EDIT: If you’re fairly certain you’re not suffering from an xml-rpc attack, check out our article about debugging Wordpress database errors.

EDIT: We now have a community article covering xml-rpc attacks!

Hey there,

As @nestchris said, we have started seeing an influx of brute-force amplification attacks, both on our service and the rest of the internet at large. These attacks exploit the XML-RPC functionality in WordPress, as described on Securi.net

You can see if you are being taken down by this attack by running one of the following two commands on your droplet, depending on which webserver you’re running:

Apache:
grep xmlrpc /var/log/apache2/access.log
Nginx:
grep xmlrpc /var/log/nginx/access.log

If you get any results there, it should be easy to fix this issue. If you’ve spun up a WordPress One-Click install droplet since Dec 2015, there’s a small config file which you can enable to block such attacks. We don’t enable it by default because it breaks some commonly used plugins (like Jetpack). Here are the commands you can run:

sudo a2enconf block-xmlrpc
sudo service apache2 restart

Alternatively, Jetpack has functionality to block these requests without impacting other plugins. Check out their Protect module.

If you’re running a WordPress droplet older than Dec 2015 and you don’t want to use the Jetpack plugin from Automattic, you’ll need to configure the droplet manually to block requests to xmlrpc.php:

Apache2 (goes inside your VirtualHost config):

<files xmlrpc.php> 
order allow,deny 
deny from all 
</files>

Nginx (goes inside your server{} block):

location /xmlrpc.php { 
deny all; 
}

Alternatively, if you are not having this issue due to an XML-RPC brute force attack, your Droplet might not have enough memory to deal with your traffic reliably. We recommend users run WordPress on a 1GB droplet or larger.

In order to help reduce your memory usage, you’d first want to check for any plugins using a large amount of memory. There are a lot of memory hungry plugins out there! You may also want to optimize your configuration files.

Below are some great resources that we commonly recommend in support tickets, explaining how to do so, depending on your setup:

Apache:
https://www.digitalocean.com/community/tutorials/how-to-optimize-apache-web-server-performance
https://github.com/gusmaskowitz/apachebuddy.pl

Nginx/php-fpm:
https://www.digitalocean.com/community/tutorials/how-to-optimize-nginx-configuration
http://blog.chrismeller.com/configuring-and-optimizing-php-fpm-and-nginx-on-ubuntu-or-debian

MySQL:
http://mysqltuner.com/
http://www.percona.com/blog/2014/01/24/mysql-server-memory-usage-2/

If none of this works, you may have no choice but to resize to a larger sized droplet that can handle your traffic.

Best,
Eris
Platform Support Specialist

As we know, a shared host like Bluehost has more direct control over your server. On DO, you are required to maintain the server, and that means things that you might not need to deal with on other hosts come up here.

I was made aware of this exploit by the DO team.
https://blog.sucuri.net/2015/10/brute-force-amplification-attacks-against-wordpress-xmlrpc.html

After implementing the apache.config solution, my problems stopped.

[Apache VirtualHost Config]
<VirtualHost>
…
<files xmlrpc.php>
order allow,deny
deny from all
</files>
</VirtualHost>

This attack is affecting all wordpress sites, but if you were using another host before, you may not have noticed. In my experience with Bluehost, I have had them suspend my account because I was attacked by comment spam, I suspect something similar to that would have occurred in this instance as well.

Unfortunately DO has been very slow to address this issue with the community as it seems everyone is in the dark about it.

I have had the same problem and its not a fault of digital ocean. A little bit knowledge in server maintenance is what is actually required.

Ubuntu 14.04

Check if your mysql is running
mysqladmin -u root -p status

and then enter your mysql root password

if it’s running properly u get a message
Uptime: 4 Threads: 1 Questions: 62 Slow queries: 0 Opens: 51 Flush tables: 1 Open tables: 45 Queries per second avg: 15.500

or an error message

mysqladmin: connect to server at ‘localhost’ failed
error: 'Can’t connect to local MySQL server through socket ’/var/run/mysqld/mysqld.sock’ (2)’
Check that mysqld is running and that the socket: ’/var/run/mysqld/mysqld.sock’ exists!

This error message is the reason for you to get Error establishing a database connection

fix this by starting the mysql with the command
sudo start mysql

I hope my solution helps you in fixing your issue.

I am not a real tech guy, but I was able to fix our website, www.clienteleitsolution.com, by turning off the power of the droplet and turning it on.

please tired all week my site falls due this connection error will not change cloud service more

Problem solve with this code.. Very simple :)

sudo start mysql

Can someone recommend a good Wordpress host because this is a fucking disaster.

I Have done all of these things; spoke to support who kindly suggested it ; and i am STILL getting this issue. I am thinking about leaving DO because this is a business I am running and random error establishing a connection shouldn’t be happening, especially if I go to sleep and find out at 1AM it went down and didn’t find out until I got up at 7AM! This has been going on for now three months. I also upgraded and I STILL have the issue. Please DO , fix it, I have tried on my end and it is STILL happening. I am frustrated about it. I have opened a support ticket and sill no fix.

I can not solve this error somehow, I tried almost every method. I started having this problem on better performance (2 cpu - 2 GB Ram) droplets also..

If you’re faced with the excruciating Database Connection Error on your WordPress website, there are a number of factors which could be causing it. Perhaps you overlooked something in trying to fix it. Also check out this blog post which shares some good tips to overcome the Database Error. URL: Error establishing a database connection in WordPress

Having the same issue now as well. Worked fine for about 3 weeks now getting this issue multiple times per day and restarting the whole server works but is not a solution… Unreal that this hasn’t been solved by now.

My problem was Mysql user privileges. My user didn’t have database privileges.
This simple piece of code helped me. Try this before anything.

<?php
$link = mysql_connect('localhost', 'root', 'password');
if (!$link) {
die('Could not connect: ' . mysql_error());
}
echo 'Connected successfully';
mysql_close($link);
?>

me too… i don’t know what happen, when i try to open my website, the text “Error establishing a database connection” was there… hope there are some support from DO.

Support forums where amateurs try to fix supplier problems are an excellent way to cut support costs. What a smart industry paradigm.

Digital Ocean posted this helpful article to hell stop XML-RPC attacks, that can cause errors like mentioned on this thread.

I installed Jetpack (and turned on “Protect” as recommended as the Method 1) and it appears to be working so far.

Link to the helpful article: https://www.digitalocean.com/community/tutorials/how-to-protect-wordpress-from-xml-rpc-attacks-on-ubuntu-14-04

This article helped me finally fix this problem: http://hubpages.com/technology/DigitalOcean-Error-Establishing-a-Database-Connection-Finally-Solved

This error can also be triggered because of a DOS attack, please make sure you are not the victim of this attack before investigating for other causes. Learn more here

Solved.
Every time i did anything within the Wordpress admin, like adding a plugin it would topple my server. The database connection would be stopped for about 2minutes after any update and i’d have to wait for the page I was editing to reappear again. Drove me nuts.

After powering off the server and resizing it from 512mb to 1gb I can now manage 2 Wordpress sites with ease without it falling over every 5 mins when editing. On one website, I have woocommerce and buddy press running, which i’ve found is too much for the 512mb server setup.

As i’m new to managing my own server, I underestimated the server requirements needed for managing Wordpress websites & all the hungry plugins.

I’m relieved I found a solution (albeit one requiring more spend on the server set-up) In the process of elimination I added more memory for mysql, checked many log files for attacks and errors. Added additional defence against brute force attacks.

Here’s my error logs from this exact issue:

170503 7:58:36 [Note] Plugin ‘FEDERATED’ is disabled.
170503 7:58:36 InnoDB: The InnoDB memory heap is disabled
170503 7:58:36 InnoDB: Mutexes and rw_locks use GCC atomic builtins
170503 7:58:36 InnoDB: Compressed tables use zlib 1.2.8
170503 7:58:36 InnoDB: Using Linux native AIO
170503 7:58:37 InnoDB: Initializing buffer pool, size = 128.0M
InnoDB: mmap(137363456 bytes) failed; errno 12
170503 7:58:37 InnoDB: Completed initialization of buffer pool
170503 7:58:37 InnoDB: Fatal error: cannot allocate memory for the buffer pool
170503 7:58:37 [ERROR] Plugin 'InnoDB’ init function returned error.
170503 7:58:37 [ERROR] Plugin 'InnoDB’ registration as a STORAGE ENGINE failed.
170503 7:58:37 [ERROR] Unknown/unsupported storage engine: InnoDB
170503 7:58:37 [ERROR] Aborting
170503 7:58:37 [Note] /usr/sbin/mysqld: Shutdown complete

You can use this script to restart your server whenever it gives database connection error.
https://github.com/Gurnoor30196/Launch-Mysql

Nothing helped me guys. The only solution i got, power off the droplet and power on again. It will solve your problem. I got result.