A few times over the past year and a bit, we’ve had issues with high steal times on Slicehost, with the only cure being to move to another host server (and about an hour of downtime). The disk latency also had got worse, and we kept bumping up against the memory limits.

After a while discussing it, we finally decided to move to Linode for the following reasons:

• London based data center. Which is handy, as most of our customers are this side of the Atlantic. And we like giving our customers a snappy response
• Twice the memory for the same price.
• So far, the steal time has been at 0% and iowait time has been low too.
• The admin console is a bit more comprehensive too, but that wasn’t a reason for picking Linode

So far, things seems fine but I’ll keep an eye on it and report back if I have any issues.

It is better to use > log_file.log as that will allow the file system to free the space that is used by the open file.

df -h Shows how much disk space is available. The -h means that you get G for gigabytes, M for megabytes etc.

sudo du -k / | sort -n Shows which directories are biggest (starting at the root directory /). The biggest ones will be printed at the bottom of the list. The sudo allows du to check directories that your user doesn’t have permission for. To check your home directory only use du -k ~ | sort -n

du -ka | sort -n Shows the size of all the files and sub-directories of the current directory. The biggest ones at the bottom of the list.

watch 'du -ka | sort -nr' Shows the size of all the files and sub-directories of the current directory. The biggest ones at the top of the list. The list is updated every 2 seconds. This is useful for watching disk usage in a particular directory, such as an upload directory on a web server.

Making upgrading of web apps as simple as possible is one way to make sure the users/admins keep the software up to date.

Now I just wish that I set up a different theme for this site so it didn’t wipe my adverts out every upgrade… Oh well.

tail -f query.log | grep "thing I'm interested in" --color

Now you will see any new lines added to query.log that contain “thing I’m interested in” highlighted. Any other lines will be ignored.

In this post, I will cover what forward and reverse tunneling of ports through SSH is.  Why you would want to do it, and how to do it.  Port tunneling is also known as port forwarding.

Forward Tunneling

Port forwarding allows a specific port to be mapped between one computer and another.  A forward tunnel allows you to have a remote port appear as a local one.

This can be useful if you need to connect your MySQL query browser to a remote database, but the database only listens to localhost.  Although this section only talks about MySQL, it is applicable to almost any other piece of client/server software that uses specific ports.

To forward port 3306 (MySQL’s default port) to your local systems port 3306 use the following command:
ssh user@remote-hostname -L 3306:remote-hostname:3306

If you are running MySQL locally on port 3306, or want to use a different local port to the remote port, change the port before the first colon:
ssh user@remote-hostname -L 3305:remote-hostname:3306

Then you can connect to your remote MySQL server as if it was a local one.  Use the usual username, password and database name as you would to connect to the database, but change the host to 127.0.0.1 and the port to 3305.  You should now be able to access the remote database as if it was a local one.

Reverse Tunneling

Reverse tunneling allows you to have a port on a remote machine work as if it is a port on your local machine. It’s the exact opposite of a forward tunnel.
You could use this when working on code that needed to be accessed by a service that is outside of your firewall, but not ready to be deployed on a server. For example, developing a payment gateway for a shop, when the payment gateway needs to talk to the code that is on your development machine and you don’t want to (or can’t) make modifications to your firewall.

To reverse tunnel port 8080 on a remote server to port 80 on the local machine use the following command:
ssh username@remote-hostname -R 8080:127.0.0.1:80
This will let you, or anyone else, visit http://remote-hostname:8080 and see what is visible on http://127.0.0.1:80. In this case, your local web server is available to anyone who can see remote-hostname. If remote-hostname is a server available to the whole internet, the whole internet can see into your local machine that is usually hidden behind your firewall.

Some notes

• This guide is for OpenSSH, other implementations may vary.
• You will need to set GatewayPorts to yes in /etc/ssh/sshd_config (Don’t forget to restart sshd after the change).
• When you close your ssh session,  the port forwarding stops.

For more information, read the OpenSSH documentation and Frequently Asked Questions