February 7, 20197 yr Hello Everyone, After a time of 0 erros the last 2 months, my unRAID array gives my at the parity check 5 errors. I cannot find where it came from. It suddenly appeared. The only thing is that my parity disk broke down, but I replaced it safely and by the manual, so it showing up as an error. Hereby the syslog. If you need more info about my system i.e., I would like to hear that. tower-diagnostics-20190207-1328.zip
February 7, 20197 yr Community Expert Reboot and run another check, if you get the same 5 errors it's the SAS2LP, it's one of the known problems with them.
February 7, 20197 yr Author Okay, thanks for the help. I've rebooted my server, and I'm now running a parity-check. I'll keep you'll updated, parity-checks costs around 9 hrs, so that will be tomorrow morning (CET). Greetz.
February 7, 20197 yr Community Expert After it's done, if it finds 5 errors again, and most likely it will, check if they are the same: Feb 6 20:35:27 Tower kernel: md: recovery thread: P corrected, sector=1565565768 Feb 6 20:35:27 Tower kernel: md: recovery thread: P corrected, sector=1565565776 Feb 6 20:35:27 Tower kernel: md: recovery thread: P corrected, sector=1565565784 Feb 6 20:35:27 Tower kernel: md: recovery thread: P corrected, sector=1565565792 Feb 6 20:35:27 Tower kernel: md: recovery thread: P corrected, sector=1565565800 If yes it's the HBA, replace it with an LSI.
February 10, 20197 yr Author Well @johnnie.black, the parity check gave 0 errors as return... I hope that it will stay so. But what is the HBA, and what is a LSI??
February 10, 20197 yr Community Expert the parity check gave 0 errors as return... That's good, as long as you rebooted first, if it's the SAS2LP problem it only happens after rebooting. HBA is an host bus adapter, or non RAID controller, HBAs from LSI are the ones recommended for Unraid.
February 10, 20197 yr Author I rebooted first, as you said. I've scheduled my parity-check every first day of the month, so what should give the next parity-check number of errors? Because my unRAID server is on-line 24/7... (and the SAS2LP isn't being removed...)
February 10, 20197 yr Community Expert 2 hours ago, rikdegraaff said: so what should give the next parity-check number of errors? Only acceptable number of sync errors is 0.
February 11, 20197 yr Author 17 hours ago, johnnie.black said: Only acceptable number of sync errors is 0. Okay, I will see hopefully that number on the next parity-sync, scheduled at the first of march 2019...
February 11, 20197 yr Hopefully all your disks stay healthy. Remember that sync errors mean that a failed disk will be rebuilt with those errors, so the chance of data loss is much higher. Personally I wouldn't wait to see, I'd run checks until I consistently got 0 errors. Until that happens, you are running at risk.
February 12, 20197 yr Author Okay, I'll start a parity sync right away. And I will do some more after this one. So I hope it stays on 0, if not, I shall post my diagnostics here. Thanks for the advice!
February 13, 20197 yr Author Sorry for the bump/doublepost but is it needed to check the box Write corrections to parity? It is by default turned on, but my logic says uncheck it, if a HDD-data is corrupt, it writes the corrupted data to use as parity numbers. Please let me be wrong...😫
February 13, 20197 yr Community Expert Unless sync errors are expect, like after an unclean shutdown, always do a non correcting check.
February 13, 20197 yr Author 2 hours ago, johnnie.black said: Unless sync errors are expect, like after an unclean shutdown, always do a non correcting check. Thanks for the information, because I am really a beginner with unRAID and other Linux based os'ses. The speed and accuracy of the support forums is since I am been a member, is perfect!
Archived
This topic is now archived and is closed to further replies.