December 27, 20178 yr Hello all, After having no luck with any preclear script or plugin (I tried all of them, see here for more info), I just let unRAID clear my new 8TB Archive drive. Afterward, I began the rsync to move all the data off my disk 5 onto the new disk 10. That went fine, but when I ran the checksum part to verify, the disk (disk 10) started throwing a ton of read errors. The checksum returned two files I needed to move, when I did that, it threw more errors. Happy coincidence, a regularly scheduled parity check went off at the same time, and it corrected 1185 parity errors, though it still shows 320 read errors for disk 10. There was a great deal of SMART thrashing going on during this process as well, resulting in 80 reallocated sectors and 10 reported uncorrect, although I did then run a short and extended SMART self-test which both came back fine. Bottom line, is this drive safe? Did I just write bad parity? Thoughts? Edited January 7, 20188 yr by pengrus Solved, removing diagnostics
December 27, 20178 yr Community Expert 24 minutes ago, pengrus said: Bottom line, is this drive safe? I would replace it, should still be under warranty, no? 25 minutes ago, pengrus said: Did I just write bad parity? Probably, if the data is still on the original disk just replace the bad disk and run another checksum compare.
December 27, 20178 yr I'd say you just did write bad parity. As soon as you have a drive that shows read errors - especially if combined with reallocated sectors, uncorrected etc - then your parity data becomes critically important, because it's only untouched parity that can allow the system to recompute the contents of unreadable sectors. Any incorrect data read while synchronizing parity means you get one sector of bad parity. And that means if you replace the bad disk, the updated parity will recreate the same incorrect read.
December 27, 20178 yr Author 4 minutes ago, johnnie.black said: I would replace it, should still be under warranty, no? That's a whole other bag of annoying, it's been on my shelf for almost exactly a year, and when I put it into Seagate's warranty checker, it says it's OEM and I have to go back to the seller. Which is some third-party fulfilled by Amazon. Headache ensues... 6 minutes ago, johnnie.black said: Probably, if the data is still on the original disk just replace the bad disk and run another checksum compare. Could I be reasonably certain that the errors came from the files that failed the checksum the first time? Or is my data on disk 10 undeniably bad and should be discarded. Also looking forward to how to replace that disk if I have to, do I let the probably bad parity rebuild onto a new disk and then checksum with disk 5? Thanks much!
December 28, 20178 yr Community Expert 7 hours ago, pengrus said: do I let the probably bad parity rebuild onto a new disk and then checksum with disk 5? Yes, as long as you have the original data you can just compare after the rebuild and fix any corrupt files, parity will be updated accordingly.
January 7, 20188 yr Author Yeah, there's something funky about this particular system. And not in a good way. But, I let the drive rebuild, checksummed, and fixed the applicable files, so we're good to go. Thanks to @pwm and @johnnie.black!
Archived
This topic is now archived and is closed to further replies.