Migrated from eDJGroupInc.com. Author: Babs Deacon. Published: 2013-01-14 04:00:56Format, images and links may no longer function correctly. I recently hit a snag during a migration project when a client’s back up tapes could not be restored due to physical tape failure. At the time I thought, “here we go again, unreliable tape backup”. Then I started to wonder, was this a real problem or just an urban legend like alligators in the sewers? I realized I didn’t truly understand the scope of the problem because all of my information was anecdotal. Anecdotal isn’t helpful when trying to compare backup and disaster recovery options. Cloud storage is the newest entry and stakeholders may be tempted to flock to it as the panacea for all of their storage ills. This is an important decision and it requires hard metrics, not just a vague feeling that one or other of the methods is unreliable or not secure.
I started with Google: “rate of backup error tape” assuming I would find a large number of articles, blog posts, etc. all reporting different statistics, but I didn’t know that I would stumble upon a bit of controversy on the subject.
Several articles and blog posts, in answer to my search, mentioned a 2006 report listing a very high incidence of tape failure, well north of fifty percent. (I’m not repeating the number so as not to add fuel to the internet search fire). Early posters sited the report and figure as gospel (without actually linking to the original report) and later coverage questioned not only the figure but the very existence of the article. Additional articles mentioned tape failure rates just below fifty percent. It sounded like a big problem but I needed a reality check.
I reached out to two groups with long histories of dealing with backup tapes and asked one question, “What percentage of backup tapes that you have worked on have failed?”
Scott Zimmerman, Assistant Vice President of Legal Solutions at RenewData, came back with a surprisingly small number, given the rates I had seen on the web. Renew’s records showed that, “Of the tens of thousands of tapes we have extracted from 2006 – present – approximately 1.4% failed.” Importantly, Scott noted that Renew believes its error rate is especially low because they have such a long history of expertise in this area. “We regularly invest in our technology, have experienced project managers that monitor the results, and run many tapes several times if necessary. In short, we do have an abnormally low error rate, but that’s because of several intentional and significant efforts on our part to keep that error rate so low.”
Index Engines’ number was also a lot lower than my web research. Jim McGann, Vice President of Marketing, responded, “We process a lot of tapes in our lab – it really depends on the age of the tape. For older tapes, i.e. DLTIV, including blanks, IO error, corrupt….. we are probably looking at around 10-15%. For newer tapes the number is closer to 5%.”
Well, there you have it. This information answered my question and I hope it will be useful when assessing the reliability of tape as a storage method.
eDiscoveryJournal Contributor and eDJ Group Director of Strategic Consulting – Babs Deacon