Ready for expansion (RN Pro Pioneer Edition)-Seeking Advice

Reported Uncorrect 12 ATA Error Count 12

These are the two concerning stats, and it is interesting that they match. It's difficult to get much real information on exactly what is being Reported as Uncorrect btw. If this were a newly purchased disk and if you were still in the return window I'd have suggested exchanging it. But that's not the case. Personally I wouldn't replace it yet, instead I would follow up with another scrub.

winger13 wrote:

StephenB wrote: ... However, in your case the drive has all of a sudden generated a burst of errors - so firmware driver issues are not the cause. It might be ok, but it needs more investigation. Most drive failures also result in other errors (CRC errors on the interface, reallocated or pending sectors, etc). Though I did have one drive that started generating ATA errors with no other bad stats. I ended up replacing it. In this case, was it increasing ATA errors or one big burst like what I saw on mine (0 to 12 during the volume expansion due to adding the two new 3TB disks)

It my case it was an ongoing issue, not a static event. The counts started rising, with a couple of new ATA errors generated every couple of days. Generally I retire disks that start behaving badly, even if I am not seeing data loss yet...

winger13 wrote:

StephenB wrote: FYI a disk scrub on the 4.2.x firmware reads every sector of the data volume, and verifies that the RAID parity blocks are all correct. So it ensures that the data volume has full integrity, and as a side benefit acts as a drive diagnostic of sorts. I run them regularly every 16 weeks, others here run them more frequently. However, if the data volume is already known to be at risk, then you probably shouldn't run one. IF during the scrub, data volume integrity issue is found, what does the NAS software do? Since I am running dual redundancy, does this factor in what the NAS does to remedy a data volume issue?

Netgear doesn't clearly say, so you do need to read between the lines a bit. You can get a better sense here: https://wiki.archlinux.org/index.php/So ... #Scrubbing

Basically, if there is an unreadable block, the normal process kicks in. Parity is used to reconstruct it, and the block is rewritten. If the write fails, the disk will reallocate the sector, keeping all the data intact. Too many bad sectors will cause the NAS (or the drive itself) to declare it has failed. That's one reason why an up-to-date backup is especially important if you have suspect disk(s).

If three or blocks can't be read from the same group (with dual redundancy), then the volume is corrupt, since it can't be corrected. That's a second reason for the up-to-date backup. I'm not sure exactly what the ReadyNAS does in that case. It could simply warn, or it make the volume inaccessible. My hope is that it simply warns.

If all the blocks are readable but the parity is inconsistent with the data, the the parity block is re-written (the feature is "disk scrubbing with auto parity fix") The system has no idea if the parity block is wrong or the data is wrong, so volume corruption potentially results. That's a third reason for the up-to-date backup. Though this scenario (a read succeeds but data is wrong) is generally not caused by hardware failure. Disks include error correcting codes with each block, which are designed to detect bad data (correcting if it can, reporting if it can't). The main cause here is a power glitch, where some cached writes simply didn't happen before the power failed. A UPS is the obvious solution btw, if you don't have one you should. A secondary cause is when the NAS crashes for some reason, again with writes cached in memory that didn't all occur.