Disk Failure Detected...

[mdgm]

kiwi have you contacted NetGear? If so do you have a case number? Best to work with them on an issue such as this. See if you can pay for support.

[bartw]

Let me shatter your hope ...

We have an U6 at a customer, with 4 of these disks (ST2000DL003-9VT166)

System failed with a bad disk 3 hours after we left it at our customer (even thug we do 48 hrs burn ins under load).
Picked the system back up, swapped drives, extended disk test, extended load test, system again OK . Install it at the customer (did the 4.2.20 upgrade there onsite), 8 hours after we leave it there 2 disk disappear ...
After a reboot one of the disks seems to come back online and a resync happens ... except , no volumes available.

Netgear, get your act together, this thread is 2 months old and we still do not have a fix ... you're loosing my confidence ...

[CitizenPlain]

Bartw, note that this thread is actually a YEAR and two months old. I first posted on it on March 27, 2011. The OP was on March 13th, 2011. I've had this problem ever since then. Have tried everything and still no dice. I'm usually safe if I don't do any heavy copying to the drive, but the problem persists. It sucks that it hasn't been addressed yet.

[mdgm]

As I understand it from what a NetGear rep posted, SeaGate recently admitted that there is a firmware issue with the ST2000DL003 drives and have a test firmware available. Don't know when a production update will be available.

[bartw]

A small update from our side ...

Checking configs at customer sites trying to anticipate issues with these disk, I notice that we have at least two more configs with these disk in the field ... and guess what ... no issues reported on those ... also Seagate ST2000DL003-9VT166 ... where is the variable then ?
The disks don't seem to be the decisive factor if I look at this ...

[capaust]

A small update from our side ...

Checking configs at customer sites trying to anticipate issues with these disk, I notice that we have at least two more configs with these disk in the field ... and guess what ... no issues reported on those ... also Seagate ST2000DL003-9VT166 ... where is the variable then ?
The disks don't seem to be the decisive factor if I look at this ...

In my experience, the decisive factor is disk load. The heavier the load, the greater the chance of failure. That being said, we used ours for 6 months before we had a problem. The disks with no issues--what kind of use are they getting?

[bartw]

Hm, you might have a point there.

Most failures happen in the beginning when the first (truck- load of data is sent to them. Ones past that stage, they seem to be honky dory ...

[bartw]

Small update. I've noticed that we have a customer where we haven been running with these disk for over a year, without one single hick up ...
The system is an NV+.

[thestumper]

Just an observation here, but I was one of the people here that was seeing failures every 10-14 days depending on load. About 7 weeks ago, I converted some of my volumes to iSCSI and downloaded an iSCSI initiator for my Mac clients. Since converting the Time Machine targets, I have yet to have a failure. I know that there have been failures reported that do not involve Time Machine; I think I may have posted one myself. Still, my system has been stable for longer than it has been since purchase over a year ago. I realize that iSCSI isn't for everyone, but for any clients that are hard wired with Gigabit Ethernet, it might at least buy you some stability. The downside to testing this is that iSCSI for Mac is pay-for-play. The GlobalSan Initiator I am using set me back $90.00 US, but so far has been worth it. I don't have any hard data, but overall file transfer performance to and from the iSCSI/block volumes seems better overall.

Note that I am still using AFP for shares, and I also stared using CIFS for Windows clients on my network - these are still primarily accessed via the wireless network. I haven't been moving large amounts of on the share for a while now, so again, this evidence is anecdotal at best. But anything that helps users with this problem is worth exploring IMO. As disk prices come down, I may start moving to new drives if/when I do have another failure.

In theory, it would probably be best to abandon RAID-5 moving forward. Having all the failures I've experienced, I've realized that the rebuild times are so long on the 2TB drives that I really am setting myself up for catastrophic failure (2 drives). If I keep my current ReadyNAS, I might do RAID-10; if I step up to another NAS, I may suck it up and go for more drives to do RAID-6.

Hope this helps someone!

[jrpNAS]

Seagate's 2TB Green disk have Flush Cache problem. If you turn off the disk cache, then the NAS are more reliable.
My Ultra 4 has worked of February 22 to May 26, without any problems. I enable to the write cache May 21 and May 26 one disk was "die".
After volume rebuild I turn write cache off and NAS is still running without problems.

In February I reghance one disk because Seatools find several bad MBA's. That disk died many times even if write cache was off.