When Red Dead Redemption 2 was released, my best friend bought it and she enjoyed it so much I wanted to play with it too. But my Xbox One hadn't been started in months, and it took a while to update the beast. After installing the game, I took my favorite controller, which is my Xbox One Elite Controller. It's a fantastic piece of electronics, although a bit expensive if bought without a discount.

While playing, I immediately felt that something was wrong. The A, B, X and Y buttons where all a bit sticky. It felt like the buttons weren't getting back up until all the pressure was removed (ie. until you removed your fingers completely). Usually, you could keep your fingers on the buttons, and just apply more pressure to activate them, but not this time. This drove me a bit crazy, because in some games you really need some reactivity, and the controller just wasn't cutting it anymore.

What's the problem ?

After searching for a while, it looks like it's a cleaning problem. The controller needs to be cleaned to work properly again. Most of this blog post applies to the standard Xbox Controller.

Let's tear down this controller

Just a little disclaimer before starting: This will completely void your warranty, and, obviously, I can't be held responsible if you break your controller while following this page's instructions, you're doing this at your own risk.

🔧 The tools you'll need:

1. A security Torx TR8 screw driver (the ones with a hole in the middle)
2. A standard Torx TR6 screw driver
3. A soldering iron
4. Some cleaning alcohol
5. And, for the most organised / tooled, a magnetic mat for the screws

I'm personally using an iFixit screwdriver kit that contains everything you'll ever need to repair this kind of electronic devices. If you want to know which one I used, please let me know in the comments.

Remove the side covers

Those side covers are hiding some screws, so we need to take them off to start with. You want to gently pry them open from the side, and work your way around. I did it with my nails, but some others used credit cards or opening tools. They should come off themselves pretty easily. Don't forget to:

• Remove the battery
• Remove the joystick pads
• Remove the arrow pad.

Little note: there a little adhesive strap inside to help them stay in place.

Remove the screws

Using the Torx TR8 screw driver, remove the 5 screws holding the front and back panel together. The fifth one, as indicated, is hidden behind the official sticker, so, by unscrewing it, you're effectively voiding your warranty.

Now that you've done this, the front panel should come off on its own, with literally no help. Please take care of removing the "profile selector" button (the one right in the middle of the controller), or it will fall off at the next step and you might loose it.

Removing the back panel

First, before removing the back panel, you must take off your joystick adapters. Those should come off easily by gently pulling them.

Just a side note, look at the joystick components: they are from ALPS, probably the RKJXV1224005, and are usually high quality. Thumbs up for putting good components in there. Also, as they are standard, they are easy to find and order, in case you need a replacement.

Now let's work around the back panel. Before starting, you must know that there is a ribbon cable running from the back panel to a PCB in the middle frame. So don't be too violent during this part or you might break the ribbon. This ribbon is used to carry the signals from the 4 optional buttons under the controller, so if you're repairing a standard Xbox controller, just forget about it.

Now, that said, let's remove that back panel. Please note that it's a bit harder than the front panel..

• Start by gently lifting the back panel from the top of the controller. By top, i mean the side where the battery is.
• The microphone port that is kind of holding the panel in place. So you'll have to make the panel slip over it
• Again, be careful with the ribbon !

You should end up with something like the next picture.

Now, we have to disconnect the ribbon. To do that, on the middle frame PCB, identify the little connector. Gently lift up the black part to free the ribbon, and pull it out of the connector.

After removing the back panel, you should end up with your controller looking like this. Note that the vibration motors aren't attached to the middle frame, so they might fall off on their own. That's not a problem.

Removing the top assembly

The top assembly includes the LB and RB buttons and the center plastic part which goes around the Xbox button. This part is also fragile, like many parts in this controller, so please be gentle with it. From the "buttons" side of the controller, you'll see that the middle plastic part has two holes, which helps it stay in place, and so does the Xbox button.

To remove it, lift the plastic part between the two hole, push the Xbox button, and try to make the plastic part slip over it. Once done, the whole contraption should come off as easily as the front panel

If you look at how the LB and RB button works, you'll see that they have a little wing that goes in between two black plastic guides before hitting the actual button. Remember this when you'll reassemble your controller, because if you don't make the little wings go between the guides, they'll break or you won't be able you use these buttons.

You'll also have to remove the sync button (the green button in the middle, in the above picture). If you don't, it will fall off when we'll remove the last PCB, and you might loose it. Now you should end up with your controller looking like this:

Removing the first PCB

There are two PCBs in there, which are connected to each other through some headers. Now, to get the first out, we need to un-solder the vibration motor cables, and then remove the two screws holding it in place.

There are four motors to un-solder, which means that there are 4 pairs of two cables. They are easily identifiable:

Using your soldering iron, heat-up the solder area until you can pull the wire out of the PCB. Repeat this for each wire. If you feel like you don't know enough about soldering, there's a nice video on YouTube about it : https://www.youtube.com/watch?v=Qps9woUGkvI

Congratulations 🎉 You've done the hardest part ! Now that you've removed the cables, let's remove the two screws holding the board in place using the TR6 screwdriver:

After removing the screws, you'll see that the board stays in place. It does because of the connectors between the first and second board. To remove it, you must pull gently the first board around the two hidden connectors:

With that first board removed, you should end up with something like this:

Note that the jack isn't soldered. It's normal that it moves. It just sits between the two boards, and makes contacts with the first board. Don't forget to place it back when you reassemble the controller !

Removing the second board

This is the easiest part. There are six screws to remove and you're done. Use the TR6 screw driver.

After removing the last PCB, you'll get access to the plastic conductive layer, and to the buttons.

Cleaning the controller

After all this hard work, the only remaining task before reassembling the controller is to clean the button chambers, and the buttons themselves. To do that, use some cleaning alcohol and a Q-tip and make sure to clean everything correctly. Make sure it's dry before assembling.

Pro tip: When you're done cleaning, do not reassemble the whole controller. Just put back the buttons, the plastic conductive layer and the first PCB, screw it back and test if the buttons are still sticky. If not, clean again.

Reassembling the controller

Please read this post upside down. There's no special trick used to reassemble it. Just make sure all the tiny things are put back while assembling:

• The sync button
• The profile selection button
• The jack port

I tell you this because I happened to forget each one at least one time 😅.

TL;DR

I hope you've enjoyed this post, and if you have any questions or feedbacks, do not hesitate to tell me in the comments below.

Header Picture by wu yi on Unsplash
I took all the pictures used in this post, and if you want to use them please contact me first.

Recently, I felt the urge to draw, and really wanted to play around with an active pen. I wanted to use some good tools like Photoshop on the go, without having to carry a Wacom tablet. There were many different choices available to me:

• The new iPad pro (3rd gen, 2018) (11 / 13 inches)
• The HP Spectre x360 (13 / 15 inches)
• The Dell XPS 9575 2-in-1 (13 / 15 inches)
• The Samsung Notebook 9 (13 / 15 inches)
• And, of course, the surface book 2 (13 / 15 inches). I didn't consider the surface pro, as the 6th version is still not out in France to date.

All those devices have pros and cons. While iPads have an extremely low pen latency, they are also fragile, and they run on a mobile OS, which isn't great for classic tools. Also, I'm not fond of standard consumer laptops from brands like Samsung and HP (amongst others), as they tend to make more and more disposable electronics. Of course, all of those statements are subjective, and everyone here will have its own opinion.

This brings us to the XPS versus Surface Book fight. The XPS has plenty of advantages, like its lower price point, higher resolution, strong build quality, the Dell d+1 not-that-expensive warranty, and doesn't have everything glued together, meaning you can swap out the SSD quite easily. Still, there are some reasons why I've chosen the Surface Book 2 over the XPS:

• The aspect ratio, 3:2 means more code on screen (hey, i'm a developer)
• The fact that it's completely integrated with the surface pen, and everything has been engineered to work together flawlessly.
• The GTX 1060 graphics card. I won't be using it often, but that's always good to have nearby.
• And finally, you can detach the screen and use it autonomously.

How much does it cost ?

It really depends on what options you choose. It starts at 1399€ with a 13.5 inches screen, an Intel i5, 8gb or ram and 128gb of SSD, all the way up to 3799€ for the top-of-the-line model with a 15 inches screen, and intel i7, 16gb of ram and 1tb of SSD. The version I bought, the 15 inches, Intel i7, 16gb RAM and 512gb SSD costs around 3300€. Hopefully I got a nice discount and paid around 2700€.

I find these prices to be a bit high. Microsoft doesn't have the pricing power of Apple , and even with the discount, I still find it really expensive. Actually, at least least for the 15 inches model, Microsoft aligned its prices with Apple's MacBook Pro. If you select the same components on both sites, you'll end-up with the same price.

Note: the CPU and RAM are soldered onto the motherboard, and thus, you won't be able to upgrade later. The SSD is a standard Samsung NVMe 2280 drive, and can be replaced. Well, only with a lot of hard work, as it involves removing the strongly glued front panel.

Let's unbox it !

After running it for the first time, I spent nearly a full day trying to update it. It was shipped with Windows 1709, and it tried to install the 1803 version. At some moment in the process it would just put a high read pressure on the SSD, and continue to do that for hours. I left it work overnight and by the morning it had finally finished.

For the geeks: I tried to understand what was going on there. It seems like as a part of the upgrade plan, it must take some ESD files and process them. One for each language (there were around 15 of them), and each one was just painfully slow. The dism process would just spend it's whole time unpacking tens of thousands of tiny files and then work on them.

My first impressions

My biggest fear was that, due to the weight of the tablet versus the weight of the base, that it wasn't really "lapable". But to my greatest pleasure, it doesn't fall on the back easily, even at the maximum screen inclination.

The screen

It's amazing. The colors are great, the aspect ratio is really comfortable, and the resolution is perfect. The laptop is shipped with a 200% UI scaling, which I reduced to 175% and it now feels perfect. The brightness is automatically adjusted following ambient light changes, like our phones do. (I knew this feature existed before, but I've never seen it working on a windows laptop.)

Some tests mentioned that the screen's response time is high, but i didn't really experience this in a significant way myself. I've been used to 60 fps standard screens, so I'm not the best to judge on this particular problem.

The pen

You can start using it directly without pairing anything or whatsoever. You just put the tip over the screen and you'll see the cursor move. That's it. It just works. Now for other features like the "eraser button" to work, you need to pair it via Bluetooth. The manual says you can just put the tip on the screen for ten seconds, but it didn't work for me, I had to pair it manually.

The pen is not rechargeable. However, there's a tiny battery inside that can easily be replaced. The advertised battery life is about 1 year, but I used it for the first time yesterday so I can't really talk about this.

It's also really balanced. This makes it a pleasure to use. It's not too light, and not too heavy, meaning you won't mind using it for hours, while still feeling it in your hands.

The pen has 2 buttons, one located in the eraser, and one on the side, which acts as a right click. It's also worth noting that the pen has magnets and is made to stay on the side of the screen, so you don't lose it.

Note: To pair it manually you have to press the eraser button for long enough to see the little green led show up and blink. At first I was too soft with this button and waited like 1 minute with my finger on it, but you really have to hear a "click" when pressing it, which requires some additional pressure. I think they wanted to avoid false clicks when erasing on the screen (and randomly triggering the ink workspace).

The keyboard

The keyboard is also really amazing, it feels great to use, and I'm actually using it to write this review. The pressure that's needed to activate the keys is just right, which means that it's hard enough to feel it, but light enough not to make your fingers tired. It doesn't make much noise like other laptop keyboards. Last but not least, it's backlit, and you have 3 levels of lighting available.

I miss the media keys from this keyboard. There's no "previous track" and "next track" key, which is quite annoying as a huge Spotify user. I also needed some time to adapt to the Fn key, because it acts as a toggle button. You click once on it, then click on your favorite function key, and finally click again on the Fn key to get back to the normal behavior. I have to say that it's a bit too much to just refresh a browser (F5).

There's still room for improvement

While I was super happy with this device, I've also discovered some little downsides.

The screen/keyboard connection

If the tablet wobbles too much, then it may disconnect from the base, requiring you to detach and attach again to get some functionality back. Its a little frustrating when using in bed for example, because if you move a bit too much inadvertently, then you have to do the whole process. Well, at least, we don't have to reboot.

The plastic strap on the top

There's piece of plastic on the top of the screen, where there are the power and volume buttons. Well, for the volume button, the little strap is not attached and quite long. This is a problem because if your finger moves out of the glass panel at this place, you end up bending it.

The magnetic surface connector

The surface connector is great, don't get me wrong. It's a fantastic piece of technology that allows you to connect magnetically a whole docking station, with all its wires and complexity. But, it's round.

It happened to me that it just escapes the connection while still being magnetically attached to the laptop. Which gives us something like this :

TL;DR

This is an amazing piece of technology. We really feel all the engineering that has been put in this laptop. The build quality is great, and there are many little details that haven't been left off. For a second generation product, the downsides are not that important and we may easily deal with them. But i still feel like the price I paid for this, even with the discount, is still a bit too much. I would have wanted 1tb of storage at this price point to be happy.

My original goal was to script some parts of a Go application using javascript. Knowing chromium components were portable, and that Nodejs was built with v8, I wanted to try it as an experimental platform. This article relates my (little) journey through building the v8 engine on an up-to-date arch linux. In this post, I assume that you already have a working build system (with clang, etc). See the base-devel group for more information.

Update 1: After building v8, I ran into some issues to build the v8 samples myself, so I added the relevant steps that had to be done to get them to link.

Get the depot_tools

The first step is to get chromium's depot_tools, as this package gives you all the needed tools and commands to get and build the code. Actually, it's pretty simple to get them, just run through the following commands, from the folder of your choice :

git clone https://chromium.googlesource.com/chromium/tools/depot_tools.git
export PATH=`pwd`/depot_tools:"$PATH"

You might want to add the export line at the end of your .bashrc (and replace the pwd with an absolute path) in order to persist the path modification.

Fetch the v8 source

To get the v8 source code, choose the folder you want to work in, and run

fetch v8
cd v8

After some time, you'll end up with a ready-to-use™ file tree, ready to be built. It might weight over 1 gigabyte. Now, to build v8, the first thing to do is to generate the build files, and then launch the build :

tools/dev/v8gen.py x64.release
ninja -C out.gn/x64.release

Static build

As mentionned by @stricker in the comments, you may want to edit the default build arguments in order to have a static build :

gn args out.gn/x64.release

And then add/edit the following lines :

is_component_build = false
v8_static_library = true

This is the default workflow to build v8. If you want to get more information about it, or even customize it, please refer to this document on the v8 wiki. But I wouldn't write this post if it was that easy. If you launch this, you'll actually end-up with some errors early, telling you that libtinfo.so.6 doesn't exist. And that's a problem because you can't just pacman it.

Finding nemo libtinfo.so.6

As libtinfo.so.6 doesn't exist in the standard packages, we have to find another way to get it. Hopefully for us, it's available on AUR (Arch User Repository). We'll need two AUR packages actually:

1. libtinfo AUR link
2. libtinfo5 AUR link

If you know how to build AUR packages just skip to the next section.

Now, if you visit the link of those packages, you'll see that there's a git repository linked to each of them, and that inside these repositories, there's a PKGBUILD file. This file is a description of the process needed to build the package, and is used by makepkg. Here's the process to follow:

git clone <the aur repo>
cd <were you cloned>
makepkg -Acs

The makepkg will automatically use the PKGBUILD file in the current folder.

The -A option ignores the target Arch architecture. The -c option cleans up the directory after makepkg is done, and -s installs the needed dependencies.
stackexchange

Once done, you'll end-up with an archive next to your PKGBUILD file. You can install it using pacman -U <archive file>. And then you're done ! :)

Going back to v8

Once you've installed libtinfo and libtinfo5, you'll have to try again the ninja command that we used before. Normally everything should be fine and build smoothly. Enjoy :D

Building the samples

After building v8, I wanted to check that my v8 build was actually working. Following the Getting Started with Embedding, I tried to run the provided command after building the static libraries:

g++ -I. -Iinclude samples/hello-world.cc -o hello-world -Wl,--start-group out.gn/x64.release/obj/{libv8_{base,libbase,external_snapshot,libplatform,libsampler},third_party/icu/libicu{uc,i18n},src/inspector/libinspector}.a -Wl,--end-group -lrt -ldl -pthread -std=c++0x

Which gave the following output:

As you can see, all the (really long, 2793 lines) output is made of errors. But all of them are undefined references that refers to the std:: namespace, which indicates a standard library problem. Adding -lstdc++ doesn't change anything, so it must be the libc++ library that causes the problem. Again, it's an AUR package that you have to install, using the same process that we used to build libtinfo. Once built and installed, just slightly change the build command to link against our brand new library:

g++ -I. -Iinclude samples/hello-world.cc -o hello-world -Wl,--start-group out.gn/x64.release/obj/{libv8_{base,libbase,external_snapshot,libplatform,libsampler},third_party/icu/libicu{uc,i18n},src/inspector/libinspector}.a -Wl,--end-group -lrt -ldl -pthread -std=c++0x -lc++

And that should be enough to get it to build. Just don't forget to copy the natives_blob.bin and snapshot_blob.bin files from the out.gn/x64.release folder to the same folder as the final executable.

If you have any questions, or something is incorrect, tell me in the comments.
Header picture

Recently, I was looking to rent a server, either dedicated or virtual. I then compared the prices of several big players in this field, like DigitalOcean, AmazonWebServices, Azure, OVH and it's sub brands Kimsufi and SoYouStart, and many others, including Online.net know for it's Dedibox servers. While browsing Online.net's web site, I came across some cloud servers with really attractive prices, but they seemed to be provided by a partner named Scaleway. So I went there, to have a look. I wasn't really sure of what I would find, but it doesn't harm to add another competitor to the list. Here's what their home page looks like when I wrote this article.

My motivation behind this article is not to promote the service, but rather to write a review about it. As it's not a big player on the server market, I thought it would be interesting to present it and talk about this one.

A bit of history

Scaleway is the company that has been created from Online.net's labs. Basically, they created their own hardware to reduce the costs, increase the server density and hence decrease the prices. Here's a quote from their post about their public preview:

When we started the project, two years ago, we had a simple idea: create a platform thought for the cloud, not just an evolution of virtualization. We wanted to avoid the performance cost of virtualization. We wanted to be sure we would not get any performance impact from noisy neighbors.

(source)

Key moments: (found by crawling their blog)

• Public Preview started on the 14th of October 2014
• Public launch on the 2nd of April 2015
• Expansion to another datacenter in Amsterdam, on the 27th of October 2016

Here is their presentation video, which I found interesting and funny:

The offer

Scaleway has two different offers : Dedicated and VPS servers. Their originality lies in the fact that they offer an ARM dedicated server, named C1 (they claim to be the firsts to do so). It has 4 cores, 2GB ram and 50GB SSD. The remaining part of their offer is more or less what we're used to see: x86 cores, good amount of memory, and for each server also 50 GB SSD. If you need more space, you may ask up to 150 GB, with 50 GB increments.

Volume are available per 50GB increments up to 150GB, with a maximum of 15 volumes per server.
Volumes are billed €1 /month per additional 50GB of SSD

Bare Metal Servers

VPS Servers

All plans comes with unlimited transfer, but the speed will depend on the plan you choose, starting from 200mbps, up to 800mbps for the most expensive server. They also state that if you're contacting another server located inside the datacenter, then you may use the full physical link speed (from 1 to 5 gbps), but only for bare metal servers. What's interesting here is that they bill everything individually : the IP address, the SDD and the server itself. This added flexibility is a good thing :

• If you want to upgrade to a bigger server, you may reuse your volume as if nothing changed.
• If you want to cancel your server but keep your data, you can (and even your IP if needed).
• If you want to make you server only available to other servers in the datacenter, you can release your IP (or explicitly ask not to have one when you create the server).

Another good thing: on the "Bare Metal" servers, you can mount up to 15 volumes at the same time. Sadly, this is not possible with the VPS ones, but they give an easy process to move from a vpn to a bare metal one. They also offer an Anti-DDoS, which is really nice for the price. If you upgrade your account (monthly fee required), you also get some bonuses like the access to the API, if you want to automate a bit your account, or priority support.

The pricing

All the prices are shown on the previous screenshots. Some of you may have noticed that the price per hour doesn't match the monthly price, because 720h * 0.002€/h = 1.44 €. But there's an explanation for this: to get the hourly price, you have to add the hourly price of the SSD, that is 0.002 €/h, and the price of the IP Address, also 0.002€/h, to the base server price. Here's my current "draft bill", for the server this web site is currently being hosted on: we can clearly see the three different prices for each resource.

Now you may say that there's another problem with this price : 1.44 € * 3 = 4.32 € and not 2.99 €. Well, in fact, there's another pricing rule. Once you use your server for more than 500h, the price is capped until the end of the month. And we can confirm this if we compute it ourselves : 0.002€/h * 3 (IP+Disk+Server) * 500 = 3.00 €. We finally got it ! This pricing might seem complex at the beginning, but there's no surprise in the end. Here are some excerpts from the documentation:

Phase 1: The resource is billed by the hour. Each additional hour of usage increase the total price of the resource.

Phase 2: Once a running resource reach the cap of 500 hours of continuous usage, the monthly rate is applied. The resource’s price is frozen to the fixed price for the rest of the month, and doesn’t increase your costs.

(EDIT 3 oct. 2018) Be careful: Restarting your server will reset the cap, as it's currently counted as continuous usage. If you have to restart, better do it at the beginning of the month.

(source)

Let's try it !

Now that we introduced our provider, let's give it a try. To create your account, you must give your credit card number, but that's a common practice among providers (DigitalOcean does that too, and some other providers). Once you're done, you'll get access to your control panel, that looks like this:

Creating a sever through the web interface is a smooth experience. There are only a few simple steps. Just don't forget to set your SSH key before creating your server, otherwise you won't be able to access it. But don't worry too much, if you don't, a big yellow popup will annoy you until you do so.

1. Give it a name

Or use the default one.

2. Choose the datacenter

Currently, you may only choose between either Paris (France) or Amsterdam (Netherlands).

3. Choose which server you want

As you may see, each server has it's own product name, composed of some letters, which have the following meaning.

• "V" for "Virtual"
• "C" for "Compute"
• "S/M/L" for "Small/Medium/Large"

If a server is out of stock on a data center, you may want to check out the other one to see if it's available there. Currently, there's no C1 offering on the Amsterdam datacenter. But the server C2L is available there, whereas in Paris, it's out of stock.

4. Choose what you want to install on it

You may install a raw linux distribution, or a ready-to-use image from the image hub (there are even more advanced options, but i'll let you find this out).

The currently available distributions:

The currently available images from the ImageHub:

As you can see there are many options. But it would have been nice to be able to have FreeBSD in the distributions, or a WordPress in the images. You may create you own image / snapshots to base your future servers on, but as I said before, i'll let you find out how to do so.

5. Configure the last parameters

As you can see below, you may add more hard drives to your instance, enable IPv6, choose if you need to have a public IP, or even add some tags, if you feel fancy.

Once you're done, you'll see that the bottom bar updated to reflect your choices, and will display your final monthly price:

Once you're done, the system should take you to your server's page, were you'll be able to see it's current status and parameters. Here's what mine looks like:

Some interesting information

My current server is a VC1S. I installed Debian 8.2, and just the bare minimum : nodejs, ghost and MariaDB (dropin replacement for MySQL). I ran a network benchmark (a command line speedtest.net) and here are my findings:

root@scw-35de9f:/home/random# ./speedtest-cli 
Retrieving speedtest.net configuration...
Retrieving speedtest.net server list...
Testing from Scaleway (163.172.152.32)...
Selecting best server based on latency...
Hosted by NEOTELECOMS (Paris) [1.88 km]: 3.099 ms
Testing download speed........................................
Download: 901.10 Mbit/s
Testing upload speed..................................................
Upload: 229.23 Mbit/s

We seem to have 1gbps of download, and 230mbps of upload, which matches the advertised bandwidth of 200mbps. Now here's my cat /proc/cpuinfo of this server:

processor       : 0
vendor_id       : GenuineIntel
cpu family      : 6
model           : 77
model name      : Intel(R) Atom(TM) CPU  C2750  @ 2.40GHz
stepping        : 8
microcode       : 0x1
cpu MHz         : 2393.902
cache size      : 1024 KB
physical id     : 0
siblings        : 1
core id         : 0
cpu cores       : 1
apicid          : 0
initial apicid  : 0
fpu             : yes
fpu_exception   : yes
cpuid level     : 11
wp              : yes
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ss syscall nx rdtscp lm constant_tsc arch_perfmon rep_good nopl pni pclmulqdq vmx ssse3 cx16 sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes rdrand hypervisor lahf_lm 3dnowprefetch tpr_shadow vnmi flexpriority ept vpid tsc_adjust smep erms arat
bugs            :
bogomips        : 4787.80
clflush size    : 64
cache_alignment : 64
address sizes   : 40 bits physical, 48 bits virtual
power management:

processor       : 1
(same as processor 0)

We can see that this server is having an Intel Atom C2750 (cpubenchmark: 3800), with 8 cores. You may have more information about this CPU by clicking on the links. I then used a nice tool to ping the server from different locations around the globe, and here are my findings :

Overall Experience

I found Scaleway's servers to be cheap and powerful enough for some simple websites. For example, with this blog, I'm using only 2-3% of the CPU, despite having the blog (nodejs+ghost) and MariaDB installed, on only 2 cores of an Intel Atom CPU. I also liked the management interface. It's clean and fast, and I got my server in less than a minute.

But there are some downsides. It misses some good features of OVH's manager, like the CPU usage / Memory graph directly integrated into the control panel. I also miss the possibility to have proper backups. Currently, you have to either use their snapshots system (which is a bit painful), or do it yourself. I'm going for the second one. But it would have been really nice if we could just buy a "backup" subscription, that does it automatically, but maybe this will come in the future, as the company is quite young.

Header picture

If there's something wrong / incomplete, or if you want some more information, just ask me in the comments.

Update 11 nov 2016 Added the ping results from around the world

Recently, people started to realize that the Spotify desktop client was writing way more data than it should reasonably do. It could easily go up to 200 gb on my personal laptop after a day running in background, without playing music. Yes, you read well, without playing music. Let's have a look at it.

Update: 11 nov. 2016

We received another reply, telling us that the 1.0.42 update has successfully been rolled out to all users. Many users on the thread/issue are saying that this update brought the necessary changes to solve this issue. But it should be noted that some users are having some troubles installing the update.

Update: 10 nov. 2016

We got a reply from a community manager saying that they fixed the problem in version 1.0.42. Let's wait for the update and check if the problem has really been solved or not.

The problem

Some people, including me, have seen that spotify has been using a fairly high amount of disk writes lately. While we don't really know the source of those writes, we could monitor the writes to to find out which files were affected by those writes.

Here are some cases reported by users (thank you!) of the Spotify community (see the links at the end of this article for the threads were these pictures were originally posted). Please bear in mind that we don't have Spotify's uptime on each system, so we can't really compare them.

On a Mac :

On Windows :

As you may have already seen, the amount of IO Write Bytes (Bytes Written on mac) are fairly high. If you look well you may see on the Windows screenshot that spotify already wrote 640 GB. To compare, the whole GTA V game (which is already quite big) is 62 GB, which is 10 times less that what Spotify wrote.

But let's be more precise about what IO Write Bytes means on Windows. This counter is not only counting disk writes, but also network writes and other kinds of communication. So you may say, that it may be normal for these counters to take such values. But that's not. In fact, if you look closely, you see that these number are actually orders of magnitude higher compared to the other programs running on the system.

I/O Read Bytes - The number of bytes read in input/output operations generated by a process, including file, network, and device I/Os. I/O Read Bytes directed to CONSOLE (console input object) handles are not counted. (source)

After some research using the resources monitor (resmon.exe), we could find the culprits. Two files, named mercury.db and mercury.db-wal, are getting the largest share of writes. But what are those files ?

/mnt/r$ file mercury.db
mercury.db: SQLite 3.x database

Following the file utility, these files seems to be SQLite databases. Let's have a look at this database with "DB Browser for SQLite". The following pictures shows that the database is indeed a valid one, with three tables: lock_ids, locked_keys and storage.

The mercury.db-wal file is special file used in conjunction with the main database file. For more informations about this topic, please have a look here. But why are those simple database file so much written to ? Well, after some research done by members of the community, it seems that Spotify is calling periodically the "VACUUM;" SQLite function. As you may already think, the VACUUM command is used cleanup unused space on the disk.

The VACUUM command rebuilds the database file, repacking it into a minimal amount of disk space

Here is an excerpt for the documentation:

The VACUUM command works by copying the contents of the database into a temporary database file and then overwriting the original with the contents of the temporary file. When overwriting the original, a rollback journal or write-ahead log WAL file is used just as it would be for any other database transaction. This means that when VACUUMing a database, as much as twice the size of the original database file is required in free disk space.

This suggests that VACUUMing the SQLite database is going to require a lot of IO operations. But this wouldn't be a problem if this wasn't executed as often as it's currently done. Some other member of the community was able to draw a graph, that shows the IO operations over the time:

By looking at this graph, we're able to deduce the time interval between each write. With a rough estimation, it seems like it's done every 5 minutes. 5 minutes means 12 times an hour (60 / 5 = 12). Depending on your mercury.db size, this may represent a big amount of data actually being written to your hard drive. In fact, it's exactly 24 times the size of database file according to the documentation of VACUUM : 12 vacuum per hour * 2 * size of the database = 24 * size of the database. In addition to this, we may add the normal operation of spotify :

• Downloading music & caching, including the accesses to the database for the records about them.
• Synchronizing different stuff (playlists, what our friends are playing, ...)
• Looking for updates
• ...

We could point out what is the problem with the Spotify desktop client right now. But you shouldn't worry too much about the issue, and continue your playback as before. If you want some temporary solutions to the problems, jump to the "Possible solutions" section of this article.

Don't worry too much

Let's do some calculus, to find how much the life of your SSD is going to be impacted. First, we have to define our environment :

• A laptop open 8hours a day with spotify running in background playing music
• Spotify writes 9 GB per hour (upper limit of user's reports)
• We have a 1TB drive with a 150TB warranty (Samsung 850 Pro)

Using spotify 8 hours a days makes the daily amount of writes to be : 8h * 9GB/h = 72GB
The number of days you can do so before hitting the warranty is : 150000GB / 72GB = 2083 days
2083 days mean you could do this continuously for 5 Years, 8 Months and 2 Weeks before hitting the warranty because of spotify.

In pratice, the 1TB Samsung 850 Pro could be pushed to 9PB (that is, 9000TB) before failing. That gives you a lot of time before frying it :)

That's why you shouldn't worry too much. While the amount of data being written to your drive is high, frying it is definitively going to take some years, if not more. But that doesn't justify, and even less means that we should accept such a behavior from software companies.

Possible solutions

There are some solutions available for this problem, that might help while we're waiting for some patch to be released. I will show 2 of those :

1. Deleting the database
2. Redirecting spotify's database to a ramdisk

There are other ones that I could find on the community threads, including modifying spotify's executable directly, or using some program found on the internet, but I definitively do not recommend to use these methods. Even if they work, nobody knows the downsides of doing so.

Both solutions involve playing around with Spotify's database. Don't worry, all your personal data (authentication, and some other stuff), are located in another directory. That's why after using these methods you're not logged out, or don't even loose what you were playing. We're just going to move / remove cached data.

Solution 1 : Delete the database

The first solution is to try to delete the database. In fact, over time, spotify's database gets messy, and it takes more writes for it to maintain it. A good solution is to just delete the folder, and have spotify generate a new one. Please note that you should completely close spotify before deleting the database, to avoid problems.

Deleting the database is easy.
On windows, just delete all the contents of the C:\Users\<username>\AppData\Local\Spotify folder. On macOS, delete the content of this folder ~/Library/Caches/com.spotify.Client/storage, althrough i'm not really sure about it, because i don't own a Mac. If you do, please let me know.

Here are some findings by dnsbty on spotify's community using this solution:

Before reinstalling:
5 mins idle:
95.9 MB written, 114.1 MB read
5 mins listening to playlist:
207.6 MB written, 24.3 MB read

After reinstalling:
5 mins idle:
3.9 MB written, 19.6 MB read
5 mins listening to playlist:
23 MB written, 39.1 MB read

Solution 2 : Use a ramdisk

The idea is simple : if we can't do anything about the writes, let's move them somewhere harmless. That somewhere would ideally be the RAM, that can handle this amount of writes easily (at 25.6 GB/s, it's a piece of cake for it). The downside is that after each reboot we'll get a new database, which means that you can't have offline playlists (but there's a solution for that, too, more on that later).

==Little warning : To use this solution, I strongly suggest you to have 8GB or more of RAM installed in your computer. Not having enough ram could lead to Windows being short on memory, thus swapping too much on the SSD, which is what we're currently trying to avoid. ==

For this, we're going to use a ramdisk. For those who doesn't know what it is, it's a way of creating a small hard drive that lives in RAM. Because of it's nature, if you reboot, it's content get lost. But don't worry, you may configure it to be persistent. The following is for windows only, but may be easily ported to Mac.

Here are the steps needed to go through this :

1. Download and install ImDisk
2. Create a folder with a base structure for the ramdisk
3. Create the ramdisk and configure it
4. Setup the symlink
5. Check that everything works

1. Download ImDisk

You may find the latest version of the imdisk toolkit on sourceforge. After having downloaded the version that matches you windows architecture (x86 or x64), run the installer.

2. Create the base folder for the ramdisk

Now that we have installed the ramdisk, we must create a base structure for the ramdisk. This structure will be load when the system starts into the ramdisk. It's necessary to have it as the symlinks we're going to use later need their folders to exists. Otherwise, they just doesn't work (trying to go through them displays an error). Here are the folders that you should have :

RamdiskTemplate
└───Spotify

You may create this wherever you want. Just make sure not to delete it.

3. Create the actual ramdisk

After installing the ImDisk, you should see a new icon on your desktop named "RamDisk Configuration". Open it. There you should see :

Choose the size to be whatever you feel is right. I choose 4GB because I'm also putting my temporary files there, and doing some work inside. I strongly recommend that you check the "Allocate Memory Dynamically". Doing so will ask the driver to use only the memory needed to store the files. If have 100MB inside it, it will use around 100MB of RAM, with the upper limit being what you set just before. Remember the "Driver Letter" that you choose, you'll need it later. You can also configure it to be used as your temp, but i'll let you have a look around on your own.

Then, you should click on the "Data" tab.

Here, under the "Load Content from Image File or Folder" label, indicate the path were you RamdiskTemplate folder is in you system. On mine, it's in D:\RamdiskTemplate.

If you want Spotify's database to persist after a reboot (and keep offline songs), then you should use "Synchronize at System Shutdown". Basically, just check Synchronize at System Shutdown and Delete data removed from the RamDisk. Using these options will make the ImDisk driver to copy it's content into the RamdiskTemplate folder when the system stops. But there's a downside of using this : when the system boots and the ImDisk driver starts, it will have to copy all the data store in that folder back into memory, which may take some time depending on the number of files and their sizes. It will also take some time for it to synchronize the ramdisk with the RamdiskTemplate folder.

When you're done choosing your settings, click "OK". This will trigger the creation of the ramdisk, and will save your settings. If you see a warning about a volume being unmounted and data lost, don't worry about it, you can safely click OK. Then, an information window will open, and you can safely close it too. Now you're ready to go on.

At this point, you should have a new drive showing up in the explorer, having the size that you choose before. If you don't, please follow carefully the instructions. And if it's still not working, then ask me in the comments.

4. Setup the symlink

Now that we have a new and shiny ramdisk, fully working, we should ask spotify to put his files inside. But he's not really a gentleman, and doesn't have the necessary options. (It does have an option to move the cache, but that doesn't move most of the files that are involved into the high IO). So we have to use a special trick : the symbolic link. Yes, my dear unix friends, windows got that too ;)

First, remove the spotify folder in your AppData\Local. Be careful, do not remove the Spotify folder in the AppData\Roaming folder ! If you do so, you'll have to reinstall it.

Then open an "Administrator" command line. There are many ways to do this, but the quickest is to press Super+X (Super is usually the one with the windows logo). In the menu that appears, click on "Command Prompt (admin)" (I'm sorry, my laptop is in french).

Once you're done, you should get something like this :

Then, enter the following commands (replace with your username, and the drive letter from step 3):

cd C:\Users\<your username>\AppData\Local
mklink /d Spotify <DriveLetter>:\Spotify

If everything is correct, then, if you go again inside your AppData\Local folder, you should see a new Spotify folder. If you open it, it should be empty, and no error must be displayed. If this is your case, then you have successfully redirected spotify to your ramdisk !

5. Test

This is the last part and the easiest one. Just launch spotify and listen to some music. Then open your ramdisk : <DriveLetter>:\Spotify. If this folder is not empty, ie looks like the next picture, then you're done, it works !

Results

After running with this solution (ramdisk) for some hours with Process Explorer running in background, here are the results I got. As you can see, only 500 KB were written to an actual hard drive. All other writes are being redirected to the ramdisk, hence the really low amount of residual writes.

Links to the posts on the Spotify community

1. The forum post
2. The issue
3. The header picture

This is a temporary fix, but it might help some people waiting for a patch to be released. I can't be held responsible if something went wrong on your computer (Usually nothing goes wrong), you’re doing this at your own responsibility.