Smart Log Errors

[Ryan Gillette]

http://pastebin.com/qrWu10dQ

These are errors I am getting any where from 1 to 4 days of FreeNAS being rebooted. Nothing looks wrong in the outputs, is this an issue with FreeNAS or am I not seeing something and need to get the drives swapped out from WD.

I have setup short and long self tests to run automatically since these outputs. I have also ran short and long tests manually on all my drives. Nothing appears to be wrong with the drives that I can tell. Also I added 3 fans to the case to get the temps under control.

WD said that I can send in the drives for replacement, but I would rather not replace the drives if there is nothing wrong, or get charged for new drives if WD finds nothing wrong with them.

Here is a more recent output of smartctl -a on all the drives.
http://pastebin.com/61iqZzNc

FreeNAS 9.3 64bit (latest update as of July 24th)
Hardware.
ASRock Rack C2550D4I
16GB ECC RAM
Onboard SATA ports, no raid setup on motherboard.
3x WD Red 1TB drives in raidz1
3x WD Red 3TB drives in raidz1
Volume spanned across both raidz1's

 

[cyberjock]

Can we get all that stuff the forum rules require when you create a post?

Thanks.

 

[Ericloewe]

For future reference, please post a new message when adding relevant information, since updates don't show up as new messages. Also, second link is broken.

Some drives are not running the recommended regular SMART tests. In any case, they all look fine. Are you using the Marvell SATA ports? If so, switch to the Intel ports.

 

[Ryan Gillette]

Here is the updated link. Not sure what happened to the other one. I will have to look but I think I have more drives then the Intel ports. I did just see that there is a firmware update for the Marvell chip.

http://pastebin.com/wLDcEtSQ

 

[BigDave]

Your disk marked ada2 reading high temps!
194 Temperature_Celsius 0x0022 094 090 000 Old_age Always - 56

 

[Ryan Gillette]

Yea, I was running with no fans for a little while. I have 3 low noise 120mm fans in the case now.

 

[Ericloewe]

Yea, I was running with no fans for a little while. I have 3 low noise 120mm fans in the case now.

Well, 56 degrees Celsius is about 16 degrees too high. And the drive has seen 60 (!) in the past!

Those high temperatures can also wreak havoc.

 

[Ryan Gillette]

Well, 56 degrees Celsius is about 16 degrees too high. And the drive has seen 60 (!) in the past!

Those high temperatures can also wreak havoc.

Looks they they are running at 28 - 36 degrees Celsius now that I have the fans.
Should I replace the drives that are coming up with the smart log error in FreeNAS?

 

[cyberjock]

Looks they they are running at 28 - 36 degrees Celsius now that I have the fans.
Should I replace the drives that are coming up with the smart log error in FreeNAS?

They won't qualify for an RMA unless you actually have a problem, and you don't have a problem that is proven to be the drives.

You'll also find that if the manufacturer finds the drives were too hot, then the warranty is void. So keep that in mind too.

 

[Ryan Gillette]

They won't qualify for an RMA unless you actually have a problem, and you don't have a problem that is proven to be the drives.

You'll also find that if the manufacturer finds the drives were too hot, then the warranty is void. So keep that in mind too.

I sent that pastebin to WD. They said they are in warrenty and that I could RMA them.

 

[Ryan Gillette]

I just reconfigured my drive locations and they are now running between 29 to 33 degree Celsius. WD Red operating temp in their PDF says 0 to 70 degree celsius. So they never made it out of operating temp.

 

[Ericloewe]

I just reconfigured my drive locations and they are now running between 29 to 33 degree Celsius. WD Red operating temp in their PDF says 0 to 70 degree celsius. So they never made it out of operating temp.

Consider the temperature listed in the datasheet the "it's a miracle the drive is still mostly readable" threshold.
70 degrees will quickly toast any drive.

 

[Ryan Gillette]

Consider the temperature listed in the datasheet the "it's a miracle the drive is still mostly readable" threshold.
70 degrees will quickly toast any drive.

Yea, I was just meaning, that if I was to warranty the drives, they could not say I was out side of their "operating temperature". I still can not figure out whats causing the errors in FreeNAS GUI. Saying SMART log read error, and if its something that I should worry about. I did flash an update on the Marvell chip that is suppose to fix some harddrive stability.

 

[Ryan Gillette]

I am going on 10 days uptime now with the new firmware, and drives running around 30 - 32 degree C. No errors have shown in the Alerts window. So not sure if it was the older Marvell firmware or the addition of couple of fans that cured the issue.

 

[DrKK]

Almost certainly the fans.

People think that just because the spec sheet says the drive will operate at 60 or 70C or whatever, that this is OK.

It's not. This is not some kind of conservative NAS-nerd thing. Our experience, and the research, on this is very, very clear. These drives react poorly, and often permanently damaged, when they reach a temperature like 60C...or even before.

Even 40C would get my attention **IMMEDIATELY**.

Most of us will start flipping out if our drives go above 35-38C, and we'd consider "code red" to start somewhere in the high 40's. If I ever had a drive that even momentarily got to 55C or more, it would be coming right out of service. I would never trust it again.

The drives throwing errors and doing all sorts of crazy stuff is completely explained, potentially, by the thermal condition you unfortunately subjected them to. Whether or not it has caused permanent damage? Who can say. But I can say this: There is a 100% chance that you substantially reduced the lifetime of that drive, and if it has reached 60C, it has gotten way past the point where it would be considered unworthy of DrKK's NAS