Corrupted Data After Resilver Discussion

[TwinDaddyKev]

Currently running FreeNAS 8.0.4-RELEASE-p1 6x2TB Raidz2

A few weeks ago I got a SMART error stating 2 pending sector error and 1 bad sector detected on 1 of my drive.
I waited a few days before replacing the bad drive. During that time, the volume was online even after a reboot.
When the replacement drive arrived, I started the replacement process and the volume began to resilver.
After resilver, it showed that I have data corruption. One of the file was corrupted. Luckily it was a file that I didn't need.
So I deleted it from the volume and perform another scrub. There was no errors detected after that scrub.

So here's my question:

Should I have taken the bad drive offline when I get the SMART error stating pending sector error and bad sector detected?
Would leaving a bad drive running cause this type of issue?

Thanks for any input.

TdK

 

[William Grzybowski]

No, there is redundancy and checksum.
You're only going to get corrupted data if 2 or more disks have silently corrupted data (pending sector for instance or offline uncorrectable sector)

Memory could be another cause, though it is rare, thats why ECC is recommended.

 

[TwinDaddyKev]

Thinking back...I did have 2 drives that was displaying pending sector and offline uncorrectable sector.
I only have 1 drive on hand so I replaced that 1 first. Resilver and everything was back online with no errors.

I got the error when I replaced the 2nd drive a few days later.

Would I still get this type of error if I have taken both drives offline? I am running Raidz2 so I would be OK if I lose 2 drives.