The methodology of backing up a server is one of the most important aspects of a well-designed system. Far too often W-Systems works with clients with lacking backup strategies. We strongly believe in having a backup strategy, and even a backup for your backup.Our backup ideology is designed to protect beyond the failure of the physical server itself. The cost of losing a CRM system is extremely detrimental to the overall health of our client’s organizations, and therefore we are extremely serious about it.
W-Systems' first line of defense is our use of RAID 10 for the file system. Hard drive failure is one of the more common causes of data loss, and RAID 10 protects us so that a failing hard drive won’t cause data loss.
The next line of defense we have is to take snapshots of the Sugar instance. These snapshots include both the database and the file system. They are stored locally on the same machine, so this doesn’t protect us from a total system failure. These are designed to make it easy for us to recover from human error.
The snapshots give us the ability to see what the database or file system looked like yesterday or even two weeks ago. If a user does something to the system that isn’t immediately obvious, we can pick a point in time to compare it to and resolve the issue without having to do a full system restore.
Every night we capture a complete image of the server. This includes the operating system and all installed applications. This image is upload-able, so we have the ability to boot that image onto another piece of hardware if needed. Also, we can set up a temporary development server to try something new before we do it on the production server.
W-Systems keeps one daily image and one weekly image at all times. If the physical server died, this image would be the easiest way for us to get a client back up and running in a very short amount of time.
Our policy is to store this image in three locations.
All three locations are in the same data center, but they are on different physical machines. Two of those machines are in separate zones of that data center. The separate zones are running with distinct power, air conditioning, and Internet connectivity. If one zone of the data center failed, the image would still be available in another zone.
Our final backup strategy is one that I think is the most important: off-site backups. Off-site backups are imperative. Even though we’ve placed our servers at one of the best data centers in the country, there is always a chance that it could go offline.
To avoid worst-case scenarios for all of our clients, we offer off-site backups to self-hosting clients. Our clients have access to the off-site backup server and can download a copy of their most recent backup whenever they want.
We run our own off-site backup server in a different city from our main servers and the off-site backup server is even run by a different company.
Even if the company running that data center failed, we have a copy of every Sugar instance at a data center run by a different company in a different city. It will be a rare event that we’ll need to use this final backup strategy in a recovery, but if that ever happens, our clients’ data will be safe because we took the time to have this additional backup step.
All of our above backup options protect loss of data from human error and server failure, but the off-site backups give true peace of mind. Even failure of the entire data center isn’t enough to cause us to lose the SugarCRM instance.
Regardless of the confidence, you may have that your existing backup plan is sufficient, I strongly encourage you to consider W-Systems' off-site backup offering for that extra protection.
You can’t afford to lose your CRM.
W-Systems has updated it’s hosting offerings since the original publication date of this blog post. To learn more about our cloud hosting, click here.