July 1, 20215 yr My server runs a parity check every last day of the month, which in this case ran yesterday. About an hour before it finished I received an email from my server about a disk (disk sdi, my parity disk) with read errors. The unraid GUI is showing the disk has 383 read errors. The parity check did however complete without any errors. The SMART info of this disk did not show any errors. To make sure the disk was fine I ran both types of SMART self-tests. SMART short self-test finished in under 2 minutes, without any errors. SMART extended self-test finished in about about 17-18 hours,(similar to the time it takes to do a parity check) ,without any errors. Can I still trust this harddrive? Could these read errors have been caused by something else other than the drive, for example the cable or the SATA controller? Some things that might be relevant: This drive is 2 years and 5 months old. It has been in the server 24/7 since and only powered off a few times (hdd sleep is disabled). This drive is connected to 1 of my 6 SATA ports on my motherboard. It has always been connected to the motherboard directly, but I've recently (2 weeks ago) replaced it's no-brand SATA cable with an 18" CablesMatters one. I've had 2 system crashes 3 weeks ago because of 1 bad ram stick. I've replaced all RAM sticks since which all passed 7 passes of mem86+. All 383 errors happened within a few minutes/seconds from each other and have been logged as "disk0 read error" in the syslog (attached). Diagnostics are attached. shadow-diagnostics-20210701-1750.zip Edited July 1, 20215 yr by ssh
July 1, 20215 yr Community Expert It's logged as a disk problem, but it's weird that it was sector 0: UNC at LBA = 0x00000000 = 0 Run an extended SMART test (or another parity check, but in that case replace/swap cables first to rule that out).
July 1, 20215 yr Author Just now, JorgeB said: It's logged as a disk problem, but its weird that it was sector 0: UNC at LBA = 0x00000000 = 0 Run an extended SMART test (or another parity check, but in that case replace/swap cables first to rule that out). Thanks for your reply JorgeB. I already ran a short and an extended SMART test which did not find any errors.
July 1, 20215 yr Community Expert Should be fine them, I would recommend replacing/swapping both cables/slot just to rule that out then keep monitoring.
Archived
This topic is now archived and is closed to further replies.