Backup Services

We pay high attention to our business stability, and constantly put much effort into ensuring that our hosting servers are secure and reliable. However, rarely system hardware fails, or some act of Nature may destroy servers and/or the data center we host our servers at. Nevertheless, we hope that such unforeseen event does not occur, but we are ready even for a worst case. And one of the most important things that give us the assurance that we can respond to these unforeseen disasters is a reliable two-level backup scheme.

Each hosting server requires three main datasets to be backed up:

• System configuration files;
• Accounts (users) files;
• Accounts (users) database.

These files are necessary to quickly configure the replacing server with an identical environment, thus ensuring that hosted web applications will keep working without reconfiguration. This information is backed up at least once a day to the backup server, and also to the special server of our core technical team.

Our backup server has entire copy of all hosting accounts. Several times a day it sequentially connects to each hosting server and checks for changes (added files, removed files, files with changed content). Such changes are carried over secure channel and merged with a copy stored at the backup server. This ensures that our backup server has up-to-date content. The synchronization software we use is smart to transfer changes fast when server is almost idle, and go slower when the server is in active use; thus, backing up does not interfere with the usual server activity. If the hosting server fails, all we need to do is simply install a similar hardware, and copy all relevant files from the backup server to the new server; once that's done, the new hosting server is immediately ready to continue providing hosting services instead of the old one.

Since hosting and backup servers are hundreds miles away from each other, chances that both hosting and backup servers will fail at the same time are miniscule, however we are readily prepared for that. Once a week the copy of accounts that is stored at the backup server is written out to a backup media. Our backup schedule is designed in a way that at each moment we keep the following full backups at the external media:

• 4 or 5 weekly backups (during the last month);
• 4 or 5 bi-weekly backups (during the last quarter but last month);
• 3 monthly backups (during the last but one quarter);
• 2 quarterly backups (during the last but two quarters)

Thus, at any moment we have at least 13 backups, plus a very fresh backup at the backup server - total, at least 14 backups. In addition, during the last 2 weeks we keep at least 14 (usually 25 to 35) incremental backups. This provides excellent safety against backup server and backup media failures, and also allows us to resolve problems like accidentally or maliciously deleted or modified files.

This is the most important, and the most dynamic information with most hosting accounts, so we put even more effort into backing up database. We have installed realtime replication of database to the remote server, thus any database changes appear within few seconds on the backup server. This is like using two servers instead of one! Let's once again remind you that the servers are hundreds miles apart - it's virtually impossible that they are affected by a single disaster at the same time.

Once database information has appeared on the backup server, it undergoes the same backup schema as account files - i.e., at least 13 backups at external mediae plus a live copy at backup server hard drives. Two important additions are:

• We keep archive of replication logs, thus are able to restore your database as it was at any point in the past (for example if you suddenly executed SQL statement "DELETE FROM userlist").
• Files backup also backs up database files. In other words, database is double-backed up having at least 28 backups.

Thus ensures that your information will survive almost any disaster our good old Earth can survive.

Nonetheless please note that not all features of SQL are replicable. Databases that use features not compatible with realtime replication are excluded from realtime replication. Such databases still undergo periodic file level backups.

Obviously, backups are the highly sensitive information, so good security is very important. Throughout all backup processes, we use modern cryptographical software. Only a couple of people have access to backup location, backup software and backup mediae. If ever a disaster should occur, they will work 24/7 until all affected domains are online again, with as few loses as possible (our backup schedule allows to avoid losses completely in many cases). Most likely, your hosting account is covered by our downtime compensation policy - in other words, you will be refunded money for the period of unavailability of your domain.

If you need to restore backup even if the hosting server works ok (for example you lost your password and someone messed with your hosting account), you will be charged for the job at rate $100/hour, with 30 minutes minimum billable. In certain cases if at no fault of yours restoring backup is necessary, we might decide to waive the fee once, but for no longer than the first 30 minutes (which should be sufficient in all cases unless you require something unusually tasking, like rolling back huge database, or merging backup with your current account content).

Do not hesitate to contact us if you have any questions regarding our backup procedures.