While developing Leanservers we experienced the same exact issue. So we thought of creating one of our less known and coolest features: backups.

Behind that naive name a killer feature is hiding:

• frequent: our backups run daily;
• automatic: no action needs to be taken by the end user other than signing up;
• full: each of our backups is full and independent from each other which minimizes the risk of data loss. It’s not fun when you have an incremental backup split in 10 pieces and one of the pieces gets corrupted: automatically all the other parts become useless. And that is your precious data.
• offsite: we use backup servers that are located remotely (i.e. thousands of kilometers from the servers that hold user account data) on different datacenters run by different companies and all these machines do is… backups. Yes, they are dedicated.
• don’t hinder performance: thanks to having this remote machines, all the heavy IO happens somewhere else. Our user account data is periodically synchronized securely over the Internet with our backup servers. Folder and databases.
• accessible anytime through a browser: we keep the latest 30 full backups accesible through our control panel. You just need a browser. This can be used as a version control system giving the possibility of reverting back to any point in time within the last month of work. Even on databases.

If you are thinking that this is the same as having a plugin installed on your CMS of choice (WordPress, Drupal, Joomla, etc…) that runs periodically and creates a full backup and keeps it somewhere and/or sends it to you through email… sorry, it is not. The action of performing backups requires a significant amount of server effort. Which translates into slower websites. Additionally, keeping backups on the same device where your data resides is looking for trouble. Hard disks fail. We hate it, but they do.

At the moment this article is being written, Leanservers has infrastructure in Canada, USA, Germany and Malaysia. Sites hosted with us are currently being served from Canada or Malaysia and backups are being stored offsite on USA and Germany. That is pretty remote. And we are expanding continuously.

The data of a user that has chosen to go for backups is stored four times: server where his account is hosted, replicated fresh copy on USA, archived backup on USA, duplicate of archived backup on Germany. Without going into too much (boring) detail, that’s a lot of redundancy! Possibilities of having his data lost are virtually zero.

Those multiple levels of redundancy on different and remote locations are the main reason that makes us charge for backups: the infrastructure needs to be supported. Following our philosophy we prefer to allow our customers to chose what exactly they want to use and only pay for that. No grey areas.

We also experienced issues with other hosting companies in the past that are supposed to have backups but you never know until the day you really need them… and you (sadly) discover that they are not in place. And that makes you loose all your data.

Here at Leanservers we have extensive experience with VMware, Xen/Citrix, VirtualBox and OpenVZ; but our favorite beast is undoubtedly OpenVZ.

You can benefit similarly using any other virtualization technologies but OpenVZ is the one that works best for us primarily because of its amazing performance that allows you to run virtual machines at virtually native speeds. Its only limitation fits us perfectly: it can only run Linux OS.

Let me quote the excellent explanation of the different virtualization techniques from OpenVZ.org:

There are several kinds of virtualization techniques which provide similar features but differ in the degree of abstraction and the methods used for virtualization.

1. Virtual machines (VMs)

Virtual machines emulate some real or fictional hardware, which in turn requires real resources from the host (the machine running the VMs). This approach, used by most system emulators, allows the emulator to run an arbitrary guest operating system without modifications because guest OS is not aware that it is not running on real hardware. The main issue with this approach is that some CPU instructions require additional privileges and may not be executed in user space thus requiring a virtual machines monitor (VMM), also called a hypervisor, to analyze executed code and make it safe on-the-fly. Hardware emulation approach is used by VMware products, VirtualBox, QEMU, Parallels and Microsoft Virtual Server.

2. Paravirtualization

This technique also requires a VMM, but most of its work is performed in the guest OS code, which in turn is modified to support this VMM and avoid unnecessary use of privileged instructions. The paravirtualization technique also enables running different OSs on a single server, but requires them to be ported, i.e. they should «know» they are running under the hypervisor. The paravirtualization approach is used by products such as Xen and UML.

3. Virtualization on the OS level, a.k.a. containers virtualization

Most applications running on a server can easily share a machine with others, if they could be isolated and secured. Further, in most situations, different operating systems are not required on the same server, merely multiple instances of a single operating system. OS-level virtualization systems have been designed to provide the required isolation and security to run multiple applications or copies of the same OS (but different distributions of the OS) on the same server. OpenVZ, Virtuozzo, Linux-VServer, Solaris Zones and FreeBSD Jails are examples of OS-level virtualization.

Short comparison

The three techniques differ in complexity of implementation, breadth of OS support, performance in comparison with standalone server, and level of access to common resources. For example, VMs have wider scope of usage, but poorer performance. Para-VMs have better performance, but can support fewer OSs because one has to modify the original OS.
Virtualization on the OS level provides the best performance and scalability compared to other approaches. Performance difference of such systems can be as low as 1…3%, comparing with that of a standalone server. Virtual Environments are usually also much simpler to administer as all of them can be accessed and administered from the host system. Generally, such systems are the best choice for server consolidation of same OS workloads.

So, why using virtualization is so beneficial? Basically it allows you to do a bunch of operations safely, such as:

• Backup and clone a server without stopping it.- in general, there are ways on most virtualization platforms of doing a full backup of a running machine without stopping it. This is really a piece of cake using OpenVZ and this backup can be spawn as a clone with different name, resource allocation and IP address(es) easily.
• Creating a checkpoint/snapshot that allows you to revert to it if something bad happens.- one of the most common problems for system administrators across the globe is when and how to update the system; should it be automatically or manually? Supervised or unattended? Virtualization allows you to easily and quickly take a snapshot of a running system that can be reverted to in case an update goes terribly wrong.
• Add/remove/edit IPs easily with no downtime.- it’s just a pleasure to add/remove IP addresses to a server running OpenVZ: fast,reliable, no service restart required (it’s all done automatically for you).
• Modify the resource allocation and quotas dynamically.- this is especially useful if, let’s say, you are consolidating servers and are running a monitoring system together with your DNS server and your monitoring system goes crazy and starts eating up all the CPU… Well here is virtualization to rescue you, set limits properly and sleep tight at night knowing that secondary services won’t eat up all the resources letting your main boys starve to death.
• Faster reboot times.- this might be a OpenVZ specialty but… When you reboot a OpenVZ container it’s way faster than normal since is not that the whole machine is rebooting; only the services need to be restarted, pretty much. You can cut our restart times from minutes to seconds. This is awesome when you need to reboot a machine that’s so critical that cannot go down… So instead of waiting 5 minutes you wait 15 seconds.

Even though all our services run on dedicated servers we use OpenVZ as an extra layer of management to make day to day operations easy, fast and reliable. We need to spawn a new node in Europe? No problem, we clone and relocate. Our mail server needs a critical update? Easy, we just create a checkpoint and troubleshoot. There has been a new critical bug reported on production? Fine, clone the server and run it somewhere else safe an quietly where our techies can take a deep look at it.

It was a tough decision because although we found several other places very well suited for Leanservers we had to pick just one. So we finally did.

The new facilities are located in Hong Kong. A very interesting city-state, quoting from Wikipedia:

As one of the world’s leading international financial centres, Hong Kong has a major capitalist service economy characterized by low taxation and free trade, and the currency, Hong Kong dollar, is the eighth most traded currency in the world.[30] The lack of space caused demand for denser constructions, which developed the city to a centre for modern architecture and the world’s most vertical city.[31][32] The dense space also led to a highly developed transportation network with public transport travelling rate exceeding 90 percent,[33] the highest in the world.[34] Hong Kong has numerous high international rankings in various aspects. For instance, its economic freedom, financial and economic competitiveness, quality of life, corruption perception, Human Development Index, etc., are all ranked highly. Citation.

We are at the moment finalizing the setup and beginning our thorough testing to make sure everything is as it should. Once that is complete we will open to the public, hopefully by the end of 2011 or beginning of 2012

Feel free to contact Leanservers Support as early as now if you need a fast NGINX shared account in Hong Kong. We will be happy to assist you with the process.

After several months of full dedication we have been able to pull together a major milestone for the Leanservers project.

We engaged a design studio that sat and talked to us; they understood what our product is about and the philosophy of the company. Then they designed.

The results are substantial; the new look and feel is slicker and leaner. Text can be read comfortably and the images fit perfectly. Everything has been thoroughly thought. Nothing is random.

Our new slider provides an overview of our differentiators: affordability, quality, no-contracts and modularity. This slider has been easily implemented thanks to Drupal’s versatility and robustness. It’s a pleasure to develop new features with such an advanced framework.

We have added and polished a number of user specific features, such as: My Account, My Details, My Usage, My Domains, My Backups and added shortcuts to backend controlling pages: Webmail, Email Admin, DNS Admin and phpMyAdmin. All of these are now conveniently accessible through our elegant LEANMENU, a JQuery based user menu. Extremely slick.

It is now possible to know how many webroots we have, how much bandwidth we’ve used so far, manage our passwords, import unlimited sized MySQL databases and a significant number of other things that deserve a blog post on their own.

One of our strengths is modularity: you can increase or decrease the level of services with us anytime. Plus we bill you monthly. We know that some people (specially companies) prefer to pay yearly. That is also possible. We always wondered why people had to commit for 1 or 2 years of hosting when they didn’t know whether they would need it for that period of time.

All our payments are being managed at the moment through 2checkout. We don’t have access to any credit card information. They do it for us. We can simply stop recurring payments and refund money. That’s pretty much all. They take care of passing the fraud checks as well.

We always feel proud of letting people know that our servers have an average of 4% of CPU utilization. And that is where our clients host their sites. The load is minimal.

Our provisioning system deserves some air time as well. Everything is automated, there is no human intervention that’s why it’s so fast. Including the decision on where to create the account. Automatic.

Soon after I started with my assignment I was able to understand that NGINX is a web server on steroids. I would say that its main functionality is serving static files via http or https protocols but it has a bunch of other amazing features which make it unique because of the incredible performance that it is able to pull off.

That is the time when I absolutely fell in love with NGINX. Some of the results from my tests showed NGINX consuming several thousand time less memory than the world’s most popular web server to date: Apache.

A few weeks after I realized how beneficial NGINX would be for a web hosting provider where resource utilization is absolutely key. A system that is able to carry thousands of times more load would be just fantastic. It is important to understand as well that this is not just better for the provider; it also is a great advantage for the client. Today’s hosting provider scene is packed with companies that base their offer in quantity, not quality. Keeping similar users per machine ratios as any other traditional hosting should be thousands times faster. So let’s assume that it is only a few times faster since the world is not perfect.

So I began looking for NGINX based shared web hosting providers. But I did not find any. That is the time we started conceptualizing Leanservers.

Today Leanservers is truly a fast and reliable NGINX based web hosting. Just the way we thought it should be. We had an idea and removed the fat.

Technical Perspective

In a nutshell, NGINX is excellent in doing most of the things that people need; other web servers out there have more features but… do you really need those? The most used and common features are included: SSL, content compression (gzip), rewrite rules, reverse proxy, server side includes, load balancing, WebDAV, access control, advanced logging capabilities, fast-cgi… And more coming since there is a significant number of third-party community contributed modules under active developed: NGINX 3rd Party Modules.

One of its main strengths is its event-driven architecture which allows to scale extremely well under heavy load making use of Linux’ (e)poll system calls. This allows NGINX to handle more than 10,000 clients simultaneously, addressing the C10k problem.

Another big benefit of using NGINX is how simple its configuration is, allowing a new user to be up and running in minutes.

Finally I should mention that if you run NGINX you really never have to stop it even for upgrading it!!! Your can be serving thousands of requests per second while you compile a newer version of the source code. After you are done you can gracefully restart NGINX into the newer version.

References

• Nginx: the High-Performance Web Server and Reverse Proxy
• NGINX site
• Why use NGINX?
• Using epoll() For Asynchronous Network Programming
• Apache Web Server Project

Today I would like to focus on the most outstanding features of its Control Panel. Why? Well, it has been specifically developed for its platform so there is nothing superfluous. No fat. Everything is absolutely lean and centered on usability and performance, taking on average 2 seconds to completely render any page.

At the time this post is being written, Leanservers’ Control Panel features the following pages: My Account, My Details, My Usage and My Backups.

My Account page

This page gives a snapshot of what services are currently enabled for a user account. Given the modular nature of Leanservers it can be very useful to see what you have and what you could have if you need.

The screenshot shows the page rendered on Google Chrome running on Mac OS X, but it can be opened on any major browser/system.

My Details page

This page allows an authenticated Leanservers user to edit her email address and passwords.

The passwords section opens up the possibility of changing any of the 3 system passwords independently: Control Panel, SFTP or MySQL. By clicking on any of the “change” links the user will be taken to a sub-page where the specific action(s) can be fulfilled. The next screenshot shows the MySQL Password Change sub-page:

The page shown includes an important reminder for the user because modifying the database password may break any running applications. This can be prevented by making sure any application that connects to the database uses the new password.

My Usage page

This page contains detailed information of the account’s disk and bandwidth usage:

The total used disk space is computed by adding up the user’s home folder and the MySQL database size. It can be clearly viewed on the upper part of “My Usage” page along with the maximum allowed for this particular user and the percentage consumed so far.

The bandwidth usage can be divided according to its source into: HTTP, SFTP and PHPMAIL. Leanservers system gathers fine grained data which allows the backend servers to generate such a detailed report. A user can easily understand how the bandwidth is being utilized and even report cases of malicious intrusions into the system which typically manifest as unexpected increases on PHPMAIL traffic (which is typically should be fairly low). Any other unexpected traffic increases should be reported to Leanservers Support who will be happy to help.

Leanservers is able to keep track of each of the emails that the running applications send and link them to the generating user account. This is a very powerful tool to prevent spammers from abusing the system since all the data is being monitored 24 hours a day, 7 days a week.

The most important value of this page (“USED”) displays in nice green when the usage is under 80% but turns into yellow or red if it goes over certain threshold (80% – 90% yellow, above 90% red). This makes it easy for the user to anticipate the need to increase the monthly bandwidth usage allowance. As stated on Leanservers site, any service can be enabled/disabled/modified anytime, as you see fit. No commitment.

My Backups page

This page allows a user to access one of Leanservers’ most wanted features: automatic daily full backups:

Leanservers takes a full daily backup of the files and all databases of each user and presents it on a friendly way through this page. Obviously the backup files are only accessible from the account they belong to, after the user has been authenticated.

This feature is very powerful because it delivers version control over files and databases and piece of mind for the user. A Leanservers user will never forget to do a backup… because the system will automatically do it on her behalf.

Related Links

• Leanservers homepage
• NGINX Web Server
• PHP’s mail function