SATA reset

Have you looked at the SMART stats under health?

Seatools doesn't care about the format, so it can check the drives.

Smart stats are fine on each drive - I was monitoring them throughout.

Chris

SeaTools for Windows will not run properly to check a drive if it is connected via USB. It must be connected via an SATA port of the motherboard.

PapaBear thanks for the tip. Looks like I will have to put them into my production pc then.

Chris

I also notice that the resync seems to run in two passes - the first one at around 60MB/sec, and then it starts again, first at <1MB/sec speeding up slowly to around 12MB/sec. Is that normal?

Chris

Not in my opinion. I really haven't watched very closely that often. Currently my second NVX is doing it's weekly disk scrub with is basically a resync. Now it usually takes about 10 hours (starts a 4AM and finishes about 2PM). That calculates out to about 120MB/sec, but then it is an x-86 based unit with X-Raid2. Sunday night I lost a 1TB drive on my first NVX (mixed 2x3TB and 2x1TB) when I installed a spare and it went through the resync, it only took 6 hours, but that could be because it only did the 1TB layer.

If you don't have a spare SATA cable, you can temporarily use the one that normally goes to the DVD drive.

Lat resync took 9 hours and this looks similar. It looks more likely one of the disks is faulty.

Chris

I have a few conclusions from this little adventure:

1. When I tested my disks, it turned out that both had failed, which showed up right at the start of SeaTools’ test sequence in each case. So a SATA reset (whatever it is) does seem to be a possible indicator of disk failure. To their credit, Seagate replaced both drives immediately.

2. Both disks failing simultaneously is not supposed to happen! It defeats the protection of a RAID array. So this reiterates the straplines of several contributors who are a lot more knowledgeable than me – a NAS is not a substitute for a backup. Luckily I had one.

3. But why did the two disks fail together? It could have been coincidence, a bad batch of drives, some exogenous event, or some event that came from the NAS itself. I don’t know, and that is slightly worrying.

4. Throughout, the SMART statistics and the ReadyNAS frontview panel (which probably feeds off the SMART statistics) showed both disks as healthy, so you cannot rely on those indicators.

5. The current version of SeaTools can test through a USB enclosure, which makes life easier.

6. And now the second shocker – although I had a backup of the full ReadyNAS to a USB drive made with the RAIDiator 4.1.8 software, there appears to be no function of that software to restore a backup. What is the point of a backup if it can’t be restored? Netgear please take note! I ended up doing a manual restore using a client PC to pull the data off the USB, through the ReadyNAS, and then write it back onto the ReadyNAS. Even with a 1GB network card, 500GB data was not fun.

I hope these notes help others in a similar position.
Cheers
Chris

Multiple drive failures in a RAID array are not that uncommon.

Generally the drives are the same model, purchased at the same time, placed in an identical physical environment, and run under identical loads. Considering all that, it is not very surprising that they often fail in batches. Quite similar to having your automotive brakes on both front wheels needing to be replaced at the same time. (See http://www.pdl.cmu.edu/PDL-FTP/Failure/ ... 06-111.pdf ... "We also present strong evidence for the existence of correlations between failures")

Also the SMART stats generally are good until the disk actually starts to fail. There often is not much lead time. So it is good to have a spare or two handy. And as you discovered, having a backup is quite important.

Stephen

Interesting, and thanks for the pointer to the Carnegie research. I am still a bit wary as to whether I am missing something, since my drives died after around 10,500 hour of use.

Chris