PROVIDING SOLUTIONS
SINCE 1983.
Mendon Associates Inc. 
 For PublishersFor AssociationsSolutions & IdeasAbout Us
 For Research  For Systems

 

Emergency Data
Recovery Planning

Problem
Definition

What would happen to your business if you lost your computers and the data on them tomorrow? How dependent are you on your systems? To find out, do a simple systems usage survey. Go to each computer in your organization and ask the operator the following questions:

  1. Could they do their job if you took their computer away from them?
  2. If they're hooked up to a central computer or network, could they do their job if you took that connection away from them?
  3. Now, could they do their job if you took away access to all of their existing data files, including word processing files, spread sheets, data bases and all the others that they take for granted in a given work day?

Plot the answers on a matrix similar to the following:

  Can Perform Job with Loss of:
Work Station Computer Network
Connection
Data
Access
Operator #1 NO NO NO
Operator #2 YES YES YES
Operator #3 NO YES YES
Operator #4 NO YES NO

If there are any "NO" answers in this matrix, you have a potential system problem which requires some sort of Emergency Action Plan. Simply defined, an Emergency Data Recovery Plan is a set of pre-determined procedures designed to handle the loss of one or more of your computer system capabilities or resources, and return you to full operation in as short a period as possible, with minimal disruptions. Any plan you might decide upon must take into account a hierarchy of catastrophes. This hierarchy can usually be mapped to five levels of problems ranging from mild to catastrophic, as illustrated here:

Level I Minor inconvenience, such as lost of network connection through faulty wiring, network cards or network hub, but with no real loss of data.
 
Level II Minor loss of data due to accidental deletion or corruption of files.
 
Level III Moderate loss of data due to crash of operator's hard disk or computer.
 
Level IV Major loss of data due to crash of network server hard disk, or complete system failure.
 
Level V Complete loss of most or all computer systems, such as what might happen due to a fire, natural disaster or if someone decides to back a moving van up to the rear door of your office one weekend, and help themselves to its contents.
 

Action Plan

The focus here is on a Data Recovery Plan, because in the end, hardware can be replaced, but lost data usually can't unless precautions have been taken first! The only real insurance against data loss is a backup procedure which is followed in a regimented fashion. You can rely on some system's capabilities to undo file deletions, or to recover damaged hard disks through sophisticated file copying redundancy, such as is found in RAID (Redundant Array of Inexpensive Drives) systems featured on more costly network servers, to a certain extent. Your action plan, however, must allow you to be able to recover from a Level V disaster to be worthwhile.

A regimented backup procedure requires four elements in order to be worthwhile:

  1. A backup device which is reliable.
  2. Backup media which has sufficient capacity to store all required information, both in terms of server and operator data.
  3. Backup software, or methods which allow quick and easy recovery of damaged or lost files.
  4. A set of procedures which ensure both sufficient redundancy in backup to recover data that may be lost for some time before the need to recover is identified, and yet is secure enough to withstand a Level V disaster.

Each of these four elements are discussed with specific suggestions below.

Backup Device

The device you choose to backup onto is crucial to your success in surviving a systems disaster. Picking a backup device which is unreliable is worthless. Backup devices are varied, so you need to select one which fits your requirements. A simple option is a device such as a Zip drive which can utilize 100MB or 250MB disks. This solution can cost you under $150, including media. Another solution is larger removable media, such as portable hard drives or Jazz type drives. Yet another solution is a tape backup system which can hold 12GB or more of data. Mendon Associates Inc. uses a DAT tape backup system due to the speed, portability and low cost of the media.

Whichever device you choose, make sure that you can obtain a backup for it too. If you have a single old cartridge backup system, make sure that if it were damaged or stolen, you could replace it quickly. If the system is so antiquated that you can't obtain a replacement, then your backup tapes could be worthless. There are service bureaus that can read most media for you, however, ideally you want your backup solution to be an exclusively internal one to your company. For this reason, it is recommended that you upgrade your backup device as you upgrade systems, especially servers, in order to ensure that you have the latest devices which can be readily replaced.

Whatever you choose for backup purposes, test it on a regular basis to ensure that it remains reliable. If you are using a tape drive system, clean the heads periodically as per manufacturer's recommendations. Equally important is the location of the backup device. Generally speaking, installing it on the server is the preferred method. Wherever it is set up, make sure that it is configured to be able to back up all systems which contain crucial data. If you cannot see some of your operator's hard disks, make sure that they are backing up individually, or only store data on the server's hard disks.

Backup Media

The media you select should be portable so that it's easy to store. Equally, it should have sufficient storage capacity to hold all of your data, preferably without needing to use multiple volumes. As with devices, media can degrade over time, so periodic testing is a vital procedure which should be part of any backup regime. If your backup software has a verification option, where it checks the data stored on the backup media against the data on the network for comparative purposes, always use this feature. Replace media following manufacturer's recommendations for expected life spans.

Above all else, keep you media away from magnets!! Many paper-clip holders use magnets, so storing tapes near or around these holders can put them at risk too. As well, store the backup media in a secure location away from the server or backup device!

Backup Software/Method

The software you select will determine how sophisticated you can be about backing up. Some software allows you to select specific drives, directories and times for backing up. If you are using a disk type of backup, you could use your system's simple copy commands to copy from the source to the target drive. If you have Windows 95, 98 or higher, there is a backup utility which comes with the operating system that also does file compression, which allows you to get more onto a disk. It also allows you to verify data as mentioned above. Windows NT Servers also come with backup software which will recognize most popular drives and has verification features as well. These packages also have simple restore procedures which allow you to restore all or some files from a backup media.

Backing up is a monotonous and boring task, but it is a crucial step to ensure your success in surviving a systems disaster. For this reason, we strongly recommend that you purchase software that has built-in script writing capabilities and which can be timed to start independently. In this case, all an operator need do is replace the media every morning.

Backup Procedures

Setting up your backup procedures to ensure that you are covered in case of an emergency is important. It is too easy to fall out of good backup habits, so whatever you decide to do, write the procedures down and re-read them on a regular basis to remind yourself what to do, or to make changes to them, if necessary.

It is a mistake to backup onto the same set of disks or tapes every time. Often if a file is removed, it is not noticed for some time, so backing up onto only one set of media is not sufficient. There are tens of thousands of users out there who have purchased a sophisticated tape back up system, but only one tape this is nothing more than a disaster waiting to happen! Some users have three sets of media that they rotate through, but even this is not good enough.

We recommend as a minimum, an eight tape system, where you have one tape designated for each of Monday through Thursday, and four tapes labeled "Friday I" to "Friday IV". In this way, the Friday tapes are rotated so that you have individual backups for each of the last five days, and for each of the last four weeks.

As a further precaution, and in order to meet the requirements of a Level V disaster, Friday IV tapes/media are sent off-site in a rotation of three tapes marked for Friday IV-A to IV-C. The Monday through Thursday tapes are replaced with new tapes annually, while the Friday tapes are replaced every other year. These old tapes are used for one-of backups, such as when large project directories are removed from one of our servers, as a permanent record. Note that the verification option of our backup software is always set to on. Some clients rotate one tape every month permanently out to an off-site storage facility. Given the low cost of tapes, this is an inexpensive form of data insurance.

Backups should be done daily, and should be timed to happen at the lowest point of usage on your network or systems, which is generally at night. All backups should be verified every morning, and backup tapes or disks replaced at that time with the next one in sequence.

Conclusion

What we have presented here are basic procedures to help you in the design of your own plan. The larger you are, the more in-depth you must design your plan. We have glossed over the requirement to have hardware replaced, and haven't even discussed problems with peripherals; all of these will need to be addressed in a formal way if you are to have a complete and effective Emergency System Plan. We also have not touched on the need to have backups of your original software in case they ever need to be re-installed, as well as the need for keeping track of their serial numbers, which are required with some installations.

Whatever you decide on for your Emergency Data Recovery Planning, make sure you execute it consistently. Then to be sure, review and test all elements of your plan regularly. Remember, if you fail to plan, you are planning to fail. If you need guidance in developing your own plan, or want to have one of our professionals review your existing procedures, just drop us a line and we will be glad to assist you.

 

 

 

 

 

4195 Dundas Street West, Suite 340, Toronto ON M8X 1Y4
(416) 239-9661 Fax: 416-239-1076 E-Mail: info@mendon.com

 TOP | HOME | CONTACTS © Copyright 2008 Mendon Associates Inc. All rights reserved.