|
Building Sun Data Redundancy |
|
|
|
|
Page 1 of 3 Despite the most compelling design, and innovative programming, there are many reasons that a Sunystems can fail. One of the keys to minimizing SunSystems failure is redundancy. The power and data redundancy discussed here are a good start to ensure that your Sun is always available, but other types of redundancy also should be considered.
Backup
Server crashes and hard drive failures are inevitable due to equipment failure, lightning, power outages, simple age or defect-related failures, bugs, hackers, viruses, and, of course, human error. The most basic method of saving your data is to "backup". All backup solutions have one thing in common — they involve copying data from your hard drive(s) to a second media, from which you can restore your data in the event that your hard drive(s) or your server fails.
If you create and maintain your own SunSystems site, keep local and up-to-date backups of all files. At a minimum, a daily backup should be made of all the data on your server, so that if the data is erased or modified in error, much of it can be restored. Backup should be a ritual and scripted so it becomes a part of your daily routine.
There are two kinds of modified backups, each with significant differences:
Differential Method (one full backup + several differential backups): With this method, you back up only the data that has been modified since the last full backup. This is done when you set the backup software so that it will leave a file’s “archive flag” unchanged after the back up is completed. This method gives you redundancy — the original full backup and the most current differential backup. Using the full backup and the latest differential backup, one can safely restore an entire hard drive. This method requires a lot of tape space, however.
Incremental Method (one full backup + several incremental backups): This method backs up the files that have been modified since the last backup, whether full or incremental. Setting the backup software to clear the file’s “archive flag” after it is backed up does this. This requires minimum tape space but may require several tape backup sets to find a lost file. To restore an entire hard drive you must restore the full backup and then restore each incremental backup in the cycle.
Get the biggest and fastest backup system you can afford.
When mapping out a backup plan, consider all the costs including hardware, software, and staff time. It usually takes at least an hour to back up a single server. Manually backing up several servers weekly, semi-weekly, or daily represents a considerable amount of man-hours.
|
System Administration 
