There has been a lot of discussion in the hosting community lately about backups and the protection of customer data. This follows a large budget host losing entire servers worth of customer data after a failure and the very public collapse of the popular blog hosting provider Journalspace. Neither of these hosts had a proper backup strategy in place.
Rochen originally launched our Rochen Vault enterprise managed backup platform in May 2007 and it is included as a free of charge service with all of our shared and reseller hosting plans. It can also be added to full managed servers as well for a small monthly fee.
One of the most popular pre-sales questions we receive is: “what is different about Rochen Vault to other backup solutions?”
I hope I can help fill you in on Rochen Vault and why we think it is such a revolutionary solution through this blog post. Firstly, I will start by running through a few of the other backup solutions that are on the market and commonly used by many hosts and then explain Rochen Vault.
File Backups to a Secondary Drive
Most web hosting providers use low-cost backup solutions where your data is simply backed up once daily to a secondary drive in each of their servers. This works by simply copying your account files from one location to another on the server often using a free tool such as rsync. While this is a good fall back solution it is not a good primary backup method for several reasons: –
a) your data is backed up to a drive in the same physical server so if the server is fried by a power surge, brownout or another catastrophic issue the chances are the backup hard disk is going to be destroyed along with the primary.
b) with a simple disk-to-disk file copy there are no integrity checks on the backups to ensure they are running properly and they are not corrupted.
c) in the vast majority of cases there will be less than a handful of diffrent points in time where your data can be restored from.
d) copying large amounts of files can take a lot of time and if you are hosted on an oversold server then it can raise the server loads for many hours slowing down your website and impacting performance while the backups run.
e) if the entire server needs to be restored in a disaster recovery scenario then the operating system (OS) first has to be re-installed, the control panel system has to be re-installed and then finally services like Apache and MySQL need to be re-configured. This all has to occur before a single file can be restored from backups.
Pros: Easy to implement, low cost solution
Cons: Low level of data protection, no integrity checks on backed up data, very few number of restore points, slow disaster recovery
Rochen Position: Here at Rochen we backup all of your data to a secondary drive in our servers once nightly. This is only used as a secondary backup system and a fall back to Rochen Vault though.
File Backups to a Remote Server
Some hosts are now taking the initiative of further protecting customer data by copying files with a tool like rsync to a remote server instead of to a secondary hard disk. While this is a good step to take, as it provides additional protection, it is still far from perfect. There are still no integrity checks on the data, very few points in time that you can restore to, slow restores in a disaster recovery situation and on oversold hosts it will still create load problems.
Remote backups also open another can of worms in that you are unlikely to be able to access the backup data on the remote server directly yourself so you will need to open a support ticket with the provider to restore for you. With some less than stellar hosts this is not as simple as it sounds. You could be waiting hours for what should be a simple file restore but is keeping your site offline until it is completed.
Pros: Easy to implement, provides additional protection over local disk-to-disk backups
Cons: No integrity checks on backed up data, very few number of restore points, slow disaster recovery restores, can often take a long time to get data restored
Rochen Position: Here at Rochen we push all of your data to an off-site remote location twice weekly. This along with local disk-to-disk backups is only used as a secondary backup system and a fall back to Rochen Vault though.
RAID is great but it is *NOT* a backup solution!
RAID is a great system as it mirrors your data over multiple hard disks but it is not a backup solution on its own. The reason RAID is not a backup solution is that it mirrors your data across all of the hard disks in the array in real-time. This means that if you mistakenly delete a file, your account is compromised resulting in files being wiped out or if simple data corruption occurs across the server then you are sods out of lock as the RAID system will replicate this to all drives in the array instantly with no way of recovering the previous state.
Don’t get me wrong. As I said above I think RAID is a fantastic and something all hosts should be running but it should not be mistaken for a backup solution. This is the mistake Journalspace made as they thought they had a solid backup strategy in place by running RAID and mirroring their hard disks but they ended up losing everything and their business has subsequently collapsed. Where RAID can help though is if a hard disk suffers a failure it can be swapped out with minimal to no downtime without any need to restore the server from backups. It is better to think of RAID as a redundancy system to improve site availability than as a backup platform.
Pros: Can improve site uptime in the event of a hard disk failure, can help improve disk throughput performance when running RAID10.
Cons: It is not backup solution, can be costly to implement if using RAID10 and/or Serial Attached SCSI (SAS) hard disks
Rochen Position: Here at Rochen we run between 4 – 6 Seagate Cheetah 15,000 RPM Serial Attached SCSI (SAS) hard disks in RAID10 to provide the highest level of reliability and performance possible in a shared and reseller hosting enviroument. Our RAID10 arrays are also all backed by hot-spare disks.
Rochen Vault
Rochen Vault works by taking a block level snapshot of the entire hosting server (including your files, MySQL databases and email) and storing it off-server on high performance Network Attached Storage (NAS) appliances. It runs at the kernel level and “images” each block on the hard disk to create a snapshot instead of backing up individual files. This provides the following benefits –
a) your account on the server is backed up multiple times per day instead of just once daily as with most traditional backup systems. This type of enterprise backup schedule is becoming more important for customers these days who may be hosting Joomla installs and running other dynamic scripts like WordPress or phpBB where MySQL and other data is changing more frequently than in the past with static HTML sites.
b) because Rochen Vault runs at the block level as a kernel module and not at the higher file system level it can take multiple backups per day without impacting system performance or adding drastically to the overall system load. While protecting customer data at Rochen is extremely important so is maintaining a fast and speedy hosting platform.
c) you can login to the Rochen Vault recovery site at anytime via: https://vault.rochen.com to restore data under your account yourself without even having to open a support ticket with Rochen.
d) unlike traditional backup solutions where only 1 – 3 different backups will be kept you have the ability with Rochen Vault to restore from dozens of recovery points over the past 30 days. You can restore individual files or an entire account. In fact you could even restore one file to a backup taken two weeks ago and another file to a backup taken 12 hours ago.
e) in the event that we need to restore an entire server in a disaster recovery situation we can do so rapidly without having to first re-install the operating system, control panel and reconfigure the services. We can kick off what is called a “bare metal restore” allowing us to re-image the entire server from a Rochen Vault snapshot.
f) all data Rochen Vault stores is not backed up on the local server but to dedicated high performance Network Attached Storage (NAS) appliances. We run a private network connecting all of our hosting servers to the Rochen Vault appliances and all backup data passed over the network is fully encrypted for good measure.
Here are some screen shots of the client side interface of Rochen Vault –
Brad Baker from our team recently demonstrated Rochen Vault at the Melbourne JoomlaDay and received some extremely positive feedback regarding the system.
We are continually looking to improve Rochen Vault and over the coming months we will be adding client-side restore support for MySQL databases as well as files. MySQL data is already all backed up but a ticket most be opened to have us restore databases right now.
Next time you need to restore some data give Rochen Vault a try as we think you will be pretty impressed. If you have already used Rochen Vault then leave a comment under this blog letting us know what you think.
Thanks for reading and until next time!
– Chris