Eating boot drives

[Whattteva]

OCZ, like Seagate, uses a 48-bit value for their RAW_READ_ERROR_RATE.

The most-significant 16 bits are the error count, the least-significant 32 bits are the number of reads. So in this case, it's actually 0x00000D5A7F51, and dash-separated for visibility it's 0000-0D5A7F51 - or "zero errors in 224,034,641 reads" - and the latter 32-bit section loops around.

Once that first section jumps to 0001-0D5A7F51, then you have an error. In decimal value, once you break 2^32 (4,294,967,296) then be concerned.

That's interesting. Are you just supposed to know this or is there some documentation about how the breakdown of the bits?

PS: Interesting how they use a really odd int size. I don't think 48-bits is even a standard word size in any C compiler.

 

[WI_Hedgehog]

That's interesting. Are you just supposed to know this or is there some documentation about how the breakdown of the bits?

PS: Interesting how they use a really odd int size. I don't think 48-bits is even a standard word size in any C compiler.

Based on their rather excellent S.M.A.R.T. reporting as compared to other venders in general, I'm going to guess Seagate put a lot of thought into it and OCZ saw it and thought, "Yup, good enough, we're doing that too!" Out of their very interesting history I'll quote the following:

OCZ-branded SSDs were notable for high failure rates. Two lines from the old OCZ Technology Group, the Petrol and the SATA II versions of the Octane, had return rates of over 40% at one anonymous French technology online retailer. The SATA II version of the 128 GB Octane had a return rate of 52.07%. No other company had any SSDs with a return rate of over 5% from this retailer in the data set that was published on March 5, 2013.​

To that point, Toshiba bought OCZ so the quality should be significantly improved.

Regardless, and back on topic, even cheap USB drives should be lasting longer, and I'd tare that system down to the bones and build it back up piece-by-piece, testing it every step along the way. (You may think I'm kidding, but I just did that a third fourth time time with my TrueNAS HP server. It's pretty effing solid right now, hopefully it'll be Production in two weeks--I mean, if that doesn't cut into my sleep schedule too much.)