GNU / Linux May 2015 Desktops


Recommended Posts

GNU / Linux May 2015 Desktops - Posting Guidelines

- Embedded images should be smaller than or equal to 1440x900 (preferably 1024x768).
- The file size for image previews should be no larger than 750 KB.
- Images linked externally should be no larger than 3 MB.

- If you plan to img-embed your desktop, please try to keep it under 100k (It's recommended that you link it, externally).
- Stay within the board rules, no matter how yummy you may believe it to be.
- If you are going to quote, there isn't a need to quote an img-embeded picture - Just don't do it. Please just reference the post by page number/page link/etc.
The desktops thread is not to be used as a means to post images of a pornographic or otherwise provocative nature. We do not support a babes forum and neither will we in an unofficial capacity. Please help to keep the thread relevant and safe to browse for everyone.
- The moderators of this forum reserve the right to edit or remove your posts as necessary so they may comply with our Community Rules and so that this thread remains on topic.

- In the interests of keeping the thread on topic it is strongly urged that you format your posts in the following order:

Themes: Name/where you got it (with link would be nice).
Wallpaper: Name/where you got it (with link would be nice).
Icons: Name/where you got it (with link would be nice).
Programs: Name/where you got it (with link would be nice).

Linux System Stats Script - Produced by our own Neowin members, Haggis and Karl L (+xorangekiller)!
For finding mystery wallpapers, consider using Google Images Reverse Search or Tineye.

Link to comment
Share on other sites

I found this on deviantart. I know it's Arch and I also recognize Plank, but can anyone tell me what the rest of the stuff is? I mean the icon theme, the overall theme, the weather thingy, the music notification, etc. I asked the uploader about this, but I didn't get any reply. :(

 

arch_linux__cinnamon__plank_by_igarunya8

Link to comment
Share on other sites

  • 2 weeks later...

I found this on deviantart. I know it's Arch and I also recognize Plank, but can anyone tell me what the rest of the stuff is? I mean the icon theme, the overall theme, the weather thingy, the music notification, etc. I asked the uploader about this, but I didn't get any reply. :(

 

arch_linux__cinnamon__plank_by_igarunya8

The music widget might be CoverGloobus. The weather thing again looks like a widget. Although it could be conky I suppose. Does Cinnamon come with widgets?
Link to comment
Share on other sites

The music widget might be CoverGloobus. The weather thing again looks like a widget. Although it could be conky I suppose. Does Cinnamon come with widgets?

 

It looks like Plank and Cinnamon, and yes, users can install Widgets/Desklets directly from Cinnamon Settings.

Link to comment
Share on other sites

It looks like Plank and Cinnamon, and yes, users can install Widgets/Desklets directly from Cinnamon Settings.

Oh cool, thanks for the explanation.
Link to comment
Share on other sites

  • 2 weeks later...

post-67054-0-95783800-1432866450.png

 

Technically this isn't a screenshot of a Linux desktop. It's not even a Linux machine, as you can guess from my hostname. But indulge me. I'm working with Linux and this setup would have been exactly the same on a Linux machine.

 

What's going on?

That's a fullscreen terminal running tmux with 3 horizontal split panes. The first pane has a Vim session with 3 vertical split windows, and a NERD tree browser on the left. This is on a 13" laptop screen, and to maximise screen real estate, I've got Vim set up so that the current active window will automatically expand and the others shrink as I navigate back and forth. The middle tmux pane shows real time logs from 4 linux nodes, and the bottom pane is just a bash session.

 

I'm working on an automated setup for provisioning a Hadoop cluster and launching applications on it. The endgame would be to enter one command into the terminal, and have a cluster up and running locally in a virtual machine or in the cloud. The cluster in that screenshot is running locally on the laptop, in a VM with 9GB of RAM. Each node is run in a Docker container.

Link to comment
Share on other sites

attachicon.gifHadoop Cluster on Docker.png

 

Technically this isn't a screenshot of a Linux desktop. It's not even a Linux machine, as you can guess from my hostname. But indulge me. I'm working with Linux and this setup would have been exactly the same on a Linux machine.

 

What's going on?

That's a fullscreen terminal running tmux with 3 horizontal split panes. The first pane has a Vim session with 3 vertical split windows, and a NERD tree browser on the left. This is on a 13" laptop screen, and to maximise screen real estate, I've got Vim set up so that the current active window will automatically expand and the others shrink as I navigate back and forth. The middle tmux pane shows real time logs from 4 linux nodes, and the bottom pane is just a bash session.

 

I'm working on an automated setup for provisioning a Hadoop cluster and launching applications on it. The endgame would be to enter one command into the terminal, and have a cluster up and running locally in a virtual machine or in the cloud. The cluster in that screenshot is running locally on the laptop, in a VM with 9GB of RAM. Each node is run in a Docker container.

 

i can follow you on the layout which looks good but i am clueless otherwise. :/

Link to comment
Share on other sites

Hey, that's alright. You're running those nodes from a 12.04 Ubuntu VM, so it counts.

 

I would, however, advise that you clean up the more sensitive information such as hostnames, ports, and such in your screenie unless you have already changed those. Any Sysadmin worth their salt would never expose that information publicly. Using Dropbox as shown above is a unique approach to say the least but I advise strong, proactive security practices to protect your setup. I'm sure you've already recycled everything by now, so there's no need for a long diatribe about it. :yes:

 

(Assuming, of course, you're only using Dropbox to store extra configuration files for future reference and not dynamic configuration for loading each time -- which nobody should ever do, hehe!)

 

Nicely laid out and thought out interface. I'm sure it serves your needs well.

Link to comment
Share on other sites

Hey, that's alright. You're running those nodes from a 12.04 Ubuntu VM, so it counts.

 

 

Close :) Each node is running Ubuntu 12.04 in a Docker container. The VM is running boot2docker.

 

I would, however, advise that you clean up the more sensitive information such as hostnames, ports, and such in your screenie unless you have already changed those. Any Sysadmin worth their salt would never expose that information publicly. Using Dropbox as shown above is a unique approach to say the least but I advise strong, proactive security practices to protect your setup. I'm sure you've already recycled everything by now, so there's no need for a long diatribe about it.  :yes:

 

 

 
Yes, this is not a production setup. It's a proof of concept at this stage.
 
Having said that, only linked containers can talk to each other over the ports defined in a link, and the hostnames only work if you are a linked host. Nodes that are not participating in the setup can't do anything meaningful with the information.
 

 

(Assuming, of course, you're only using Dropbox to store extra configuration files for future reference and not dynamic configuration for loading each time -- which nobody should ever do, hehe!)

 

That path is just where I've checked out the project. The canonical store for the configuration is a Git repository.

 

 

 

Nicely laid out and thought out interface. I'm sure it serves your needs well.

 

Thanks :) tmux is almost like a tiling window manager for the terminal. I've got it set up so that I can jump from Vim splits to tmux splits using the same shortcut keys. It's pretty seamless.

  • Like 1
Link to comment
Share on other sites

vhane, it's a carefully crafted work of technical beauty. I'm pleased that a proof-of-concept and testing implementation like this is being worked on. There have been a few "Full Session Containers" in the past that didn't quite make the grade.

 

It's got some real potential, no doubt about it.

 

The way you have it set up is very, very elegant without being overtly complicated or outright gnarly. And that UI is lovely, and (I assume) customizable to some degree. Any trouble with information overload? I'm fairly interested in your experience with several workloads going at once on those containers -- any performance drop-outs? That was a problem with the other Full Session Containers.

 

If this is simply for your own use, of course you may feel perfectly free to tell me to kindly bugger off. :laugh:  I'm just interested in this sort of thing for future reference purposes for a cluster of my own that I'm eyeballing.

Link to comment
Share on other sites

vhane, it's a carefully crafted work of technical beauty. I'm pleased that a proof-of-concept and testing implementation like this is being worked on. There have been a few "Full Session Containers" in the past that didn't quite make the grade.

 

It's got some real potential, no doubt about it.

 

The way you have it set up is very, very elegant without being overtly complicated or outright gnarly. And that UI is lovely, and (I assume) customizable to some degree. Any trouble with information overload? I'm fairly interested in your experience with several workloads going at once on those containers -- any performance drop-outs? That was a problem with the other Full Session Containers.

 

Thank you for the kind words. I'm building on the shoulders of giants (Docker, Ambari, Hadoop), so I'm not sure I deserve this kind of credit. Pardon my ignorance, but what are full session containers? I'm not really an ops guy. I'm more of a developer, so I may not be familiar with the parlance. Are you referring to the fact that the cluster runs on a single VM? My development setup does, because running a single VM locally requires less resources. In production the containers can be spread out to run on different physical machines. This is currently not quite as simple and seamless as the proof of concept, but will be in the future, once Docker Swarm is able to link containers across different Docker hosts.

 
As for actually running workloads, I haven't gotten that far yet. I've got to the point where I have automated the creation of a Hadoop cluster. The next step is to setup of the applications that I want to run on it. In the short term this would be the GeoMesa database, which runs on top of the Accumulo database. In the long term there are other technologies in the Hadoop ecosystem that I want to use, e.g. Kafka and Apache Samza for realtime stream processing. In order to automate the setup of a Hadoop infrastructure, I'm looking at Ambari Blueprints. So, I have my Hadoop cluster up and running, including a container for the Ambari Shell, and I run the blueprints against the shell. If everything goes well, at the end of this process I have the Hadoop applications running maybe half an hour after typing one single "docker-compose up" command. That's the end game.
 
As for the UI, it's not part of the project. It's just how my screen looked like when I took the screenshot ;) It's how I to work in the console: Tmux with various tabs and splits, Vim with various tabs and splits. The magic sauce is being able to move (Ctrl + h, j, k or l) between Vim and Tmux panes as if they were all part of one single IDE.

 

If this is simply for your own use, of course you may feel perfectly free to tell me to kindly bugger off.  :laugh:  I'm just interested in this sort of thing for future reference purposes for a cluster of my own that I'm eyeballing.

 

I'll probably put this up on GitHub once I'm happy with the setup. Obviously I won't publish parts that are specific to my project, but a generic recipe to quickly launch a Hadoop cluster could be useful to other people. There are a few out there already, but they are either tailored for a single-node cluster (not very useful for production), harder to understand, or outdated.

 

I'm aiming for something like:

  1. Configure the number of nodes in the cluster
  2. Add your Ambari blueprint (or don't, if you'd rather use the Ambari admin UI that launches on the Ambari server node)
  3. docker-compose up

I'll be able to simplify mine because of recent developments in Docker (docker -machine, -compose, -swarm) and Ambari (Blueprints). Like I said, standing on the shoulder of giants.

 

What sort of cluster are you looking at? Will it be based on Hadoop? I'll probably be building the custom parts of the system using Akka, which is also new to me. I'm still new to the big data world. I'm having fun being a novice and having to research/learn everything from scratch again :)

  • Like 1
Link to comment
Share on other sites

CentOS 7/KDE Plasma Desktop/"ElementaryLuna" Window Decoration/"Air for netbooks" Desktop Theme/HighContrast icons

 

"Search and Launch" Desktop Layout

post-450287-0-40539500-1433060316.png

 

Dolphin

post-450287-0-60604000-1433059822.png

  • Like 1
Link to comment
Share on other sites

This topic is now closed to further replies.