Miraculously healing drives

[Segfault]

This is third drive what does it to me. First zpool drops it, bad unreadable sectors. Indeed, smartctl test also fails. Then I scrub the pool, clear it and the drive has no read errors and no relocated sectors. Passes smartctl test. All drives are WD Red 3 TB CMR drives. I have bought a new drive every time, the data I care more about than some 90 bucks. Now I have three drives, each about 25,000 hours on it, smartctl tells they are good. Are they?

 

[toast0]

Sometimes unreadable sectors are fine if they're rewritten. Maybe the sector is fubar, but maybe it just got a iffy write for whatever reason. While it's unreadable, it'll be shown in smartctl, but if the write is successful, all is forgiven (by the drive logging anyway). If the write fails, then it would be reallocated.

I don't know that I would forgive it but my threshold for bad sectors is higher than a lot of people. 10ish for drives where you don't check frequently, and 100ish for drives where you have active monitoring and a reasonable backup plan.

 

[Dan_D]

You can get false positives on bad sectors that go away when they are overwritten or when the drives are formatted. I've seen it many times over decades of dealing with hard drives.

 

[Segfault]

Thanks for responses. False positives, eh? Now I have three spare drives, I guess. My RAID is RAIDZ-2, so I can afford putting them back one at the time and see how they serve. I have off-site weekly backups, too. Considering inflation (PPI just jumped up again, will hit CPI soon) collecting hard drives may prove a good hedge. LOL

 

[Segfault]

Maybe this is not so good after all, for instance for one drive smartctl shows read error rate 22, threshold is 51. Below is from Acronis knowledge base.

Description
Read Error Rate S.M.A.R.T. parameter indicates the rate of hardware read errors that occurred when reading data from a disk surface. Any value differing from zero means there is a problem with the disk surface, read/write heads (including crack on a head, broken head, head contamination, head resonance, bad connection to electronics module, handling damage). The higher parameter’s value is, the more the hard disk failure is possible.

Recommendations
This is a critical parameter. Degradation of this parameter may indicate imminent drive failure. Urgent data backup and hardware replacement is recommended.