Category Archives: OS

Backup your systems with urBackup

In addition to my ZFS snapshots I decided to implement a secondary backup system. I decided to land on urbackup for ease of use and, more importantly, it was easier to set up.

Server Install

Assuming a Cent-based system:

cd /etc/yum.repos.d/
sudo wget http://download.opensuse.org/repositories/home:uroni/CentOS_7/home:uroni.repo
sudo yum -y install urbackup-server
sudo systemctl enable urbackup-server
sudo systemctl start urbackup-server

Open up necessary ports for the server:

sudo firewall-cmd --add-port=55413-55415/tcp --permanent
sudo systemctl reload firewalld

By default urbackup listens on port 55414 for connections. You can change this to port 80 and/or 443 for HTTPS by installing nginx and having it proxy the connections for you.

sudo yum -y install nginx
sudo systemctl enable nginx
sudo setsebool -P httpd_can_network_connect 1 #if you're using selinux

Copy the following into /etc/nginx/conf.d/urbackup.conf (make sure to change server_name to suit your needs)

server {
        server_name backup;

        location / {
                proxy_pass http://localhost:55414/;
        }
}

Then start nginx:

sudo systemctl start nginx

You should then be able to access the urbackup console by navigating to the IP / hostname of your backup server in a browser.

Client Install:

Urbackup can use a snapshot system known as dattobd. You should use it if you can in order to get more consistent backups, otherwise urbackup will simply copy files from the host which isn’t always desirable (databases, for example)

Install dattobd (optional):

sudo yum -y update
# reboot if your kernel ends up being updated
sudo yum -y localinstall https://cpkg.datto.com/datto-rpm/repoconfig/datto-el-rpm-release-$(rpm -E %rhel)-latest.noarch.rpm
sudo yum -y install dkms-dattobd dattobd-utils

Install urbackup client:

TF=`mktemp` && wget "https://hndl.urbackup.org/Client/2.1.15/UrBackup%20Client%20Linux%202.1.15.sh" -O $TF && sudo sh $TF; rm $TF
#Select dattobd when prompted if desired

Configure Firewall:

sudo firewall-cmd --add-port=35621-35623/tcp --permanent
sudo systemctl reload firewalld

Once a client is installed, assuming they’re on the same network as the backup server, they will automatically add themselves and begin backing up. If they don’t show up it’s usually a firewall issue.

Restore

Restoration of individual files is easily done through the web console. If you have a windows system, restoring from an image backup is also easy.

Linux hosts

Recovery is trickier if you want to restore a Linux system. Install an empty system of same distribution. Give it the same hostname. Install the client as outlined above, then run:

sudo /usr/local/bin/urbackupclientctl restore-start -b last

Troubleshooting

If for some reason the client not showing up after removing it from the GUI: Uninstall & re-install client software

sudo /usr/local/sbin/uninstall_urbackupclient
TF=`mktemp` && wget "https://hndl.urbackup.org/Client/2.1.15/UrBackup%20Client%20Linux%202.1.15.sh" -O $TF && sudo sh $TF; rm $TF

Create & Mount disc images in Linux

When working with hard drives it is always a good idea to back the entire thing up before proceeding. I wanted to write down the procedure so I don’t keep forgetting it.

Create disc image

dd does the trick here.

sudo dd if=/dev/<drive device file> of=image.img bs=64M

If you wish to see the progress of the above dd command you can open up a separte window and issue the kill command

kill -USR1 `pidof dd`

Mount disc image read only

You can now disconnect the drive and work with its image instead (great for forensics or dealing with a dying drive.)

In later versions of Linux you can do this with losetup and partprobe.

sudo losetup -Pr -f <path to image file>
sudo losetup #find which loop device file corresponds with your image here
sudo mount -o ro /dev/<loopdevice>p<partition number> <mountpoint>

For example, this is what I did on my system for my aunt’s laptop (I was interested in the 2nd partition on her drive, the one containing Windows files)

sudo losetup -Pr -f susan-ssd.img
sudo losetup

NAME SIZELIMIT OFFSET AUTOCLEAR RO BACK-FILE DIO
/dev/loop0 0 0 0 1 /home/partimag/susan-ssd.img 0

sudo mount -o ro /dev/loop0p2 mount/

When you’re done, unmount the image and delete the image mapping:

umount <path to mount directory>
sudo losetup -d <loop file obtained earlier>

CentOS7, nginx, reverse proxy, & let’s encrypt

With the loss of trust of Startcom certs I found myself needing a new way to obtain free SSL certificates. Let’s Encrypt is perfect for this. Unfortunately SophosUTM does not support Let’s Encrypt. It became time to replace Sophos as my reverse proxy. Enter nginx.

The majority of the information I used to get this up and running came from digitalocean with help from howtoforge. My solution involves CentOS7, nginx, and the let’s encrypt software.

Install necessary packages

sudo yum install nginx letsencrypt
sudo firewall-cmd --add-service=http --permanent
sudo firewall-cmd --add-service=https --permanent
sudo systemctl enable nginx

Inform selinux to allow nginx to make http network connections:

sudo setsebool -P httpd_can_network_connect 1

Generate certificates

Generate your SSL certificates with the letsencrypt command. This command relies on being able to reach your site over the internet using port 80 and public DNS. Replace arguments below to reflect your setup

sudo letsencrypt certonly -a webroot --webroot-path=/usr/share/nginx/html -d example.com -d www.example.com

The above command places the certs in /etc/letsencrypt/live/<domain_name>

Sophos UTM certificates

In my case I had a few paid SSL certificates I wanted to copy over from Sophos UTM to nginx. In order to do this I had to massage them a little bit as outlined here.

Download p12 from Sophos, also download certificate authority file, then use openssl to convert the p12 to a key bundle nginx will take.

openssl pkcs12 -nokeys -in server-cert-key-bundle.p12 -out server.pem
openssl pkcs12 -nocerts -nodes -in server-cert-key-bundle.p12 -out server.key
cat server.pem Downloaded_CA_file.pem > server-ca-bundle.pem

Once you have your keyfiles you can copy them wherever you like and use them in your site-specific SSL configuration file.

Auto renewal

First make sure that the renew command works successfully:

sudo letsencrypt renew

If the output is a success (a message saying not up for renewal) then add this to a cron job to check monthly for renewal:

sudo crontab -e
30 2 1 * * /usr/bin/letsencrypt renew >> /var/log/le-renew.log
35 2 1 * * /bin/systemctl reload nginx

Configure nginx

Uncomment the https settings block in /etc/nginx/nginx.conf to allow for HTTPS connections.

Generate a strong DH group:

sudo openssl dhparam -out /etc/ssl/certs/dhparam.pem 2048

Create SSL conf snippets in /etc/nginx/conf.d/ssl-<sitename>.conf. Make sure to include the proper location of your SSL certificate files as generated with the letsencrypt command.

ssl_certificate /etc/letsencrypt/live/example.com/fullchain.pem;
ssl_certificate_key /etc/letsencrypt/live/example.com/privkey.pem;
ssl_dhparam /etc/ssl/certs/dhparam.pem;

Here is a sample ssl.conf file:

server {
        listen 443;

        ssl_certificate /etc/letsencrypt/live/<HOSTNAME>/fullchain.pem;
        ssl_certificate_key /etc/letsencrypt/live/<HOSTNAME>/privkey.pem;
        ssl_dhparam /etc/ssl/certs/dhparam.pem;

        access_log /var/log/<HOSTNAME.log>;

        server_name <HOSTNAME>;

        location / {
                proxy_set_header Host $host;
                proxy_set_header X-Real-IP $remote_addr;
                proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
                proxy_set_header X-Forwarded-Proto $scheme;

                proxy_pass http://<BACKEND_HOSTNAME>/;
        }
}

 

Redirect http to https by creating a redirect configuration file (optional)

sudo vim /etc/nginx/conf.d/redirect.conf
server {
	server_name
		<DOMAIN_1>
                ...
		<DOMAIN_N>;

        location /.well-known {
              alias /usr/share/nginx/html/.well-known;
              allow all;
	}
	location / {
               return 301 https://$host$request_uri; 
	}
}

 

Restart nginx:

sudo systemctl restart nginx

Troubleshooting

HTTPS redirects always go to the host at the top of the list

Solution found here:  use the $host variable instead of the $server_name variable in your configuration.

Websockets HTTP 400 error

Websockets require a bit more massaging in the configuration file as outlined here. Modify your site-specific configuration to add these lines:

# we're in the http context here
map $http_upgrade $connection_upgrade {
  default upgrade;
  ''      close;
}

server {     proxy_set_header Upgrade $http_upgrade;
    proxy_set_header Connection $connection_upgrade;
}

 

Fix FreeNAS multipath error with USB drives

I have a Mediasonic 4 bay USB drive enclosure plufgged into a FreeNAS system as a backup. This enclosure randomizes hardware ID information which  confuses FreeNAS. The main symptom of this is the error message about multipath not being optimal.

The way to fix this is do drop to a command line and destroy the mistaken multipath configuration (thanks to this site for the information)

For Freenas 9.x:

gmultipath destroy disk1 disk2 disk3
reboot

For FreeNAS Corral it’s a bit different. You have to brute force remove the kernel module for multipath (thanks to this site for the explanation)

mv /boot/kernel/geom_multipath.ko /boot/kernel/geom_multipath.ko.old

Keep in mind that you may have to do this after each update.

An additional problem is that for some reason GUI doesn’t see the drives but the OS does. If you want smart checking on those other drives you have to do a bit of hackery by creating a SMART job for the visible drives, then manually dropping into a shell to add the invisible ones. Below are my notes for when I did this (I’m not sure these changes survive a reboot)

sudo vi /etc/local/smartd.conf
/dev/da2 -a -n never -W 0,0,0 -m root -M exec /usr/local/www/freenasUI/tools/sma
 rt_alert.py -s S/(01|02|03|04|05|06|07|08|09|10|11|12)/../(1|2|3|4|5|6|7)/(00)

ps aux|grep smartd

kill -9 (pidof smartd)
sudo /usr/local/sbin/smartd -i 1800 -c /usr/local/etc/smartd.conf -p /var/run/smartd.pid

 

Install Linux on Chromebook Pixel 2 (Samus)

I’ve run crouton on my Chromebook Pixel 2  (2015, codename Samus) for some time now but I’ve found myself wanting more. Virtualbox, kernel access, graphics, and more don’t perform well in a chroot. Thankfully it’s actually pretty easy to dual boot Chrome OS and Linux on your chromebook thanks to chrx (pronounced “marshmallow”?)

Installation

The first part of installation is identical to setting up crouton:

  • Enter developer mode:
    Press ESC, Refresh, power simultaneously (when the chromebook is on)

    • Every time you power on the chromebook from now on you’ll get a scary screen. Press CTRL-D to bypass it (or wait 30 seconds)
    • If you hit space on this screen instead of CTRL+D it will powerwash (nuke) your data
      A scary screen will pop up saying the OS is missing or damaged. Press CTRL D, then press Enter when the OS verification screen comes up.
  • Wait several minutes for developer mode to be installed. Note it will wipe your device to do this.

Enable SeaBIOS:

Open up a shell (CTRL + ALT + T, shell, enter) and enter the following

sudo crossystem dev_boot_usb=1 dev_boot_legacy=1

and reboot.

Next, download and run the chrx script twice. The first run will partition and powerwash your system; the second run will actually install GalliumOS (or Ubuntu or Fedora) alongside ChromeOS.

cd ; curl -Os https://chrx.org/go && sudo sh go

reboot after partition, run shell again. You can specify a number of arguments to the go script; I wanted to use Cinnamon on Fedora so these are the ones I used:

cd ; curl -Os https://chrx.org/go && sudo sh go -d fedora -e cinnamon -r latest -H <hostname> -U <username> -Z <timezone>

Fedora took quite a long time to install (1 hour in my case.) Just let the script do its thing. Once complete you can reboot and press CTRL + D for chromeOS or CTRL + L for Linux.

After that, reboot into your new linux environment!

Cleaning up

There were a few samus-specific things I needed to do.

Locale

For some reason my locale was set to an African country.  Correct by doing this:  (thanks to here) I added SELinux commands because for some reason I was getting permission denied errors.

sudo setenforce 0
localectl set-locale LANG=en_US.utf8
sudo setenforce 1

Audio doesn’t work (no sound)

This issue stems from the fact that the sound card is not presented as the first available card. The system defaults to HDMI sound instead. Fortunately this page has instructions on how to fix this. If you’re running GalliumOS default you can follow the instructions from the link above. In my case I had to get a bit creative.

  1. Download the samus patches from here
    wget https://github.com/GalliumOS/galliumos-samus/archive/master.zip
  2. Extract subfolders inside said zip file to root directory
  3. Reboot
  4. run the following:
    1. cp -r /etc/skel/.config $HOME
      sudo samus-alsaenable-speakers
      sudo samus-touch-reset

Success! You can now dual boot between Full blown Linux and ChromeOS on your Chromebook Pixel.

Touchpad / touchscreen stop working after resume

Occasionally my touchpad and touchscreen stop responding after resuming from sleep. The galliumOS-samus fix mentioned above has a handy reset script that fixes this. Simply run:

sudo samus-touch-reset

and your touch functionality is restored. I bound this command to a key shortcut to make things easier.

Virtualbox won’t start

After installing virtualbox I got a strange error message when trying to start VMs:

Failed to load VMMR0.r0 (VERR_SUPLIB_OWNER_NOT_ROOT)

I found this mention saying that /usr has to be owned by root. Easy enough of a fix:

sudo chown root:root /usr/

Simple network folder mount script for Linux

I wrote a simple little network mount script for Linux desktops. I wanted to replicate my Windows box as best as I could where a bunch of network drives are mapped upon user login. This script relies on having gvfs-mount and the cifs utilities installed (installed by default in Ubuntu.)

#!/bin/bash
#Simple script to mount network drives

#Specify network paths here, one per line
#use forward slash instead of backslash
FOLDER=(
  server1/folder1
  server1/folder2
  server2/folder2/folder3
  server3/
)

#Create a symlink to gvfs mounts in home directory
ln -s $XDG_RUNTIME_DIR/gvfs ~/Drive_Mounts

for mountpoint in "${FOLDER[@]}"
do
  gvfs-mount smb://$mountpoint
done

Mark this script as executable and place it in /usr/local/bin. Then make it a default startup application for all users:

vim /etc/xdg/autostart/drive-mount.desktop
[Desktop Entry]
Name=Mount Network Drives
Type=Application
Exec=/usr/local/bin/drive-mount.sh
Terminal=false

Voila, now you’ve got your samba mount script starting up for every user.

Monitor your servers with phpservermonitor

I have a handful of servers and for years I’ve been wanting to get some sort of monitoring in place. Today I tried out php server monitor and found it was pretty easy to set up and use.

Download

The installation process was pretty straightforward.

  • Install PHP, mysql, and apache
  • Create database, user, password, and access rights for mysql
  • Download .tar.gz and extract to /var/www
  • Configure Apache site file to point to phpservermonitor directory
  • Navigate to the IP / URL of your apache server and run the installation script

The above process is documented fairly well on their website. I configured this to run on my Raspberry Pi 2. This was my process:

Install dependencies:

sudo apt-get install php5 php5-curl php5-mysql mysql-server

Configure mysql:

sudo mysql_secure_installation

Create database:

mysql -u root -p
create database phpservermon;
create user 'phpservermon'@'localhost' IDENTIFIED BY 'password';
 grant all privileges on phpservermon.* TO 'phpservermon'@'localhost'; 
flush privileges;

Extract phpservermon to /var/www and grant permissions

tar zxvf <phpservermon_gzip_filename> -C /var/www
sudo chown -R www-data /var/www/*

Configure php:

sudo vim /etc/php5/apache2/php.ini
#uncomment date.timezone and set your timezone
date.timezone = "America/Boise"

Configure apache:

sudo cp /etc/apache2/sites-available/default /etc/apache2/sites-available/phpservermon

#Modify /etc/apache2/sites-available/phpservermon server root to point to directory above, also add a ServerName if desired

sudo a2ensite phpservermon
sudo service apache2 reload

Configure cron (I have it check every minute but you can configure whatever you like)

*/1 * * * * /usr/bin/php /var/www/phpservermon/cron/status.cron.php

Navigate to the web address you’ve configured in apache and follow the wizard.

It’s pretty simple but it works! A nice php application to monitor websites and services.

 

Hot swapped disk missing in FreeNAS fix

I hot removed a malfunctioning drive in my FreeNAS unit recently. The problem is its replacement would not show up in available drives. Camcontrol devlist wouldn’t reveal the device, even after camcontrol rescan all.

I found from this site that another command exists – camcontrol reset. I found out which bus to reset (instead of resetting all of them) by looking at logs and noticing the scbus number. Once obtained, I ran the following commands (the last number refers to the bus my drive was on)

sudo camcontrol reset 10
sudo camcontrol rescan 10

That did the trick! The drive was suddenly visible by the FreeNAS system once more.

Fix vmware-view RDP issues in Linux

I’ve been using vmware view as a means to remote into my computers at work for some time now. An update to the linux client appears to have broken my ability to remote into work machines over the RDP protocol. This issue affects multiple distros.

The symptom is the fact that after you log into vmware-view and double click on a computer you wish to connect to over the RDP protocol, the screen flashes for a second and then takes you right back to where you started – no error message. Frustrating.

If you launch vmware-view in a console you get a little more insight into what’s going on:

RDP Client(10222): WARNING: Unknown -r argument
2017-02-26 14:06:53.817-07:00: vmware-view 7858| RDP Client(10222): 
2017-02-26 14:06:53.817-07:00: vmware-view 7858| RDP Client(10222): Possible arguments are: comport, disk, lptport, printer, sound, clipboard, scard

After much frustration I was able to combine documentation from vmware and freerdp in order to finally get the right combination of arguments to get things working again. I read that freerdp works better than rdesktop with this version, so I tried launching vmware-view with this option:

vmware-view --rdpclient="xfreerdp"

Progress – at least now the error message was different.

RDP Client(20799): [14:04:04:097] [20799:20803] [ERROR][com.freerdp.crypto] - certificate not trusted, aborting.

After more investigation the culprit turned out to be crypto negotiation. Since I’m already connected to the truste work VMware server, I don’t really care about certificate validation. This is what finally got me up and running. The key components are the rdpclient and the /cert-ignore options.

vmware-view --rdpclient="xfreerdp" --xfreerdpOptions="-wallpaper /sound:sys:alsa /cert-ignore"

You can include these options in your ~/.vmware/view-preferences config file so you don’t have to manually add all those switches:

echo 'view.rdpClient = "xfreerdp"
view.xfreerdpOptions = "-wallpaper /sound:sys:alsa /cert-ignore"' >> ~/.vmware/view-preferences

Finally RDP via vmware-view is working in Linux again.

Install wordpress on CentOS7 with nginx & caching

Lately I’ve become interested in the LEMP stack (as opposed to the LAMP stack.) As such I’ve decided to set up a wordpress site running CentOS 7, nginx, mariadb, php-fpm, zend-opcache, apc, and varnish. This writeup will borrow heavily from two of my other writeups, Install Wordpress on CentOS 7 with  SELinux and Speed up WordPress in CentOS7 using caching. This will be a mashup of those two with an nginx twist with guidance from digitalocean. Let’s begin.

Repositories

To install the required addons we will need to have the epel repository enabled:

yum -y install epel-release

nginx

Install necessary packages:

sudo yum -y install nginx
sudo systemctl enable nginx

Optional: symlink /usr/share/nginx/ to /var/www/ (for those of us who are used to apache)

sudo ln -s /usr/share/nginx/ /var/www

Open necessary firewall ports:

sudo firewall-cmd --add-service=http --permanent
sudo systemctl restart firewalld

start nginx:

sudo systemctl start nginx

Navigate to your new site to make sure it brings up the default page properly.

MariaDB

Install:

sudo yum -y install mariadb-server mariadb
sudo systemctl enable mariadb

Run initial mysql configuration to set database root password

sudo systemctl start mariadb
sudo mysql_secure_installation

Configure:

Create a wordpress database and user:

mysql -u root -p 
#enter your mysql root password here
create user wordpress;
create database wordpress;
GRANT ALL PRIVILEGES ON wordpress.* To 'wordpress'@'localhost' IDENTIFIED BY 'password';
quit;

php-fpm

Install:

sudo yum -y install php-fpm php-mysql php-pclzip
sudo systemctl enable php-fpm

Configure:

Uncomment cgi.fix_pathinfo and change value to 0:

sudo sed -i 's/\;\(cgi.fix_pathinfo=\)1/\10/g' /etc/php.ini

Modify the listen= parameter to listen to UNIX socket instead of TCP:

sudo sed -i 's/\(listen =\).*/\1 \/var\/run\/php-fpm\/php-fpm.sock/g' /etc/php-fpm.d/www.conf

Change listen.owner and listen.group to nobody:

sudo sed -i 's/\(listen.owner = \).*/\1nobody/g; s/\(listen.group = \).*/\1nobody/g' /etc/php-fpm.d/www.conf

Change running user & group from apache to nginx:

sudo sed -i 's/\(^user = \).*/\1nginx/g; s/\(^group = \).*/\1nginx/g' /etc/php-fpm.d/www.conf

Start php-fpm:

sudo systemctl start php-fpm

Caching

To speed up wordpress further we need to install a few bits of caching software. Accept defaults when prompted.

zend-opcache & apc:

Install necessary packages:

sudo yum -y install php-pecl-zendopcache php-pecl-apcu php-devel gcc
sudo pecl install apc

Add apc extension to php configuration:

sudo sh -c "echo '\

#Add apc extension 
extension=apc.so' >> /etc/php.ini"

Restart php-fpm:

sudo systemctl restart php-fpm

Varnish

Install:

sudo yum -y install varnish
sudo systemctl enable varnish

Configure nginx to listen on port 8080 instead of port 80:

sudo sed -i /etc/nginx/nginx.conf -e 's/listen.*80/&80 /'

Change varnish to listen on port 80 instead of port 6081:

sudo sed -i /etc/varnish/varnish.params -e 's/\(VARNISH_LISTEN_PORT=\).*/\180/g'

Optional: change varnish to cache files in memory with a limit of 256M (caching to memory is much faster than caching to disk)

sudo sed -i /etc/varnish/varnish.params -e 's/\(VARNISH_STORAGE=\).*/\1\"malloc,256M\"/g'

Add varnish configuration to work with caching wordpress sites:

sudo sh -c 'echo "/* SET THE HOST AND PORT OF WORDPRESS
 * *********************************************************/
vcl 4.0;
import std;

backend default {
 .host = \"127.0.0.1\";
 .port = \"8080\";
 .first_byte_timeout = 60s;
 .connect_timeout = 300s;
}
 
# SET THE ALLOWED IP OF PURGE REQUESTS
# ##########################################################
acl purge {
 \"localhost\";
 \"127.0.0.1\";
}

#THE RECV FUNCTION
# ##########################################################
sub vcl_recv {

# set realIP by trimming CloudFlare IP which will be used for various checks
set req.http.X-Actual-IP = regsub(req.http.X-Forwarded-For, \"[, ].*$\", \"\"); 

 # FORWARD THE IP OF THE REQUEST
 if (req.restarts == 0) {
 if (req.http.x-forwarded-for) {
 set req.http.X-Forwarded-For =
 req.http.X-Forwarded-For + \", \" + client.ip;
 } else {
 set req.http.X-Forwarded-For = client.ip;
 }
 }

 # Purge request check sections for hash_always_miss, purge and ban
 # BLOCK IF NOT IP is not in purge acl
 # ##########################################################

 # Enable smart refreshing using hash_always_miss
if (req.http.Cache-Control ~ \"no-cache\") {
 if (client.ip ~ purge) {
 set req.hash_always_miss = true;
 }
}

if (req.method == \"PURGE\") {
 if (!client.ip ~ purge) {
 return(synth(405,\"Not allowed.\"));
 }

 ban (\"req.url ~ \"+req.url);
 return(synth(200,\"Purged.\"));

}

if (req.method == \"BAN\") {
 # Same ACL check as above:
 if (!client.ip ~ purge) {
 return(synth(403, \"Not allowed.\"));
 }
 ban(\"req.http.host == \" + req.http.host +
 \" && req.url == \" + req.url);

 # Throw a synthetic page so the
 # request wont go to the backend.
 return(synth(200, \"Ban added\"));
}

# Unset cloudflare cookies
# Remove has_js and CloudFlare/Google Analytics __* cookies.
 set req.http.Cookie = regsuball(req.http.Cookie, \"(^|;\s*)(_[_a-z]+|has_js)=[^;]*\", \"\");
 # Remove a \";\" prefix, if present.
 set req.http.Cookie = regsub(req.http.Cookie, \"^;\s*\", \"\");

 # For Testing: If you want to test with Varnish passing (not caching) uncomment
 # return( pass );

# DO NOT CACHE RSS FEED
 if (req.url ~ \"/feed(/)?\") {
 return ( pass ); 
}

#Pass wp-cron

if (req.url ~ \"wp-cron\.php.*\") {
 return ( pass );
}

## Do not cache search results, comment these 3 lines if you do want to cache them

if (req.url ~ \"/\?s\=\") {
 return ( pass ); 
}

# CLEAN UP THE ENCODING HEADER.
 # SET TO GZIP, DEFLATE, OR REMOVE ENTIRELY. WITH VARY ACCEPT-ENCODING
 # VARNISH WILL CREATE SEPARATE CACHES FOR EACH
 # DO NOT ACCEPT-ENCODING IMAGES, ZIPPED FILES, AUDIO, ETC.
 # ##########################################################
 if (req.http.Accept-Encoding) {
 if (req.url ~ \"\.(jpg|png|gif|gz|tgz|bz2|tbz|mp3|ogg)$\") {
 # No point in compressing these
 unset req.http.Accept-Encoding;
 } elsif (req.http.Accept-Encoding ~ \"gzip\") {
 set req.http.Accept-Encoding = \"gzip\";
 } elsif (req.http.Accept-Encoding ~ \"deflate\") {
 set req.http.Accept-Encoding = \"deflate\";
 } else {
 # unknown algorithm
 unset req.http.Accept-Encoding;
 }
 }

 # PIPE ALL NON-STANDARD REQUESTS
 # ##########################################################
 if (req.method != \"GET\" &&
 req.method != \"HEAD\" &&
 req.method != \"PUT\" && 
 req.method != \"POST\" &&
 req.method != \"TRACE\" &&
 req.method != \"OPTIONS\" &&
 req.method != \"DELETE\") {
 return (pipe);
 }
 
 # ONLY CACHE GET AND HEAD REQUESTS
 # ##########################################################
 if (req.method != \"GET\" && req.method != \"HEAD\") {
 return (pass);
 }
 
 # OPTIONAL: DO NOT CACHE LOGGED IN USERS (THIS OCCURS IN FETCH TOO, EITHER
 # COMMENT OR UNCOMMENT BOTH
 # ##########################################################
 if ( req.http.cookie ~ \"wordpress_logged_in|resetpass\" ) {
 return( pass );
 }

 #fix CloudFlare Mixed Content with Flexible SSL
 if (req.http.X-Forwarded-Proto) {
 return(hash);
 }

 # IF THE REQUEST IS NOT FOR A PREVIEW, WP-ADMIN OR WP-LOGIN
 # THEN UNSET THE COOKIES
 # ##########################################################
 if (!(req.url ~ \"wp-(login|admin)\") 
 && !(req.url ~ \"&preview=true\" ) 
 ){
 unset req.http.cookie;
 }

 # IF BASIC AUTH IS ON THEN DO NOT CACHE
 # ##########################################################
 if (req.http.Authorization || req.http.Cookie) {
 return (pass);
 }
 
 # IF YOU GET HERE THEN THIS REQUEST SHOULD BE CACHED
 # ##########################################################
 return (hash);
 # This is for phpmyadmin
if (req.http.Host == \"pmadomain.com\") {
return (pass);
}
}

sub vcl_hash {

if (req.http.X-Forwarded-Proto) {
 hash_data(req.http.X-Forwarded-Proto);
 }
}


# HIT FUNCTION
# ##########################################################
sub vcl_hit {
 return (deliver);
}

# MISS FUNCTION
# ##########################################################
sub vcl_miss {
 return (fetch);
}

# FETCH FUNCTION
# ##########################################################
sub vcl_backend_response {
 # I SET THE VARY TO ACCEPT-ENCODING, THIS OVERRIDES W3TC 
 # TENDANCY TO SET VARY USER-AGENT. YOU MAY OR MAY NOT WANT
 # TO DO THIS
 # ##########################################################
 set beresp.http.Vary = \"Accept-Encoding\";

 # IF NOT WP-ADMIN THEN UNSET COOKIES AND SET THE AMOUNT OF 
 # TIME THIS PAGE WILL STAY CACHED (TTL), add other locations or subdomains you do not want to cache here in case they set cookies
 # ##########################################################
 if (!(bereq.url ~ \"wp-(login|admin)\") && !bereq.http.cookie ~ \"wordpress_logged_in|resetpass\" ) {
 unset beresp.http.set-cookie;
 set beresp.ttl = 1w;
 set beresp.grace =3d;
 }

 if (beresp.ttl <= 0s ||
 beresp.http.Set-Cookie ||
 beresp.http.Vary == \"*\") {
 set beresp.ttl = 120 s;
 # set beresp.ttl = 120s;
 set beresp.uncacheable = true;
 return (deliver);
 }

 return (deliver);
}

# DELIVER FUNCTION
# ##########################################################
sub vcl_deliver {
 # IF THIS PAGE IS ALREADY CACHED THEN RETURN A HIT TEXT 
 # IN THE HEADER (GREAT FOR DEBUGGING)
 # ##########################################################
 if (obj.hits > 0) {
 set resp.http.X-Cache = \"HIT\";
 # IF THIS IS A MISS RETURN THAT IN THE HEADER
 # ##########################################################
 } else {
 set resp.http.X-Cache = \"MISS\";
 }
}" > /etc/varnish/default.vcl'

Restart varnish & nginx:

sudo systemctl restart nginx varnish

Logging

By default varnish does not log its traffic. This means that your apache log will only log things varnish does not cache. We have to configure varnish to log traffic so you don’t lose insight into who is visiting your site.

Update 2/14/2017:  I’ve discovered a better way to do this. The old way is still included below, but you really should use this other way.

New way:

CentOS ships with some systemd scripts for you. You can use them out of the box by simply issuing

systemctl start varnishncsa
systemctl enable varnishncsa

If you are behind a reverse proxy then you will want to tweak the varnishncsa output a bit to reflect x-forwarded-for header values (thanks to this github discussion for the guidance.) Accomplish this by appending a modified log output format string to /lib/systemd/system/varnishncsa.service:

sudo sed -i /lib/systemd/system/varnishncsa.service -e "s/ExecStart.*/& -F \'%%{X-Forwarded-For}i %%l %%u %%t \"%%r\" %%s %%b \"%%{Referer}i\" \"%%{User-agent}i\"\' /g"

Lastly, reload systemd configuration, enable, and start the varnishncsa service:

sudo systemctl daemon-reload
sudo systemctl enable varnishncsa
sudo systemctl start varnishncsa

 


Old way:

First, enable rc.local

sudo chmod +x /etc/rc.local
sudo systemctl enable rc-local

Next, add this entry to the rc.local file:

sudo sh -c 'echo "varnishncsa -a -w /var/log/varnish/access.log -D -P /var/run/varnishncsa.pid" >> /etc/rc.local'

If your varnish server is behind a reverse proxy (like a web application firewall) then modify the above code slightly (thanks to this site for the information on how to do so)

sudo sh -c "echo varnishncsa -a -F \'%{X-Forwarded-For}i %l %u %t \"%r\" %s %b \"%{Referer}i\" \"%{User-agent}i\"\' -w /var/log/varnish/access.log -D -P /var/run/varnishncsa.pid >> /etc/rc.local"

Once configuration is in place, start the rc-local service

sudo systemctl start rc-local

WordPress

Download, extract, and set permissions for your wordpress installation (this assumes your wordpress site is the only site on the server)

wget https://wordpress.org/latest.zip
sudo unzip latest.zip -d /usr/share/nginx/html/
sudo chown nginx:nginx -R /usr/share/nginx/html/

Configure nginx

Follow best practice for nginx by creating a new configuration file for your wordpress site. In this example I’ve created a file wordpress.conf inside the /etc/nginx/conf.d directory.

Create the file:

sudo vim /etc/nginx/conf.d/wordpress.conf

Insert the following, making sure to modify the server_name directive:

server {
    listen 8080;
    listen [::]:8080;
    server_name wordpress;
    root /usr/share/nginx/html/wordpress;
    port_in_redirect off;


    location / {
        index index.php;
        try_files $uri $uri/ /index.php?q=$uri&$args;
    }

    location ~ \.php$ {
        fastcgi_pass unix:/var/run/php-fpm/php-fpm.sock;
        fastcgi_index index.php;
        fastcgi_param SCRIPT_FILENAME $document_root$fastcgi_script_name;
        include fastcgi_params;
    }

    error_page 404 /404.html;     
    location = /40x.html {
     }

    error_page 500 502 503 504 /50x.html;
    location = /50x.html {
    }
}

Restart nginx

sudo systemctl restart nginx

Configure upload directory

If you want users to upload content, then you will want to assign the http_sys_rw_content_t selinux security context for the wp-uploads directory (create it if it doesn’t exist)

sudo mkdir /usr/share/nginx/html/wordpress/wp-content/uploads
sudo chown nginx:nginx /usr/share/nginx/html/wordpress/wp-content/uploads
sudo semanage fcontext -a -t httpd_sys_rw_content_t "/usr/share/nginx/html/wordpress/wp-content/uploads(/.*)?"
sudo restorecon -Rv /usr/share/nginx/html/wordpress/wp-content/uploads

Run the wizard

In order for the wizard to run properly we need to temporarily give the wordpress directory httpd_sys_rw_content_t selinux context

sudo chcon -t httpd_sys_rw_content_t /usr/share/nginx/html/wordpress/

Now navigate to your new website in a browser and follow the wizard, which will create a wp-config.php file inside the wordpress directory.

W3 Total Cache

Once your wordpress site is set up you will want to configure it to communicate with varnish. This will make sure the cache is always up to date when changes are made.

Install the W3 Total Cache wordpress plugin and configure it as follows:

opcode-database-object

Opcode cache: Opcode:Zend Opcache

Database cache: Check enable, select Opcode: Alternative PHP Cache (APC / APCu)

Object cache: Check enable, select Opcode: Alternative PHP Cache (APC / APCu)

Fragment cache: Opcode: Alternative PHP cache (APC / APCu)

proxy

Reverse Proxy: Check “Enable reverse proxy caching via varnish”
Specify 127.0.0.1 in the varnish servers box. Click save all settings.

Wrapping up

Once your site is properly set up, restore the original security context for the wordpress directory:

sudo restorecon -v /usr/share/nginx/html/wordpress/

Lastly restart nginx and varnish:

sudo systemctl restart nginx varnish

Success! Everything is working within the proper SELinux contexts and caching configuration.

Troubleshooting

403 forbidden

I received this error after setting everything up. After some digging I came across this site which explained what could be happening.

For me this meant that nginx couldn’t find an index file and was trying to default to a directory listing, which is not allowed by default. This is fixed by inserting a proper directive to find index files, in my case, index.php. Make sure you have “index index.php” in your nginx.conf inside the location / block:

    location / {
        index index.php;
    }

 Accessing wp-admin redirects you to port 8080, times out

If you find going to /wp-admin redirects you to a wrong port and times out, it’s because nginx is forwarding the portnumber. We want to turn that off (thanks to this site for the help.) Add this to your nginx.conf:

  port_in_redirect off;