Question SATA cable causing frequent drive disconnect & data corruption

Hrlyrdr22

Reputable
Apr 18, 2015
188
0
4,710
5
tl;dr bad SATA cable can cause drives to disconnect and corrupt your data

Over the past month I have had a 5TB hard drive completely corrupt all the data on it and exhaust all the reserved sectors with hundreds of uncorrectable errors from the SMART data. It was a shucked Seagate 5TB SMR drive to be specific with 2 months of total power on time and around 9TB of data written, all SMART data was in perfect condition and sat at 32/33c the entire time.

Then I had a brand new ~35-40 day old Seagate 4TB Barracuda drive (replacing the spot the old 5TB took, so same SATA cable) corrupt on me with 69 uncorrectable errors, no bad sectors, and same corrupted data. I was running macOS High Sierra with all drives formatted as macOS Extended Journaled, the default.

At this point I switched over to FreeNAS, copied all my data to an external HDD and created a ZFS pool with two identical Seagate 4TB Barracuda drives in mirror (RAID 1). Everything was working excellent for 2 weeks before 3 days ago I started getting bombarded with constant notifications saying one of my 4TB drives was constantly being disconnected and reconnected. I switched the drives to two different ports I have never used on my motherboard and that same drive was giving me issues. I'm confident that I have narrowed it down to the SATA cable as I have never had 2 drives experience any of these symptoms so it couldn't be the motherboard, and two relatively new drives getting corrupted and disconnected was extremely unlikely to be a drive issue. I've now replaced the SATA cable on that drive and have not received any notifications about it being disconnected. Physically, the cable appears to be in great condition. However, it is twisted near the head at a 45 degree angle to route the cable from the back of the case to the front where the motherboard is. All my cables are at most from mid-2015.

Another thing I noted was when the drive reconnected, ZFS started a resilver process because the 4TB that was having issues was missing about 400GB that the other drive in the mirror had. From my observations, it seems a bad SATA cable can cause drives to unexpectedly disconnect (which can cause data corruption if files are being written, although ZFS employs copy-on-write so this is a non issue for me) and even fail to write data at all. Missing 400GB in a mirror is far too high of a delta (from 1.8TB to 2.2TB) to just be bad luck.

Any insight and opinions would be appreciated, and to anyone experiencing similar issues, the cheapest part to replace would be the SATA cable so it's worth a try. If you want to copy all the data to another drive so you don't lose anything, I would recommend mounting the drive in read-only so it doesn't corrupt any more data. Although, I'm not sure if a bad SATA cable can also corrupt the data it reads since I copied all my data using a SATA to USB enclosure, bypassing any SATA cable or controller issues.
 

ASK THE COMMUNITY

TRENDING THREADS