Hello I have a Linux Server mounted with BTRFS in RAID-1 and I would like that someone more expert than me could solve my doubts. I have 3 HDD of 6TB each, in one of them every so often smartd detects errors in some sector. When it happens I use hdparm to read it and, if it indicates some type of error, I use hdparm to overwrite that sector, re-read it and check that it no longer gives read error. The SMART Current Pending Sector value is reduced again to 0 and finally I scrub the BTRFS system to recover the lost information. My question, do I have to worry about this HDD? (It is always the same in others does not happen to me). The disk is a WD Red NAS (the others are the same) and is still under warranty, could I process it with these symptoms or I need more symptoms to do it?
If a HDD shows pending sectors repeatedly you should replace it.
A pending sector is a sector that currently will give a read error. When it is rewritten it is (usually) mapped to a reserve sector by the drive firmware. If this happens repeatedly the number of bad and remapped sectors is growing and the drive is unreliable. I'd check with WD if this is covered by their warranty and replace the drive in any case.