Bad Blocks

[yowmemperor]

I have now had two new tapes fail a job due to bad blocks. Will a long erase register these as would formatting a hard drive so I can complete a job?

[veremin]

Not sure whether I follow you on that, could you elaborate on your question?

During tape job execution, you run into two corrupted tapes and now are asking whether this corruption would show up, if you run tape long erase, right?

Thanks.

[yowmemperor]

Correct. It gets x far in the tape write and hits bad blocks and fails. Will a long erase adjust for the bad blocks so I can then write to the tape? I pulled out the first bad tape, and just ran into another bad one.

It would be really nice if Veeam would either skip the blocks, or stop and move on to the next tape. The backup map takes 40-60 hours, then days to write to tape. The restarts are getting painful

[csydas]

If they're new tapes corrupting, you might want to check system/app logs on your tape server and do some extended write testing since it sounds like a bum tape drive or HBA causing problems. Like, maybe it's a bad batch of tapes, but twice coincidence, thrice means something's up.

A long erase isn't going to fix this; if there's bad-ness on the tapes, it's unreliable for data storage and you can't really trust it. Veeam won't know about the block until the drive does, and the drive won't until it reads. Unfortunately it's just not like a hard drive, you can't really read around bad parts of a tape.

[yowmemperor]

csydas, this is the answer I was looking for. Thank you.

It would be great though if Veeam could just load the next tape. It all worked, or appeared to work, under the old backup system we replaced. I have a new set of tapes on order. Good though to check on the drive and HBA. FWIW, the drive was just cleaned.

[csydas]

You're welcome! Hopefully it's as simple as a loose cable (had that once...frustrating week)

I think the logic behind making it fail is that usually it's indicative of a larger problem if there are bad tapes. Veeam's parlance for warnings vs errors is that a warning means there was an issue, but the backup could proceed and data integrity is not threatened. Errors are indicative of a larger issue that requires investigation. Bad tapes usually precede a bad drive, I think it makes sense to error out the jobs instead. I mean think of a fairly simple scenario where you've got a bad drive that just starts eating tapes. I guess there could be some business logic to stop after X failures, but that's still X tapes you burn through which could have been caught by a quick diagnostic from the tape testing tools.

But, that's just my take on it, I can see your position.