backup-verify vs tools verify media.

[leydon]

Retrospect SingleServer 7.0.x

 

I back up to LTO3 tapes in an autoloader. I've always got verification turned on in the backup script options. Does this essentially do the same thing as using Tools Verify Media? I ask because for all intents and purposes unless I've got a separate Retrospect system and lto3 tape drive there is really no way to use Tools Verify Media with an LTO3 tape. It looks like it will verify around 13G/hour and the tapes have anywhere from 600 to 700G on them (hardware compression). I can't see that there is any way to stop a verification then restart it where I left off. In the best of all possible worlds I would verify each tape as it fills up but can I be "relatively" safe in assuming that if the backup script doesn't report any problems with the tape during the verification phase then the tapes "should" be readable when I need them?

 

Thanks for any insight. How do people handle this sort of thing?

[rhwalker]

Quote:

---

can I be "relatively" safe in assuming that if the backup script doesn't report any problems with the tape during the verification phase then the tapes "should" be readable when I need them?

 

---

You are overlooking issues of shelf life of tapes. Tapes, like all media, degrade over time. If the tapes aren't high quality, or if they aren't stored under optimum conditions, the oxide will flake off. There's also an issue if you try to read the tapes on a drive on which they weren't made. Alignment differences can cause issues.

 

Quote:

---

How do people handle this sort of thing?

 

---

Multiple copies of backup sets, and good storage of tapes.

[Lennart_T]

Quote:

---

I've always got verification turned on in the backup script options. Does this essentially do the same thing as using Tools Verify Media?

 

---

Sort of.

The former compares the contents of the media (tapes in your case) with the backed up volume. This implies rereading all data of the tapes.

The latter just makes sure the tapes are readable, without comparing against the original data.