Lafuente

Custom Clojure helpers in your REPL

2022-03-07T00:00:00-00:00

The problem #

Over time, I wrote several Clojure helper functions I use while I'm writing new code. I want to load those helpers every time I start a Clojure REPL. How can I make those helpers available in any Clojure project I'm working on? Especially in 3rd party projects I don't control.

The solution #

Traditionally, Clojure developers put development utilities in a user.clj file, since Clojure automatically loads any user.clj found in the classpath. In recent Clojure versions, Clojure CLI was introduced, and now we can call any Clojure function from the command line. Knowing that, a better alternative is to move all our helpers to their own project, and to create an alias in our ~/.config/clojure/deps.edn for those utilities.

As an example, let's suppose this is the main file for my helpers project:

(ns local-utils)

(defn init
 [{:keys [main]]
 (println "Loading helpers...")
 (load-file main))

We can add an alias in our deps.edn file:

{:aliases
 {:user {:exec-fn local-utils/init
 :extra-deps {me.lafuente/clj-dev-utils {:local/root "$HOME/projects/clj-dev-utils"}}}}}

Now, in any project, we can do clj -X:foo:bar:user :main "path/to/main.clj". In this example, we are just printing a message before loading the main entry point, but we can do whatever we want.

My personal Clojure helpers are here: https://github.com/jlesquembre/clj-dev-utils

My init function loads some helper utilities, initializes portal, enables malli instrumentation (if present in the classpath) and starts a nREPL server. We can also provide the main function to be loaded.

A library I'm using for my helpers is https://github.com/gfredericks/dot-slash-2 With it, I can create aliases under the . namespace. For example, ./sh points to clojure.java.shell/sh. The . namespace is loaded automatically, so I can evaluate (./sh "ls") in any file without an explicit require

Web components and eleventy

2020-08-11T00:00:00-00:00

If you want to have some dynamic content in your eleventy static website, a good option is to use Web Components, they play very well together.

I'm going to explain how I added a web component to my blog, in this example I'll use model-viewer component. This is the final result:

Astronaut by Poly, licensed under CC-BY.

First, make sure that you set the html option for markdownIt to true in your eleventy config file (usually .eleventy.js):

module.exports = function (eleventyConfig) {
 let markdownIt = require("markdown-it");
 let options = {
 html: true,
 };

 eleventyConfig.setLibrary("md", markdownIt(options));
};

You also need to tell eleventy to copy the web components to your output directory. If you installed model-viewer with npm, you can copy it adding this to your .eleventy.js config file:

eleventyConfig.addPassthroughCopy({
 "node_modules/@google/model-viewer/dist/model-viewer.min.js":
 "js-modules/model-viewer.min.js",
});

But you also need to import the web components in your web pages. To do that, I'm taking advantage of eleventy's Front Matter Data and the Data Cascade. In the front matter I define the components I want to use, that way I can load different components per page:

--- title: Web components and eleventy tags: - 11ty - webcomponents modules: - "model-viewer.min.js" ---

The modules property (you can use another name) is a list with all the files I want to load in that page. You need to process that property when the template is rendered. I use JavaScript for my templates, plus lit-html-server, but you should be able to adapted the next snippet to any of the Template Languages supported by eleventy:

const { html, renderToString } = require("@popeindustries/lit-html-server");
module.exports = class BaseTemplate {
 async render(data) {
 return renderToString(
 html`
 ${data.modules.map(
 (file) =>
 html`<script type="module" src="/js-modules/${file}"></script>`
 )}
 `
 );
 }
};

You can see the real file here: blog/base.11ty.js

And now you can just write the html directly on your markdown files:

<model-viewer
 src="/models/Astronaut.glb"
 alt="A 3D model of an astronaut"
 auto-rotate
 camera-controls
></model-viewer>

This makes pretty easy to embedded custom components in your markdown documents, making it similar to writing MDX.

Custom home-manager installation with NixOS

2020-07-21T00:00:00-00:00

Recently I hear about home-manager-template. The project rationale resonates with me. In fact, I did something similiar on my NixOS configuration.nix.

Apart of trying to make the home-manager installation more reproducible, my goals are:

Install home-manager from my system configuration.nix.
Use home-manager from a git checkout. That makes easier for me to test new ideas or patches.
Don't repeat my self. I don't want to define the same variable at the system level and in home-manager, e.g.: my username or my hostname.

To accomplish that, I have the following function in my configuration.nix:

let
 user = "john";
 userHome = "/home/${user}";
 hostName = "laptop";

 home-manager = { home-manager-path, config-path }:
 assert builtins.typeOf home-manager-path == "string";
 assert builtins.typeOf config-path == "string";
 (
 pkgs.callPackage
 (/. + home-manager-path + "/home-manager") { path = "${home-manager-path}"; }
 ).overrideAttrs (old: {
 nativeBuildInputs = [ pkgs.makeWrapper ];
 buildCommand =
 let
 home-mananger-bootstrap = pkgs.writeTextFile {
 name = "home-manager-bootstrap.nix";
 text = ''
 { config, pkgs, ... }:
 {
 # Home Manager needs a bit of information about you and the
 # paths it should manage.
 home.username = "${user}";
 home.homeDirectory = "${userHome}";
 home.sessionVariables.HOSTNAME = "${hostName}";
 imports = [ ${config-path} ];
 }
 '';
 }; in
 ''
 ${old.buildCommand}
 wrapProgram $out/bin/home-manager --set HOME_MANAGER_CONFIG "${home-mananger-bootstrap}"
 '';
 });
in
{
 users.users.${user} = {
 home = userHome;
 packages = [
 (home-manager {
 home-manager-path = "${userHome}/home-manager";
 config-path = builtins.toString ../home-manager + "/${hostName}.nix";
 })
 ];
 };
}

Let's break that code to understand what is happening.

let
 user = "john";
 userHome = "/home/${user}";
 hostName = "laptop";

First I define some variables I'll use later. It is possible to define a function where you pass them as arguments, but for now I'm just using a let block.

 home-manager = { home-manager-path, config-path }:
 assert builtins.typeOf home-manager-path == "string";
 assert builtins.typeOf config-path == "string";

Now I define a function responsible for installing home-manager . It takes 2 arguments, home-manager-path (path to my local home-manager git checkout) and config-path (path to my main home-manager configuration file, usually called home.nix). I'm also type-checking the arguments.

 (
 pkgs.callPackage
 (/. + home-manager-path + "/home-manager") { path = "${home-manager-path}"; }
 ).overrideAttrs (old: {

I'm calling the derivation provided by home-manager, it is defined here:

https://github.com/rycee/home-manager/blob/master/home-manager/default.nix

Notice that it takes one argument, path, the path to my local home-manager checkout. Since I need further customization, I use the overrideAttrs function to produce a new derivation based on the original one.

 nativeBuildInputs = [ pkgs.makeWrapper ];
 buildCommand =
 let
 home-mananger-bootstrap = pkgs.writeTextFile {
 # ...
 }; in
 ''
 ${old.buildCommand}
 wrapProgram $out/bin/home-manager --set HOME_MANAGER_CONFIG "${home-mananger-bootstrap}"
 '';

These are the arguments for overrideAttrs. Since I want to wrap home-manager, I need pkgs.makeWrapper.

I'm extending the build command. I'm calling the original home-manager build command, and then wrapping it to set the environment variable HOME_MANAGER_CONFIG to a file that I'm generating with the writeTextFile helper function. More details on that later. Notice that HOME_MANAGER_CONFIG is the entrypoint for home-manager, usually that file is home.nix.

 home-mananger-bootstrap = pkgs.writeTextFile {
 name = "home-manager-bootstrap.nix";
 text = ''
 { config, pkgs, ... }:
 {
 # Home Manager needs a bit of information about you and the
 # paths it should manage.
 home.username = "${user}";
 home.homeDirectory = "${userHome}";
 home.sessionVariables.HOSTNAME = "${hostName}";
 imports = [ ${config-path} ];
 }
 '';

This is how I'm generating the main the entrypoint file for home-manager. It's a .nix file itself, which is added to the nix store. When I generate my NixOS configuration, with nixos-rebuild build, I know the value for some of the variables needed by home-manager, like my username or the hostname. I'm taking advange of that to generate a minimal home.nix file, where the values for that variables are injected, and then I'm importing my real home-manager configuration (defined by the config-path variable)

in
{
 users.users.${user} = {
 home = userHome;
 packages = [
 (home-manager {
 home-manager-path = "${userHome}/home-manager";
 config-path = builtins.toString ../home-manager + "/${hostName}.nix";
 })
 ];
 };
}

The last step, this is how I install home-manager from my NixOS configuration.nix. I'm just adding the package returned by my home-manager function. The 2 arguments to the functions are the path to my local home-manager checkout and the path to my home.nix file.

Running a local dns and web server

2020-06-24T00:00:00-00:00

As a software developer, I run multiple web services to be accessed only from localhost. While it is possible to access them on http://localhost, I prefer to have more descriptive urls, like http://docs.local.

I'm going to describe my NixOS setup to resolve .local domains to localhost and access my local web server. I'm using CoreDNS and Caddy to accomplish it. Keep in mind that while I'm showing nix configuration snippets, you could extract the configuration files and run a similar setup in any Linux distribution.

It is interesting to notice that CoreDNS started its life as a fork of Caddy.

DNS server #

While I could use dnsmasq, I wanted to try a newer DNS server, CoreDNS. Let's start the service with some custom configuration:

services.coredns.enable = true;
services.coredns.config =
 ''
 . {
 # Cloudflare and Google
 forward . 1.1.1.1 1.0.0.1 8.8.8.8 8.8.4.4
 cache
 }

 local {
 template IN A {
 answer "{{ .Name }} 0 IN A 127.0.0.1"
 }
 }
 '';

The first block will forward all the DNS requests to Cloudflare and Google DNS servers, and the results will be cached (1 hour by default)

The second block returns an A record with value 127.0.0.1 for any request for a .local domain.

So far, so good, we have a local DNS server up and running. Now we need to configure our network manager to use that DNS server. If you are on NixOS, that's pretty easy:

networking.networkmanager.insertNameservers = [ "127.0.0.1" ];

Web server #

I could use nginx, but I decided to use Caddy instead because it comes with an HTTP API out of the box. I'm not explaining it in this post, but Caddy's API opens the door to many interesting possibilities.

Let's take a look our Caddyfile configuration:

http://docs.local { root * /path/to/dir file_server }

Here we are saying that all requests to http://docs.local should serve static files located at /path/to/dir.

Until #86686 is merged, you need to create your own systemd unit to start Caddy:

let
 caddyConfig = pkgs.writeText "Caddyfile"
 ''
 http://docs.local {
 root * /path/to/dir
 file_server
 }
 '';
in
{
 systemd.services.caddy = {
 description = "Caddy web server";
 after = [ "network-online.target" ];
 wants = [ "network-online.target" ];
 wantedBy = [ "multi-user.target" ];
 serviceConfig = {
 User = "caddy";
 Group = "caddy";
 ExecStart = ''
 ${pkgs.caddy2}/bin/caddy run --config ${caddyConfig} --adapter caddyfile
 '';
 ExecReload = ''
 ${pkgs.caddy2}/bin/caddy reload --config ${caddyConfig} --adapter caddyfile
 '';
 TimeoutStopSec = "5s";
 LimitNOFILE = 1048576;
 LimitNPROC = 512;
 PrivateTmp = true;
 ProtectSystem = "full";
 AmbientCapabilities = "cap_net_bind_service";
 };
 };

 users.users.caddy = {
 group = "caddy";
 uid = config.ids.uids.caddy;
 home = caddyDir;
 createHome = true;
 extraGroups = [ "users" ];
 };

 users.groups.caddy.gid = config.ids.uids.caddy;
}

GitHub actions with Docker and Nix

2020-06-11T00:00:00-00:00

I created my first GitHub action: clojars publish action.

To see an example in action, you can go to jlesquembre/dummy-clj and take a look to the actions tab.

GitHub actions can run in Docker containers, so I created a Docker image. At first, I went to Docker hub and looked for the clojure images, but I realized that I also needed to install maven. There is also a maven image, but since I need maven and clojure on the container, the only option was to create my own image based on the official ones and install the other dependency manually.

Then, I remembered about an interesting project I hear about, Nixery. With it, you can create Docker images based on NixOS, and it makes really easy to install dependecies. In my case, I just needed this:

FROM nixery.dev/shell/maven/clojure

Done, you alredy have a container with some basic linux utilities, maven and clojure. If you need to install something else, just add /other-dep.

Since I'm already a NixOS user, it is a really good solution for me. It gives me an easy way to install dependecies, on the linux distribution I know the best.

Eleventy migration

2020-06-07T00:00:00-00:00

After 6 years from my last post entry, I'm migrating my personal blog to Eleventy.

I heard good things about it and I wanted to give it a try. So far I'm liking it a lot, the best part about Eleventy is its flexibility. As an example, I started writing my templates in Nunjucks, but since I didn't like the syntax too much, I decided to write the templates just with JavaScript. It was an improvement, but I was missing some of the functionality provided by lit-html. I found a project to render lit-html templates on the server, lit-html-server, and it was incredibly easy to plug it into Eleventy and start to use lit-html templates on my static site.

Now I write the posts in Markdown, which are rendered with markdown-it, and the result is passed to the templates to render the final html with lit-html-server

I'll also try to write regularly. I liked the idea from Lambda Island: Advent of Parens, where the author challenged himself to write a blog post for every day of December until the 24th. While I'm not writing a new post every day, I like the idea of writing smaller posts, but more often. Let's see how it goes.

GnuPG Notes

2014-04-24T00:00:00-00:00

Create new key #

Create or edit the file ~/.gnupg/gpg.conf and add the following lines to use the cryptographically stronger SHA512 hashing algorithm for signatures in preference to SHA1 algorithm.

personal-digest-preferences SHA512
cert-digest-algo SHA512
default-preference-list SHA512 SHA384 SHA256 SHA224 AES256 AES192 AES CAST5 ZLIB BZIP2 ZIP Uncompressed

Now we can generate the keys:

gpg --gen-key

The system will prompt us to generate some entropy, we can just run this in other terminal. Any expensive operation over the file system, like find would we also ok.

ls -R ~

By default we have one master key (for signing) and one subkey (for encryption)

List keys #

We can see our public keys with this command:

gpg --list-keys

And our secret keys with this command:

gpg --list-secret-keys

To see more information about the keys, we can use this command:

gpg --edit-key KEYID_or_EMAIL

The output is something like this:

Secret key is available. pub 4096R/AB58430D created: 2014-04-14 expires: never usage: SC trust: ultimate validity: ultimate sub 4096R/D5293C21 created: 2014-04-14 expires: never usage: E [ultimate] (1). John Doe <jonh@gmail.com>

Meaning of the usage field abbreviation:

Constant	Character
PUBKEY_USAGE_SIG	S
PUBKEY_USAGE_CERT	C
PUBKEY_USAGE_ENC	E
PUBKEY_USAGE_AUTH	A

Thus, for example, usage: SCA means that the sub-key can be used for signing, for creating a certificate and authentication purposes.

The primary user id is indicated by a dot, and selected keys or user ids are indicated by an asterisk

User id management #

Use the gpg ‐‐edit-key command to launch the gpg> prompt.

Select a uid:

gpg> uid n

n is the uid number. Use * to select all and 0 to deselect all.

Add a uid:

gpg> adduid

Make uid primary, before you need to select a uid:

gpg> primary

Create subkey #

For extra security, it might be appropriate to physically remove the master private key from your computer, and instead use a second generated subkey for signing files. You can sign and encrypt files as normal with your signing subkey and encryption subkey. If those keys ever get compromised, you can simply revoke them and generate new ones with your uncompromised master key.

gpg --edit-key KEYID_or_EMAIL gpg> addkey . . . gpg> save

Create subkey for SSH authentication #

You need to create a subkey for authentication. To do this, you need to run gpg in expert mode:

gpg --expert --edit-key KEYID_or_EMAIL gpg> addkey # Select (11) ECC (set your own capabilities) # Select (1) Curve 25519 # Select successively (S), (E), and (A) to remove the signing and encryption capabilities and enable the authentication capability. gpg> save

To export your public ssh key:

gpg --export-ssh-key KEYID_or_EMAIL

To make gnupg use your authentication key for ssh:

gpg --with-keygrip --list-public-keys KEYID_or_EMAIL

And add the keygrip for the authentication key to ~/.gnupg/sshcontrol

Export keys #

gpg --armor --export KEYID_or_EMAIL > keyname.public.asc

--armor exports the key in ascii format. Is also posible use --export-secret-key or --export-secret-subkeys

--export exports your whole public key. It probably doesn't make sense to only export a public subkey (public keys are supposed to be public) and various important bits of information are tied to the main key in any case. Your user id, for example, is stored on the main key.

Import subkeys #

gpg --import subkey.asc

It is not possible to delete only the secret master key. If you want to delete only the master secret key, you need to follow this steps:

Export secret subkeys
Delete secret key

gpg --delete-secret-key KEYID_or_EMAIL

Import secret subkeys

You can verify it worked by running:

gpg --list-secret-keys /home/john/.gnupg/secring.gpg ----------------------------- sec# 4096R/AB58430D 2014-04-14 uid John Doe <jonh@gmail.com> ssb 4096R/D5293C21 2014-04-14

The pound sign # means the signing subkey is not in the keypair located in the keyring.

Encrypting a file #

gpg --recipient MY_KEY --recipient OTHER_KEY --encrypt message.txt

The encrypted message is written to message.txt.gpg

I wanted to make sure I could read the message too, for this reason I have two recipents.

References #

Arch Linux, quick installation guide

2014-03-28T00:00:00-00:00

iPXE #

Install from a USB using iPXE.

Download image from http://boot.ipxe.org/ipxe.iso and copy to an USB stick.

Find your USB

lsblk

Copy the image to the USB

sudo dd if=ipxe.iso of=/dev/sdX

Now boot with the USB

Arch Linux Netboot #

Obtain an IP address

dhcp

Start Arch Linux boot, original link is http://releng.archlinux.org/pxeboot/arch.ipxe

chain http://bit.do/archipxe

Get Arch Linux Ultimate Install Script

wget http://bit.do/archis
tar xzf archis
cd <dir> && ./ais

Reboot and finish the installation. If network doesn't work do

systemctl start dhcpcd

Helper to finish the installation, original at https://raw.githubusercontent.com/jlesquembre/arch_install_script/master/ais.sh

wget http://bit.do/jlais

Using the KDE Wallet to store ssh keys with passphrases

2012-03-09T00:00:00-00:00

I have read an article about SSH passphrases (github:help), but they only explain how to use ssh-agent with msysgit and Mac OSX, no word about Linux.

This guide is about how to use KDE Wallet with ssh-agent. This way, we can use SSH keys with passphrases, but we don't need to enter a passphrase every time we use the key, KDE Wallet does the job. I assume you are using KDE.

In 3 easy steps:

1. Install ksshaskpass #

Use your favorite package manager.

2. Create script file #

Create the file

~/.kde/Autostart/ssh-add.sh

On some distributions ~/.kde4/Autostart/ssh-add.sh

And add this content

#!/bin/sh
export SSH_ASKPASS=/usr/bin/ksshaskpass
ssh-add < /dev/null

3. Make it executable #

chmod +x ~/.kde/Autostart/ssh-add.sh

And run the script (or restart the session)

Extra #

For security reasons, is a good idea stop ssh-agent when you finish your session, create the file

~/.kde/shutdown/ssh-agent.sh

With this content

#!/bin/sh
ssh-agent -k

This file should be executable

A good page for more info about SSH: Secure Shell (Arch Linux)

Autofs configuration

2011-03-11T00:00:00-00:00

Autofs uses /etc/autofs/auto.master as its default primary configuration file. We can define two kind of maps in this file, direct and indirect. An example:

# mount point config file options /- /etc/autofs/auto.direct --ghost /media/cifs /etc/autofs/auto.cifs --timeout=60 --ghost /media/sshfs /etc/autofs/auto.sshfs --timeout=60 --ghost +auto.master

First map is a direct map, the other two are indirect. Entries in a direct map contain an absolute path, instead of the relative path names used in indirect maps. The --ghost option tells the automounter to create empty directories of all the mount points. The /media folder should be empty. Let's see the other configuration files, in this example we are going to mount a remote file system using cifs and ssh.

/etc/autofs/auto.direct

# directory mount options device to mount /home/ian/music -fstype=cifs,ro,soft,user=guest ://192.168.1.10/music /home/deb/music -fstype=cifs,ro,soft,user=guest ://192.168.1.10/music

Folders /home/ian/music and /home/deb/music should be empty or not exist. In both we can access to te remote folder /music in computer 192.168.1.10

/etc/autofs/auto.cifs

# directory mount options device to mount ian -fstype=cifs,ro,soft,user=ian ://192.168.1.10/export/home/ian deb -fstype=cifs,ro,soft,user=deb ://192.168.1.10/export/home/deb

Folders /media/cifs/ian and /media/cifs/deb will be created.

/etc/autofs/auto.sshfs

# directory mount options device to mount ian_ssh -fstype=fuse,rw,allow_other,IdentityFile=/path/to/ssh_key :sshfs\#user@192.168.1.10\:/mnt/folder/

Folder /media/sshfs/ain_ssh will be created.

Using public keys with openSSH

2011-01-27T00:00:00-00:00

Generate the key #

We need to create a public key, we do it in computer A:

ssh-keygen -t rsa -b 2048 -f /home/user_name/.ssh/key-name

Press always intro when asked to generate a key without passphrase. This is less secure, but it can be really useful. Now we have in our .ssh folder the keys (key-name and key-name.pub)

Copy public key #

Copy the public key to the remote host you wish to access (in our example from computer A to B) and add it to the file

~/.ssh/authorized_keys

If that file does not exist then you should create it. Or you can use ssh-copy-id to log into a remote machine and append the indicated identity file to that machine's ~/.ssh/authorized_keys file.

ssh-copy-id [-i [identity_file]] [user@]machine

Test it #

In computer A run this command:

ssh -i key user@computerB_IP

If every thing is fine, you should be able to connect to computer B with this command.

Laziness is a virtue, using ssh config #

You can create alias for your ssh connections, this way, you don't need to remember usernames, IPs, etc...

First, create this file:

~/.ssh/config

with this content:

Host alias HostName computerB_IP User username IdentityFile ~/.ssh/path/to/key

Now it is possible connent with this command:

ssh alias

Great, isn't it? For a complete connect list for the ssh config file, type

man ssh_config

File and directory permissions

2011-01-24T00:00:00-00:00

The best way to understand the Linux permissions, is with an example. If we run the command ls -l we get an output like this:

1 2 3 4 5 6 7 8
- rw-r----- 1 user users 8 ene 24 11:01 f.txt
d rwxr-xr-x 2 root root 4096 ene 24 12:14 dir

which maps to

1 2 3 4 5 6 7 8
d uuugggooo links owner group size date name

Numbers from 1 to 8 are not really in the output, but are here to clarify every field.

First character specifies the type of the file. In the example above the hyphen (-) in the first character indicates that this is a normal file, and the d indicates that this other file is a directory. List with all file types
Next 9 characters specifies the files permission. Each 3 characters refers to the read, write and execute permissions for user, group and others. In this example, rw-r----- indicates read-write permission for user, read permission for group, and no permission for others.
Number of links, in this case is 1 for the file f.txt and 2 for the directory dir.
Owner, in this example, user for the file, and root for the directory.
Group, in this example, users for the file, and root for the directory.
Size of the file.
Last modified time.
File name.

Difference in access permission between files and folders are:

	File	Folder
Read	File contents can be read	Directory listing can be obtained
Write	File contents can be changed	Directory content can be changed
Execute	File can be executed	Directory can be accessed

Remember, give execute permission to a folder if you want to access it.

Instead of use the values r, w and x, you can use numeric values, where

r = 4
w = 2
x = 1
- = 0

It is possible combine this values, a table with all the possibilities:

Value	Text equivalent	Meaning
0	---	All types of access are denied
1	--x	Execute access is allowed
2	-w-	Write access is allowed
3	-wx	Write and execute access are allowed
4	r--	Read access is allowed
5	r-x	Read and execute access are allowed
6	rw-	Read and write access are allowed
7	rwx	Everything is allowed

To change the permissions you can use the command chmod, which has this syntax:

chmod [options] MODE FILE

Let's see some examples, to give read and write access to everyone, you can use this two equivalent commands:

chmod a=rw file.txt
chmod 666 file.txt

To give read and write access to user and group, but no access to others, you can use this two equivalent commands:

chmod u=rw,g=rw,o= file.txt
chmod 660 file.txt

If you want to change the access permissions of a folder recursively never use the -R option. Chmod doesn't differentiate between folders and files, and this is really important, because execution access has different meanings for files and folders. Instead use this:

For directories:

find . -type d -exec chmod 755 {} \;

For files:

find . -type f -exec chmod 644 {} \;

To finish, if you want to change the file owner, you can use the command chown, with this syntax:

chown [options] OWNER[:GROUP] file.txt

For example:

chown user1 file.txt

To change the owner recursively, use the option -R If you just want to change the group, use the command chgrp.

Load a kernel module

2011-01-22T00:00:00-00:00

In Linux, to load a kernel module (for example, fuse) run:

modprobe fuse

You can verify that the module is loaded using the following command:

dmesg | grep -i fuse

If you want it to to load a module, add that module to /etc/modules, an example file:

# /etc/modules: kernel modules to load at boot time.
#
# This file contains the names of kernel modules that should be loaded
# at boot time, one per line. Lines beginning with "#" are ignored.
# Parameters can be specified after the module name.

fuse

And if you want to cofigure some options for that kernel module do it in the folder /etc/modules, in our example put the options in the file /etc/modules/fuse

OpenSSH, remote host identification has changed

2011-01-19T00:00:00-00:00

Sometimes when I run the ssh command I get this error:

@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ @ WARNING: REMOTE HOST IDENTIFICATION HAS CHANGED! @ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ IT IS POSSIBLE THAT SOMEONE IS DOING SOMETHING NASTY! Someone could be eavesdropping on you right now (man-in-the-middle attack)! It is also possible that the RSA host key has just been changed. The fingerprint for the RSA key sent by the remote host is xx:xx:xx:xx:xx:xx:xx:xx:xx:xx:xx:xx:xx:xx:xx:xx. Please contact your system administrator. Add correct host key in /home/user/.ssh/known_hosts to get rid of this message. Offending key in /home/user/.ssh/known_hosts:7 RSA host key for 192.168.1.5 has changed and you have requested strict checking. Host key verification failed.

This usually happens when I change any machine in my network. The best way to solve this problem is removing all keys belonging to hostname from known_hosts file. In my case, the remote host is 192.168.1.5, so I run the command:

ssh-keygen -R 192.168.1.5

And now I can access again to the remote host.

An alternative is to set StrictHostKeyChecking to no in your ssh_config. For a better description, take a look to the end of this article: Three helpful SSH tips for developers

Give sudo access to any user

2011-01-18T00:00:00-00:00

Normally, after install the OS (Debian in my case), you don't have sudo access, and you can see this error:

user is not in the sudoers file. This incident will be reported.

To solve it, first log in as root, to do it enter:

su root

Now, as super user, you can run this command:

visudo

With this command we open the sudo configuration file, add at the end this line:

user ALL=(ALL) ALL

Replace user with your user name and save the file. Don't forget exit from the super user account, simply enter:

exit

Tip: #

If you use Debian, and want to change the default text editor for the system, enter this command:

sudo update-alternatives --config editor

A list with all text editor will appear and you can select your favourite.

If you want to know where the sudo incident are reported, xkcd offers you an explanation: