Skip to content

Linux script to automatically shutdown when load average is low

I’ve been using on-demand machines (similar to AWS EC2 and Google Cloud VM instances) to perform large computations. Those machines are both pay-as-you go and pretty expensive, so you want to terminate them as soon as your computations are done.

I’m sure there’s some fancy way, using tools provided by the cloud machine provider, to automatically shut down a machine that’s idle. But I thought, rather than looking for each vendor-specific, complicated (and likely billable) solution, I’d come up with a bash script. And here it goes:

while :
  load5M=$(uptime | awk -F'[a-z]:' '{ print $2}' | cut -d, -f1)
  echo $load5M
  if (( $(echo "$load5M < $threshold" | bc -l) )); then
    sudo shutdown now
  sleep 5

It's an infinite loop, with a 5 seconds pause, which gets the load average over the last 5 minutes, displays it, and if it's lower than the defined threshold of 0.5 (that's half a CPU core), immediately stops the machine. Simple enough, apart from the 2 magic lines needed to get the load average and to do a comparison between 2 floats (I found that surprisingly tougher than comparing 2 integers!)

A little warning though: be sure to test it, to check that your provider doesn't automatically restarts a terminated machine. That's quite unlikely, but it would be a shame ^^

Posted in Linux.

How to install an old version of a game in Steam

There are multiple guides for this already, but they’re just too insufferably verbose and long. If you need a long version, see one of them, like this one. Otherwise, here is a short version:

Concise step-by-step

  1. Go to steam://nav/console (this opens the steam console)
  2. Run download_depot [appid] [depotid] [manifestid]
    (more details, notably how to find those IDs, at the end of the post)
    NB: the download can be slow, and it won’t show any progress.
  3. Note where the game was downloaded. It will typically be located in [WhereverYouInstalledSteam]\steamapps\content\app_[appid]\depot_[appid]
  4. Find where the current game version is installed (in Steam client, right-click on the game → properties → local files → browse local files)
  5. Replace the current game files with the newly downloaded ones (keep a backup if you want – although re-downloading the latest version is trivial)

Some more details on step 2.

The app ID and depot ID are technically useless since manifest IDs are unique, but Valve decided to make them mandatory anyway. I guess they just want to make sure running commands, while still possible, is as painful as possible.

For instance, to get the June 5, 2019 – 10:55:15 UTC version of “Love Thyself”, you’ll run
download_depot 949060 949061 7335981433319872232
and it will be downloaded into

The app and depot IDs can be found here:
And the manifest IDs can be found here:

Other example, Postal Redux from 29 May 2018:
download_depot 401680 401681 4007769963011890934
List of depots:
List of manifests:

Update (2020-12-22)

It seems that after a while (and/or after a certain number of newer versions have been published), old manifests are removed. For instance, I tried to download a manifest from a bit more than 2 years ago, which had 5 newer manifests: it wasn’t available anymore.

Posted in Uncategorized.

How to catch wild pigs

You catch wild pigs by finding a suitable place in the woods and putting corn on the ground. The pigs find it and begin to come every day to eat the free corn. When they are used to coming every day, you put a fence down one side of the place where they are used to coming. When they get used to the fence, they begin to eat the corn again and you put up another side of the fence.

They get used to that and start to eat, again you continue until you have all four sides of the fence up with a gate in the last side. The pigs, who are used to the free corn, start to come through the gate to eat, you slam the gate on them and catch the whole herd.

Suddenly the wild pigs have lost their freedom. They run around and around inside the fence, but they are caught. Soon they go back to eating the free corn. They are so used to it that they have forgotten how to forage in the woods for themselves, so they accept their captivity.

It works on humans too: the government keeps pushing us toward communism/socialism and keeps spreading the free corn out in the form of programs such as supplemental income, tax credit for unearned income, tobacco subsidies, dairy subsidies, payments not to plant crops (CRP), welfare, medicine drugs, etc. while we continually lose our freedoms just a little at a time.

There is no such thing as a free lunch.

Source:, but it’s actually a pretty common story

Posted in Uncategorized.

Cutting off work-related digital distractions at work

I recently realized that I wasn’t as productive as I wished I was at work. Sure, the colleagues playing pool at any random time of the day right next to my desk, or the whistling and singing (seriously!) in the open space don’t help, but I noticed I was also distracted by something sneakier: some of my very work tools. Namely, Slack and e-mails.


Slack’s business consists in empowering users to replace their too numerous short e-mails that span long threads with… a hundredfold more numerous instant messages that fill a screenful of channels. Gee, what an improvement! Even with desktop notifications off and my phone most often in airplane mode, the red icon in the Slack browser tab, and e-mail notification if I ignore it too long, guarantee regular distractions. I eventually resorted to some drastic measures:

  • Leaving some channels where I really wasn’t relevant. Like that channel where designers configured Zeplin to send notifications every time they commit a change
  • Muting chitchat channels like #random or #music
  • Starring as few as possible important channels, and hiding by default all channels except the starred ones and those with unread stuff
  • Limiting notifications to mentions and direct messages (and keywords, but I don’t have any), when I have to have notifications on (when working remotely)

I’m down to 6 starred channels and 4 muted channels out of around 25+. I also starred 3 private messaging channels, with tiny groups of people I regularly exchange with. I didn’t leave that many channels, I’d say about 3 or 4. But even then, Slack is now a lot less distracting. Unread stuff flashes way less often, and whenever I do check updates in those less important channels, as soon as I leave them they disappear again. Out of sight, out of mind.

Note that muted channels will reappear when you have unread messages in them, only they won’t be highlighted (unlike non-muted channels). Now that I think of it, this seems logical, but at first I was a bit surprised by this.


That may be a bit trickier depending on your setup and habits. When I last changed my e-mail provider, from the start I added folders and I set up filters so that habitual incoming e-mails end up right where they belong, rather than flood my inbox. Try to do that. But not all at once: every time a new e-mail arrives, see if it’s a regular one that should fall into a folder. By regular, I don’t necessarily mean newsletters: it could also be for instance a contact with whom you exchange regularly.

Since I mentioned newsletters: ditch them. Seriously, if you do just one thing about your e-mails, I think that’s the one, and that’s easy enough. Like the incoming e-mail filter, don’t try to do it all at once, do it as they come. When a newsletter arrives, ask yourself: does it really interest me? Has this newsletter interested me at any time within the last X months? If no, hit that unsubscribe button. If yes, ask yourself if you really need to have that information pushed into your inbox, or if you can just actively consult it in your own time.
Unsubscribing is easier than ever now, as GDPR prompted newsletter managers to make sure unsubscribing is easy. Since I started the draft of this post, I think I unregistered from about 20 newsletters. My e-mail box feels so much quieter now 🙂

A last idea about your e-mails, although that one is hard to reach: try to keep your inbox empty. The previous tips are more important, and kind of a prerequisite, in order not to waste time moving e-mails around. Also, achieving a truly empty inbox might be a bad goal if you focus on it so much that it becomes in itself a distraction. But an empty, or near-empty, or at least an inbox where you can see the bottom of the list without scrolling feels quite relaxing to me. So I do try to keep my inbox to less than a screenful. Even if it means moving some e-mails into a “todo” folder that I process regularly: the inbox is where I land whenever I open my e-mail tab, a little stash out of sight in a todo folder feels better than a crowded inbox.


Slack: leave and mute channels, star the few important channels, hide non-starred channels, tune down (or fully turn off) notifications
E-mails: unsubscribe from newsletters, auto-sort regular incoming e-mails into folders, move the rest manually out of the inbox

Posted in Uncategorized.

Using Freenet with OpenJDK (AdoptOpenJDK) on Windows

Java/Oracle recently rolled out a new licensing policy. Frankly, I find it’s a mess and I don’t really understand what is and what isn’t allowed. It seems personal use and development use are both allowed, but still, downloading the SDK now requires creating an Oracle account. That broke the camel’s back. So I looked into alternatives.

AdoptOpenJDK seemed nice. It provides builds that seem regularly updated, for OpenJDK 8, 11 and 12, and it even lets you choose which Java VM you want, between HotSpot and OpenJ9. That JVM choice doesn’t seem to matter that much, from the few benchmarks I found, but still it’s appreciated.

Installation is straightforward, and I was soon able to get this in my console:

> java --version
openjdk 12.0.1 2019-04-16
OpenJDK Runtime Environment AdoptOpenJDK (build 12.0.1+12)
OpenJDK 64-Bit Server VM AdoptOpenJDK (build 12.0.1+12, mixed mode, sharing)

A nice upgrade from Java 8u201.

But, to my surprise, Freenet wasn’t able to find Java (so wasn’t able to run at all). After a brief search, I found that I was missing the registry entries for Java. Maybe I messed up during setup, but anyway it can be fixed quite trivially, by defining the following keys via Regedit (you could also just put this in a .reg file and “run” it):

Windows Registry Editor Version 5.00


[HKEY_LOCAL_MACHINE\SOFTWARE\JavaSoft\Java Runtime Environment]

[HKEY_LOCAL_MACHINE\SOFTWARE\JavaSoft\Java Runtime Environment\12.0.1]
"JavaHome"="C:\\Program Files\\AdoptOpenJDK\\jdk-"

Note that, depending on your version, you’ll want to replace “12.0.1” with whatever you have (and of course, adapt the path too). Although I’m actually not that sure whether the version number matters for real, as long as both occurrences match.

Freenet should now be able to start.

Edit: I tried on another computer, and I tried the “Javasoft (Oracle) registry keys” option during setup. It created some keys automatically, but not the ones needed for Freenet to work: it created keys under HKEY_LOCAL_MACHINE\SOFTWARE\JavaSoft\JDK, which I guess might turn out useful for development, but not under HKEY_LOCAL_MACHINE\SOFTWARE\JavaSoft\Java Runtime Environment, which is what Freenet needs.

Update (2019-07-31)

I just tried on yet another computer, where I hadn’t run Freenet in a while. Mysteriously, despite setting all this configuration properly, Freenet still wouldn’t start (it was still looking in the old path, I have no idea why… maybe a restart will help). As a last resort, I added the full path to the Java executable, in [Freenet path]/wrapper/wrapper.conf, with the follwoing line (NB: it should already exists, you just need to complete it):\Program Files\AdoptOpenJDK\jdk-\bin\java.exe

Update (2019-11-11)

I don’t know how I didn’t notice that in the first place, maybe it’s because of the “AdoptOpenJDK” name or because they put forward the JDK so much, but they provide both a full JDK and just the JRE alone. So unless you’re planning to do Java development, you can grab just the JRE, which is about 4 times smaller than the JDK. I updated the link in the post, to point to a download page where it’s clearer that you can choose between those 2 options.

Posted in software, Windows.

Installing Rust in a custom location on Windows

It’s actually pretty well described in Rust’s documentation. I’m just putting it here because the documentation is large and I appear to have a hard time finding those specific instructions in a timely manner every time I need them.

First, you could grab a GUI installer from there, but Visual Studio Code doesn’t seem to like it much. And I also remember having some issues running rustup in that context.
This is why I rapidly decided to use the “recommended” rustup-init.exe.

Before running said rustup-init.exe:
1) Set the CARGO_HOME environment variable to where you want cargo to be. I picked D:\PROG\PROGRAMMING\Rust\cargo. (NB: for a convenient way to edit environment variables, I recommend Rapid Environment Editor)
2) Set the RUSTUP_HOME environment variable to where you want cargo to be. I picked D:\PROG\PROGRAMMING\Rust\rustup (I’m so creative, I know).
3) If you’re using Rapid Environment Editor, make sure you SAVE (until you do, the environment variables that you created/modified/deleted/etc are NOT actually changed)
4) Make sure you start a new console to run rustup-init.exe. If you use a console that was already running before you added the environment variables, that console won’t have them. If by any chance you are using ConEmu, you need to close and reopen the whole ConEmu: just opening a new tab won’t do, if ConEmu was already running before you add the variables. I insist on this point, because rustup-init.exe will give you NO warning/notification as to where the install will be performed, until it’s all over. So if you’re not careful, you’ll end up with Rust installed in its default location before you can say “God fucking dammit”.

Now, you can (finally) run rustup-init.exe. Make sure to pick option 2) if you want to install the GNU/GCC version rather than the default MSVC, or if you want nightly rather than the default stable. I’m not a fan of using nightly, because it contains features that could get removed at any time, but sadly big frameworks like Rocket require it.

Posted in programming, Windows.

aToad #25: Rapid Environment Editor

A user-friendly GUI to edit Windows’s environment variables

A dozen years ago, some guy decided he couldn’t stand Windows’s default, painful environment variables editor and created this program, Rapid Environment Editor (or “RapidEE”), with a really nice interface to edit environment variables. It’s been regularly updated since then, and even though Windows 10 significantly improved the default UI for editing environment variables, it’s still quite superior IMO.

Notably, it provides a usable browsing interface when adding a folder to the PATH env variable (I just tried this with Win 10’s native UI, even with the improvements it’s hell, and I accidentally overwrote another folder name – I was able to cancel, though), and it highlights in bold red whenever an env variable points to a folder that no longer exists. If you start it as non-administrator, you’ll only be able to edit your user variables, but there’s a convenient button to directly restart RapidEE as Administrator.
Among the extra goodies, the website provides downloads for many older builds, in particular builds of the last versions that supported Windows NT/2000 and Windows 95/98/ME, and the software is available as a portable version, which consists of just a .exe file with a companion rapidee.ini.

Development seems to be pretty slow nowadays, but it’s simply feature-complete… It makes sense to have just some maintenance and translation updates.

Posted in A Tool A Day, Windows.

aToad #24: Bulk Rename Utility and Duplicate Files Finder

The names are pretty self-explanatory… A mass file renamer and a tool to find duplicate files

Bulk Rename Utility is a Windows freeware that makes it “easy” to mass rename files.
On the plus side, it has plenty of features, you can use regular expressions, use ID3 tags and EXIF meta data, change the case, add some numbering, preview file name changes, change the files creation date, etc, etc. On the minus side, that feature-richness comes at a price: it’s not very simple to use. Particularly when you don’t use it very often: since I rarely use it, most of the times I use it I have to check the help file first. Still, I find it pretty great.

Duplicate Files Finder is a multi-platform (Windows/BSD/Linux) free and open source software that allows you to detect duplicate files and delete them. Pick the directories you want to include in the scan and hit Go. Or optionally, you can include or exclude specific files names, and filter your search with minimal and maximal file sizes (excluding tiny files can speed up the search and make the results more readable).
The scan is pretty fast, as it first matches files with their size, and only then compares those that have the same size. The only big weakness of this software is that deletion of duplicates can only be done one by one, i.e. for every file that has duplicates, you have to manually select which one you want to keep. So, it’s good to deal with a few large duplicated files, but it’s not very effective to deal with many small duplicates (all the more reason to exclude smaller files from the search, IMO). Still, my typical use case is mostly on large-ish files that are not too numerous (or whole duplicate folders), so it does the job efficiently enough.

Posted in A Tool A Day.

Upgrading Ubuntu Server from a LTS to the next (non-LTS) version

The configuration file to edit should be shown when you run “do-release-upgrade”, as long as there is a new version after your LTS. But for some reason it wasn’t there (or I missed it) many times before when I previously ran this command. So at least now I won’t lose it anymore.

sudo apt-get update && apt-get upgrade (because the release upgrade command will require you to have all current updates)
sudo nano /etc/update-manager/release-upgrades
Change Prompt=lts to Prompt=normal
sudo do-release-upgrade

That’s pretty much it. Note that you can’t jump a version, so for instance if you’re on 17.10 and you want 18.10, you’ll have to upgrade first to 18.04 LTS and then to 18.10

Posted in Linux.

Renewing the Thecus N7510’s TLS certificate

The Thecus N7510 is a cheap NAS that used to be popular for its large amount of disks (7) while still being as cheap as (or even cheaper than) most 4-disks NAS.

It is powered by Thecus OS, but sadly it seems that its version of Thecus OS isn’t maintained very actively anymore. Particularly, the SSL/TLS certificate used for FTP over TLS expired about a month ago. Which is pretty annoying, because FileZilla refuses to let you permanently ignore a certificate expiration alert (for stupid reasons, but this isn’t the first time the FileZilla developers provide poor explanations for equally poor choices – we can only live with that).

So the only option I had left was to try to upgrade the NAS’s certificate by myself. Gladly, this turned out fairly easy, as I wrote a guide before on how to create your own self-signed certificate. So the only new (and minor) difficulty was to find where the current SSL/TLS certificate of the N7510 is. I quickly found that it’s named /etc/ssl/private/pure-ftpd.pem, which contains both the server private key and the signed certificate (something very slightly different from my previous guide: you just need to stash 2 files into one .pem file).

If they’re not already enabled, you need to enable SSH and SFTP from the ThecusOS control panel (the SSH & SFTP toggles are in Network Service > SSH)

Once this is done, here are the commands I used (cf the linked guide if you need more details) to generate the certificate:

cd /etc/ssl/private
openssl genrsa -des3 -out servPriv.key 4096
openssl req -new -key servPriv.key -out servRequest.csr
cp servPriv.key servPriv.key-passwd
openssl rsa -in servPriv.key-passwd -out servPriv.key
openssl x509 -req -days 3650 -in servRequest.csr -signkey servPriv.key -out signedStartSSL.crt

At this stage, you have everything you need excepted the “stashed” pem file.
At first, I tried to use nano to create it, but the Thecus N7510 doesn’t have nano 😡 So, I connected via SFTP (with FileZilla) as root (that’s why I told you to enable SFTP along with SSH earlier). Then I grabbed servPriv.key and signedStartSSL.crt, and put them both into a single text file (not sure if the order matters) name newcert.pem.

Just for the sake of clarity, newcert.pem looks like:

[base64 stuff]
[more base64 stuff]

Finally, I uploaded newcert.pem into /etc/ssl/private, renamed pure-ftpd.pem to pure-ftpd.pem.bak, and renamed newcert.pem to pure-ftpd.pem.

All is now ready, the last thing you need to do is to restart the FTP server. The easiest way to do it is to disable then re-enable it via the ThecusOS control panel (Network Service > FTP).

Now, when you connect with FileZilla to the FTP server, you’ll see your new, non-expired, certificate, and will be able to trust it permanently (that is, until it expires in about 10 years).

Posted in FTP, security, servers.