Category: Linux

One of my current projects is the configuration of a new telephone server running Asterisk and FreePBX.

The copper landline telephones of our past have been mostly replaced by cell phones, and to a lesser extent, Voice over IP (VOIP) technology, which uses Session Initiation Protocol to connect desk telephone to a server running the Asterisk PBX software, whether it be local or external. Consumers typically order 1 or 2 VOIP “landlines,” but business offices still have a large number of desk telephones still in use.

I come to Asterisk VOIP from the the perspective of an old web hosting guy. So to me, Asterisk for telephone servers is analogous to an Apache web server for web pages, or to a qmail/vpopmail toaster for email.

There are several subsystems that work together to create a telephone server:

• The server’s operating system, typically Linux for a VOIP server.
• The Asterisk VOIP server and associated plugins.
• The FreePBX web interface program.
• An external “trunk” that completes telephone calls at a service provider over the Internet.
• A set of “routes” that determine how calls are directed between local telephones, or to one or more external termination providers based on least-cost routing.
• A set of extensions, which include extension numbers, passwords, voicemail boxes, and rules on things like call forwarding.
• A set of Inbound Voice Routing (IVR) rules that create voicemail menu choices by touchtone.
• A set of recordings used by the mailboxes and IVRs.

I have the luxury of having an existing system to which I can refer. The existing system uses CentOS, Asterisk, and Elastix HTML GUI.

Although the original server was deployed 9 years ago, the desk phones in the office have all been replaced recently with Cisco SP504G phones.

After surveying options in the marketplace, it was clear that this private little PBX server with 20 clients was only costing C$50/mo. in recurring telecom, vs C$450 for an outsourced solution, based on a few quotes from vendors. So a new build would have to take place.

There are several ways to set up, including virtual appliance images and several installation ISOs that can create a pre-packaged server with much of the configuration work pre-done.

A package called the “FreePBX Distro” is actually a re-packaging of CentOS 7, with Asterisk, plugins, and FreePBX GUI all automatically installed.

There is still a significant amount of work to do, but so far I have been able to configure a “backhaul” connection uses IAX2 between the server and a service provider that will terminate (complete) calls to the outside world. I was able to connect a Cisco SP504G desk phone to make a SIP connection to an extension account on the telephone server, and I was able to make a phone call using that desk telephone to a cell phone answered by a colleague.

More work ahead, especially in terms of the creation of IVR rules and recordings. I plan to test the server with a test phone number, then when it is ready, switch over the main number and all of the desktop phones over a weekend.

Even in an age where office workers use their cell phones for most calls, an Asterisk telephone server can allow for the forwarding of calls based on a touch tone menu to an extension or department that can forward to an outside telephone line.

If you have an older Windows PC (minimum 64bit with at least 2GB RAM), you can reformat it with Linux to function as a home server.

This is only an overview

This post is intended to provide an overview of what is possible. Over time, I will be posting detailed information on how to implement each of these steps. For detailed steps on how to install OpenVPN server software, see my blog post on the subject.

A home server running Linux can perform several functions:

File server

A home Linux server can be setup to run file sharing using a free program called Samba, which makes the Linux server act as a Windows-style file share, accessible via Windows or MacOS, and of course, Linux desktops. You can stream movies stored on your home server to your iPad over wifi using VLC media player, creating your own private Netflix.

VPN server

With a port forwarded on the router to point to the home server, a free program called OpenVPN can be installed and set up to provide remote secure access to the home network. For example, if you forgot a document on one of the computers in your home, you could connect via the VPN to your home server, then become part of your home network from a remote location, able to browse files and connect to desktops within the home network. There are OpenVPN client applications available for Windows, MacOS, Linux, Android, and iOS.

LAMP web server for testing, programming

You can configure a home web server to host Linux/Apache/MySQL/PHP (LAMP) web pages, allowing you to play with the same technology used by large web hosting operations. You can learn a great deal by installing a server and configuring each of its services. This learning experience will give you insights and make you a better administrator of the productized services you manage via web hosting and VPS providers.

A home LAMP server can be used for development of new websites and web-based programs, then deployed on a VPS or shared web host on a web hosting provider for production use with the public.

You can make the server visible to the public Internet by forwarding a port on your router. You can assign a semi-permanent web address to your server by using a service like no-ip.com.

Nextcloud image management application

Nextcloud is a free image management application that runs as a LAMP web site. Nextcloud can take the place of Google Images, or Apple’s iCloud storage: at least for the management of image and media files. By running your own image management application on your own home server, you maintain your privacy, while making files available for computers within your home network and from outside, via the Internet. Nextcloud client apps are available for Android and iOS.

After a software update upgraded an Acer netbook from Fedora 26 to 27, the netbook booted with the error message “no valid boot devices.”

The system was originally installed, then updated, in UEFI Secure Boot mode. Fedora 27 installed a new UEFI certificate in the BIOS, but I had to go into the BIOS to select it as the current default.

Fedora 27 is a close cousin of CentOS 7, which is actually Fedora 19. Most of the documentation for server-centric stuff is still targeted at CentOS 7. Some topics, like how to install and configure OpenVPN, are poorly documented.

I was pleased to find this link, a tutorial on how to install OpenVPN on Fedora 26:

https://chichivica.github.io/2017/08/02/Install-OpenVPN-on-Fedora-26/

I found this guide to be useful, but found myself taking notes on corrections to the procedure. Unable to contact the author of the howto, I offer the procedure with minor corrections here. Note that my procedure was done on Fedora 27.

1) First of all install necessary dependencies

sudo dnf install openvpn easy-rsa

2) Copy rsa scripts to the home folder

mkdir ~/openvpn-ca

cp -ai /usr/share/easy-rsa/3/* ~/openvpn-ca
cd ~/openvpn-ca

3) According to this start a new PKI and build a CA keypair/cert

./easyrsa init-pki
./easyrsa build-ca nopass

4) Build Server certificate, key

./easyrsa build-server-full server nopass

5) Build Client certificate, key

./easyrsa build-client-full client01 nopass

you can omit nopass on steps 3,4,5 if you need to

6) Generate a strong Diffie-Hellman keys

./easyrsa gen-dh

7) Generate HMAC signature to strengthen the server’s TLS integrity verification capabilities

openvpn –genkey –secret pki/ta.key

8) Before openvpn server setting up we need to put appropriate keys ca.crt ca.key server.crt server.key ta.key dh.pem into /etc/openvpn/server/keys folder

sudo ln -s ~/openvpn-ca/openssl-1.0.cnf ~/client-configs/

sudo cp ~/openvpn-ca/pki/issued/server.crt /etc/openvpn/server
sudo cp ~/openvpn-ca/pki/private/server.key /etc/openvpn/server
sudo cp ~/openvpn-ca/pki/private/ca.key /etc/openvpn/server
sudo cp ~/openvpn-ca/pki/ca.crt /etc/openvpn/server
sudo cp ~/openvpn-ca/pki/dh.pem /etc/openvpn/server
sudo cp ~/openvpn-ca/pki/ta.key /etc/openvpn/server

9) Now we need to set up the server itself, firstly copy configurations

sudo cp /usr/share/doc/openvpn/sample/sample-config-files/server.conf /etc/openvpn/server

10) Modify several lines in that configuration file

sudo nano /etc/openvpn/server/server.conf

add these lines at the end of the file:

key-direction 0
auth SHA256
remove ; symbol to uncomment following lines

user nobody
group nogroup

10a) [optional] In order to Redirect all traffic Through the VPN, remove ; from the following lines:

push “redirect-gateway def1 bypass-dhcp”
push “dhcp-option DNS 208.67.222.222”
push “dhcp-option DNS 208.67.220.220”

10b) [optional] Adjust port and protocol if you don’t wish to use default:

port 443
proto tcp

and if you have server.crt and server.key with the different name point to them here:

cert myservername.crt
key myservername.key

11) Allow IP Forwarding. This is fairly essential to the functionality we want our VPN server to provide.

sudo nano /etc/sysctl.conf
and drop a line there

net.ipv4.ip_forward=1
activate that:

sudo sysctl -p

12) Set up firewalld to work with OpenVPN

sudo firewall-cmd –permanent –add-service openvpn
sudo firewall-cmd –permanent –add-masquerade

13) Now we are going to set up our systemd service.

sudo ln -s /usr/lib/systemd/system/openvpn-server\@.service /etc/systemd/system/multi-user.target.wants/openvpn-server\@server.service

sudo ln -s /etc/openvpn/server/dh.pem /etc/openvpn/server/dh2048.pem

server corresponds with the configuration file name in /etc/openvpn/server such as server.conf. So if you have myserver.conf you have to replace server with myserver

14) Now we are ready to start OpenVPN service

sudo systemctl -f enable openvpn-server@server.service
sudo systemctl start openvpn-server@server.service

15) enter in /etc/rc.d/rc.local (reminder: chmod 755 rc.local):

iptables -t nat -A POSTROUTING -s 10.8.0.0/24 -o enp3s0 -j MASQUERADE
(where enp3s0 is the name of your ethernet device)

Done! We successfully deployed our OpenVPN server, and we are ready to move on and set up the client

Client setup

As you remember we already generated client01.crt and client01.key at step 5. Now we need combine them with our general Certificates of Authority in order to build client config file.

1) First of all we need generate Client Configurations. Let’s create client-configs directory and prepare with the keys

mkdir -p ~/client-configs/files
cd ~/client-configs

we are actually going to omit these instructions, we have re-coded our batch file under client creation to avoid this issue:
# mkdir ~/keys
# cp ~/openvpn-ca/pki/ca.crt ~/client-configs/keys
# cp ~/openvpn-ca/pki/ta.key ~/client-configs/keys
# cp ~/openvpn-ca/pki/private/client1.key ~/client-configs/keys
# cp ~/openvpn-ca/pki/private/client1.crt ~/client-configs/keys

2) Next we need to copy base configuration from examples

cp /usr/share/doc/openvpn/sample/sample-config-files/client.conf ~/client-configs/base.conf

3) Open this file in your text editor

nano ~/client-configs/base.conf

4) and modify as following

remote server_IP_address 1194
# place your server address here
proto udp
# update with specified protocol
next uncomment (by removing leading semicolons)

user nobody
group nogroup

NB: If you are using CentOS, change the group from nogroup to nobody to match the distribution’s available groups
and comment out the lines because we place them directly in client’s config

#ca ca.crt
#cert client.crt
#key client.key

add these lines at the end of the file:

auth SHA256
key-direction 1

5) Next, we will create a simple script to compile our base configuration with the relevant certificate, key, and encryption files. This will place the generated configuration in the ~/client-configs/ files directory.

Note: to be consistent with the portion of this document above, I should really use ~ instead of /home/desktop in the section below. However that is how I run it:

nano ~/client-configs/make_config.sh

#!/bin/bash

# remember to run easyrsa build-client-full clientid nopass

# First argument: clientid

KEY_DIR=~/openvpn-ca/pki
OUTPUT_DIR=~/client-configs/files
BASE_CONFIG=~/client-configs/base.conf

cat ${BASE_CONFIG} \
<(echo -e ‘<ca>’) \
${KEY_DIR}/ca.crt \
<(echo -e ‘</ca>\n<cert>’) \
${KEY_DIR}/issued/${1}.crt \
<(echo -e ‘</cert>\n<key>’) \
${KEY_DIR}/private/${1}.key \
<(echo -e ‘</key>\n<tls-auth>’) \
${KEY_DIR}/ta.key \
<(echo -e ‘</tls-auth>’) \
> ${OUTPUT_DIR}/${1}.ovpn

make the file executable:

chmod 700 ~/client-configs/make_config.sh

6) Execute that file with client01 input parameter

Note you must first run the client creation from step 5 in the server setup. A repeatable procedure for client creation is as follows (using client02 as token):

cd ~/openvpn-ca

./easyrsa build-client-full client02 nopass

cd ~/client-configs

./make_config.sh client02

If everything went well, we should have a client02.ovpn file in our ~/client-configs/ directory

7) Now that file can be used on the client machine

sudo dnf install openvpn
sudo openvpn –config client02.ovpn

In theory, if you tag posts in your blog by category, those category titles should be offered as links that show a search list of matching posts, on the blog’s main page, in a section labeled “Categories.”

It turns out that, in order for this to actually work, you must first enter values under Settings | Optional | “Category base” and “Tag base” as below. Note I chose to use the value “categories” rather than the suggested value “topic” for both fields. Entering values in these fields will result in the links under Categories displaying a correct listing of the posts in each category:

Recently, I activated a very small VPS: 512MB RAM, 20GB SSD drive space. It ran WordPress well under Fedora 27. However, I encountered a problem with DNF refusing to update, exiting with a kernel panic on the executable.

Turns out that a virtual server image with minimal OS config can be born without a swap file. I found this link to be useful (note that for this subsystem Fedora 27 is close enough to CentOS 7, which itself is Fedora 19):

https://www.digitalocean.com/community/tutorials/how-to-add-swap-on-centos-7

 

Although my experience with web hosting servers goes back two decades, I have recently stopped hosting content for others, and no longer maintain DNS or email server infrastructure.

Given the trends of consolidation and cloud, what are the best practices to manage a digital footprint at the beginning of 2018? There are 4 fundamentals: domain registrar, dns server, email server, and web server. For bonus points, a test server.

Here is what I did, which is a strategy that would be suitable for a small business or personal presence on the Internet, or for a startup up to 10 or 20 people.

Domain registrar

I purchased the domain on which this is hosted from godaddy.com

DNS server

GoDaddy offers a terrific DNS control panel built-in — I encourage people to use it rather than giving up control of their DNS to a hosting ISP.

email server

A colleague continues to maintain an email server, and has been gracious enough to host my email. I would suggest outsourcing your email to a boutique ISP like integrationm.com, or a service like Gmail for Business.

web server

What works for me:

I leased a small Virtual Private Server (VPS) from Digital Ocean running Fedora Linux, and have configured it to host a few websites, including my personal blog using WordPress, and some vacation photos. Some people might want a server with cPanel in order to manage the server.

What you should consider as a hosted option:

Get a WordPress blog from a boutique ISP like integrationm.com or wordpress.com, or if you need a builder and commerce capabilities, squarespace.com to host the www.domain.com and domain.com for your site. Do not give the ISP control of your DNS, just ask them for the IP address and point to the hosts using your DNS control panel.

Test server

I have a test server at home, a Vista-class 64 bit dual core with 4GB RAM and a small SSD for the operating system, and a few big usb external drives formatted with ext4 and ntfs partions, running Fedora 27. I can supplement the limited resources of a very small VPS package by mounting, over the WAN, the file system on my home test server. Some of this can of course be done as a network mount like xfs or smb (windows style) over a VPN link. However, an even cleaner yet still secure approach is sshfs remote file system mount, which I predict will make many VPN setups redundant or greatly simplified in the future.

We may live in the age of the cloud but there is still room for individual expression and self-sufficiency. It’s never too late to reformat a piece of junk as a home test server.