- Urgent
I noticed this sometime ago, but I wasn't sure exactly what was happening. About a month ago I noticed 128 errors after a parity check was completed but saw it only after a week or so had already passed. All SMART reports were good. I ran my monthly parity check several days ago and the exact same number of 128 errors occurred. All SMART reports were again good with zero anomalies on all drives. The syslog showed a spin down an hour or so after parity check was started (FYI, my disk settings are set to spin down after an hour) and that's when the read I/O errors occurred (specifically involving only three drives at that time).
After performing an XFS repair on the three drives with zero issues, I again ran the parity last night. This morning, I noted in the syslog that again an hour after parity check was initiated, UnRAID spun down ALL drives and soon thereafter I/O errors on ALL drives were reported. Two hours later, again UnRAID spun down ALL drives and I/O errors started.
This appears to be a UnRAID 6.7 bug wherein it's incorrectly spinning down drives that may be physically I/O active. I don't believe its hardware related as I've had this particular motherboard/chassis (SuperMicro X11SPH/Chenbro 48-bay RM43348) combination running for over six months now and never experienced anything like this before.
Attached is my syslog (FYI, the parity check is still in progress and although there were numerous I/O errors, the parity check has reported ZERO errors thus far).