Effective Backup Systems for Today’s Technology

Q: What should be my top considerations in backing up my system?A: Backups are a critical component of proactive systems management and yet are quite often overlooked. Among the most common mistakes people make with backup systems:• Relying on backup systems that were set up years ago without verifying that they actually work. We’ve seen a large number of firms where someone blindly swaps tapes in and out on a daily basis, without once testing the tapes.• Having incomplete backups. When most users think of backups, they generally just back up their documents and files. We’ve seen a large number of servers (usually exchange servers) where just the MS Exchange or “shared files” folders are being backed up. Over time, additional directories or shares get added to the server, and they’re never backed up.• Untested backups. The worst form of backup is no backup. The second worst is an untested backup. Tapes fail; the overwrite function fails. Oftentimes, the backup tapes may never have been formatted or haven’t been overwritten in months because the backup software was set up not to overwrite for 90 days, or whatever the defaults are.I would recommend backing up your system as follows:• Replace tape backups with disk-based backups. Gigabyte-for-gigabyte, when you factor in the tape drive, media and recovery-time costs, hard-drive based backups are significantly cheaper.• Implement full-systems imaging backups. Full-systems imaging clones the backups onto external media–usually USB-attached drives or network shares. Full-systems imaging allows you to restore the complete operating system, configurations, licensed software and all the little things that people customize but never back up–desktop images, Internet favorites, buddy lists, et cetera.• Test the backups! Using a good virtualization product (like VMware, Xen or Microsoft’s Virtual PC), you can restore your backups into a virtual image to verify that the backups actually worked.• Use virtualized backups to test service packs–major upgrades to systems–before rolling them out in production. Oftentimes, when vendors make big changes to their applications or operating systems, it is imperative that these major changes be tested before rolling them out onto the user desktops or servers. A good backup strategy can allow you to restore selected systems, apply patches, test things thoroughly and then deploy to your users or roll back to older platforms.• Maintain remote backups. Once you know you have good backups at the office, start implementing an offsite backup service. The current fashion is to use online backup services such as Mozy or Carbonite. These may be fine for home users, but they are inadequate for most businesses.When selecting remote backups, you have to factor in the amount of data being backed up (usually gigabytes to terabytes), speed of your Internet connection and the time it will take for you to back up remotely. Remember, if it takes you five days to do a full backup, then it will take you one to five days to restore that data. That can be an expensive proposition. A better approach is to store offsite backups in remote offices, preferably within two to four hours of driving time. Or pick local vendors within driving distance.Geography and driving time are important considerations. Why? Look at the scenarios under which you might need to restore from offsite backups: The local backups in the office were accidentally deleted or corrupt. Or systems in the office are inaccessible due to fire, water, smoke or other hazards. In other words, the local systems are inaccessible due to minor disasters and you need rapid recovery. Having offsite backup somewhat nearby allows you to recover within a matter of hours. (This is not the same as having remote backups in other states or countries for disaster recovery or compliance reasons. For some companies—such as banks, governments, multinational corporations, healthcare institutions and publicly listed firms–having multiple backups in multiple states or countries might be a sensible solution. If your disaster recovery plans require you to mitigate terrorist, nuclear, biological or other large-scale threats, certainly having secondary backups across the continent may be the most sensible option.)• Finally, pick backup systems that will actually be used and that will increase productivity. The most common problem we see is, “Oops, I deleted that file on the server.” For those situations, we maintain point-in-time snapshots of the corporate file servers and give our users (or the site managers) ability to go back to a point in time (yesterday, last seven days, last 90 days) and restore the deleted or overwritten files.By making this option available to ordinary users, we’ve significantly reduced help desk calls, and most important, allowed very expensive staff to keep on working. Did you delete a proposal you’ve been working on for three months? Go to the R: drive, go to yesterday’s folder, find your file and copy it to your folder. Did someone overwrite the client presentation two weeks ago and you need it in 15 minutes? Go to the R: drive, go to the folder from 15 days ago, copy the file and confirm it’s what you need. Did someone’s desktop just die (or did they lose the laptop)? Grab a new or spare PC, restore their image and get them back to work within two hours.When evaluating technologies or vendors, ask them those questions and gauge their answers. When your backup system can solve those problems effectively, then you know you have a good backup solution.