waste 30 minutes of my time. Oh, and don't read CLI command data as part of the transfer. You suck.

Python haters, gather 'round

oh come on... In java it's all simple as 123. You build an app, you have like 200 dependencies, you pack it all in a single fat jar and only deploy that single .jar. Don't need no internet, no installs, no pip, no nothing: just your .jar file and the JVM.

So java:
- build an app
- use 200+ deps
- build your whole project into a single fat jar
- deploy your jar in the env
- install [*khem khem* scp into the server] jre
- run

Now let's look at py, shall we?
- build an app
- use 1 dep
- deploy all the 20 .py files in the env
- make sure you have internet access
- install python
- install pip
- pip install <my dependency>
- run your app

Here at the client site everyone uses Windows 7, since this is not an IT firm. They make jewellery. So I don't blame them.
The problem is their in house dev team are also forced to use Windows.

Today someone from their dev team was with me for a new project. Their senior guy sent us a mail mentioning that the project code is on AWS EC2 instance and we will have to SSH to see it.
I checked the code on my MacBook with SSH and copied it to local using SCP. The dev guy was seeing it in amazement. I asked, what's it?

He just asked, "You don't have to use PUTTY" 😮

I smirked 😏

I wanted to install this new software in our cluster but it didn't have Cmake on it, so I scp Cmake source files and then I started looking for a CMakeLists.txt file for compile it when I suddenly realized what I wanted to accomplish in the first place 😅

After working with a coworker on some odd issues, I finally decided to check on the actual ticket he needed assistance with.

From now on, we will optimize our HTML for aesthetic appeal in Chrome's dev tools. display:none is verboten.

Sometimes I wonder if I've had a stroke or if I've died and am in purgatory.

Started a new job as a dev. First days revealed no local admin rights, no right to use Linux locally and a very limited set of Software. Negotiated compromise to get a remote VM with Linux and a user who is part of sudo. VM turned out to be isolated by proxy, so I can not install anything new. At least Docker is pre-installed and I hoped it could work out. But guess what no access to dockerhub and I can not pull any images. Admin told me to copy manually the images with scp.

I'd never thought that there could be any companies out there who treats devs like that. What puzzles me most, there're lot of devs staying with that company for years, even decades already and they're good guys, please don't get me wrong.

Did you encounter anything like that? Could you make any difference there, where you met anything like it.

I reached the point after 3 weeks where I do not think I can make any difference and when it'll take ages to move people and company policy.
I do not want to give up, but I fear it is pointless to fight for change there. I am out of options and about to leave asap. Can you recommend me anything else?

Thanks in advance and for your time :)
Felt good to write it down.

And this happened today:

While on ssh session, this guy restarted network manager. According to him "his Facebook page wasn't showing him recent feeds".

DO NOT EXPORT GPG KEYS _TEMPORARILY_ AND ASSUME THAT THEY'LL BE IN THE ORIGINAL LOCATION AFTER EXPORT!

I learnt this lesson the hard way.

I had to use a GPG key from my personal keyring on a different machine ( that I control ). This was a temporary one-time operation so I thought I might be a smart-ass and do the decryption on the fly.

So, the idiotic me directly piped the output : `gpg --export-secret-key | scp ...`. Very cool ( at the time ). Everything worked as expected. I was happy. I went to bed.

In the morning, I had to use the same key on the original machine for the normal purpose I'd use it for and guess what greeted me? - *No secret key*

*me exclaims* : What the actual f**k?!

More than half a day of researching on the internet and various trials-and-errors ( I didn't even do any work for my employer ), I finally gave up trying to retrieve / recover the lost secret key that was never written to a file.

Well, to be fair, it was imported into a temporary keyring on the second machine, but that was deleted immediately after use. Because I *thought* that the original secret key was still in my original keyring.

More idiotic was the fact that I'd been completely ignorant of the option called `--list-secret-keys` even after using GPG for many years now. My test to confirm whether the key was still in place was `--list-keys` which even now lists the user ID. Alas, now without a secret key to do anything meaningful really.

Here I am, with my face in my hands, shaking my head and almost crying.

Changed the .bashrc of my headless server incorrectly, causing me to get kicked out of my ssh session as soon as I connected… Thank God for scp (learned something new today as well 😂)

Fuck you Redis

Goes in a docker container, calls bgrewriteaof, get success, checks info - no pending writes, last write success.

Tries to scp to remote, fails - Unexpected AOF

Decides to shut down the local redis to be able to port, in case it's blocking it

Calls redis-cli shudown (expect to just shut redis down rigth)

It fucking deleted all my data, now I see the docs

"Flush the Append Only File if AOF is enabled."

Why the fuck? Fuck you redis, fuck you

Fuck ssh. It does 4 things at once and i couldn't get it to do one. I have some pi's and want a shared directory on each of them. On a server i created a user for that and mounted its home directory on a pi, it worked. I did some lockdowns (no shell, only sftp allowed, login only via keyfile), but i was still able to mount it on boot.

Now i had to migrate this setup to another server. It took me a while copying all the configuration etc. All i got for that was a error-message. I figured out the users home-directory had to be owned be root, fixed that, got another error message. Somehow scp didn't use sftp but the login shell which is /usr/sbin/nologin. That made scp (and sshfs) fail, even though it perfectly works with the other server.

I gave up and removed all the setup. I'll find another distributed filesystem for that (but not samba or nfs, those are way to complicated). Those are the setbacks that depress me.

Wanted to use my ps4 controller on my pc to play a game. There's a program called scp toolkit to install appropriate device drivers for those controllers. The program provides a list of available usb devices currently connected, for which to install the drivers. Unfortunately the devices listed by scp have slightly cryptic names, which means you have to find out, which one it is you're looking for. I chose a usb device for installation I was convinced was the controller. Little did I know that the following NullPointer error made my keyboard unusable. 🙃

Just tried scp to transfer data directly from aws to my local machine, cool shit

The end of today was hell...

Another company make changes to UAT for one part of the application and refuse to use VCS. This means we have to manually merge there changes in regularly. The developer in charge of this project left a few weeks ago and when someone new to the setup decided to pickup a task to re merge they accidentally did the scp command the wrong way around...

I probably don't need to say much more but we received MANY angry email and calls.

So i'm trying to upload a file to an SSH server using node. First I try the obvious putFile method provided by the obvious node-ssh package. On any other server this would work fine but this server doesn't have sftp installed so that doesn't work.

OK, so next I learn how scp works (it runs the command "scp -t" on the remote server, and sends to stdin a command like "C0666 1234 file.txt" and then sends the contents of the file) and I write some javascript code to do this. It's pretty finicky, the first few tries I forget to close the stream right or detect the program finishing. I add some logging and that helps me figure out what the problem is, and finally I get it to not output any errors.

So I log into the server and check and the file isn't there. I try again several times, file still isn't there. I try running scp -t manually on the server, typing in exactly what my program is sending, and it works. This goes on for a while until I realize that I've been sending a file to one server and logging into a different server to check if the file was sent. grrr

It's always a matter of much is there to do and in what language...
There is the IDE-Zone, which is dominated by IntelliJ (CLion be praised when you do Rust or C++) for large stuff and heavy refactorings.

Always disputted by VS Code with synced settings. It's nice and comfy and has every imaginable language supported good enough, especially when its smaller change in native code or web/scripting stuff.

Then there is the "small changes" space, where Vim and VS Code struggle whos faster or which way sticks better in my brain...

might be you SCP stuff down from a box and edit it to re-upload, or you use the ever-present vi (no "m" unfortunately)

sometimes things are more easy for multi-caret editing (Ctrl-D or Alt-J), and sometimes you just want to ":%s/foo/bar/g" in vim.
I am sure that each of these things are perfectly possible in each of the editors, but there is just reflexes in my editor choices.

I try to stay flexible and discover strenghts of each one of my weapon of choice and did change the favorites. (Atom, Brackets, Eclipse, Netbeans, ...)

However there are some things I tried often and they are simply not working for me...

might for you. I don't care. and I'll just use some space to piss people off, because this is supposed to be a rant:

nano just feels wrong, emacs is pestilence from satan that was meant for tentacles instead of fingers, sublime does cost money but should not, gives me a constant guilty feeling (and I don't like that) that, and all the editors from various desktop environments are wasted developer ressources.

On Linux I want to scp a file from Machine B to my machine A.

Assuming B allows all connections is any key/cert file exchange needed?

I recall whenever I SSH to a new machine I need to accept a key on the first connect.

Guess general question is what is the connection process between 2 Linux machines for sftp purposes?