APLawrence.com -  Resources for Unix and Linux Systems, Bloggers and the self-employed

2005/08/06 Backup verification

When backing up to tape (or really any media), you need to verify. Anything can go wrong: tape media is much more subject to physical damage than a hard drive, so data can be written incorrectly. Simple dirt or dust can affect a tape backup too. Verification is vital.

There are two basic ways to verify. One is to take a checksum of each file before writing it to tape, write that checksum, then write the tape data. The file can then be verified by reading back the checksum and comparing to a new checksum computed from the file data read from tape. Note that the drive itself may compute independent checksums for each block of data written; if that's the case storing a file checksum may seem redundant. Most any backup software (tar, cpio, etc.) is going to store that checksum anyway.

The other method is to read back the data from the hard drive and compare to the data on tape. That's where an apparently redundant checksum could be of value: you could save time by comparing the stored checksum to one newly computed, rather than re-summing on the actual tape data.

The "checksum on tape" method has the advantage of being able to verify backup integrity at any time without any need for comparing to on-disk data. But it is also valuable to compare back to the original on disk because it may have changed do to user activity or (more rarely, of course) hardware problems.

There are three basic conditions:

Data changed sometime during the backup of a particular file. Either method (stored checksum or re-read from disk) should identify that.

Data is corrupt on tape. Either method would know that.

Data changed after this file was backed up but before getting back to verify the file. The checksum method doesn't care that the on-disk file has changed. The re-read method does.

In this last case, the value of that is often forensic: if the file wasn't supposed to change during the backup period, why did it? In one particular situation, that anomaly caused us to look at who was logged in during the backup window and ultimately led to a discovery of embezzlement.

Of course there are always files expected to change during backup. The software I use lets you specify files that don't need verification (log files, etc.). It lists the files that do change, including just time stamp changes - which don't threaten the integrity of the backup, but again may indicate unexpected use. That time stamp difference once triggered an investigation that turned up an employee stealing customer contact information after hours. His goal apparently was to take that data with him to another employer or to start his own business. He had no access to the actual data files, but could print reports, which is just what he was doing. It would have been impossible for him to print the entire database in one night, so he was doing it in sections, all the "A"'s tonight, the "B"'s tomorrow, and so on. The time stamp difference caused that activity to be noticed; the next night I set a trap to see exactly what he was printing, and he was caught before he got to the "D"'s.


Got something to add? Send me email.



1 comment



Increase ad revenue 50-250% with Ezoic


More Articles by

Find me on Google+

© Tony Lawrence







Mon Aug 8 14:28:33 2005: 940   BigDumbDinosaur


In this last case, the value of that is often forensic: if the file wasn't supposed to change during the backup period, why did it? In one particular situation, that anomaly caused us to look at who was logged in during the backup window and ultimately led to a discovery of embezzlement.

So that's how all those Enron crooks...er...executives got caught! <Grin>


------------------------
Kerio Samepage


Have you tried Searching this site?

Unix/Linux/Mac OS X support by phone, email or on-site: Support Rates

This is a Unix/Linux resource website. It contains technical articles about Unix, Linux and general computing related subjects, opinion, news, help files, how-to's, tutorials and more.

Contact us





What do such machines really do? They increase the number of things we can do without thinking. Things we do without thinking — there's the real danger. (Frank Herbert)

One day my daughter came in, looked over my shoulder at some Perl 4 code, and said, "What is that, swearing?" (Larry Wall)












This post tagged: