I have two new identical drives and am running preclears on both at the same time last night.   IE the pre-clears started less than 1 minute apart for both drives.

Very near to 12 hours in, I start getting the same smart error on both drives about the same time.  The time stamp is the same minute, it doesn't show seconds

• Event: Unraid device dev3 SMART health [199]
• Subject: Warning [TODD-SVR] - udma crc error count is 1
• Description: TOSHIBA_HDWG160_81J0A02MFDQG (dev3)
• Importance: warning

Error counts 1 - 4 happened at the same time for both drives over a 3 or so hour period

Error 5 came to one drive by itself.

Then error 6 for one drive and 5 for the other happened again at the same time.

That was the last error sent and it was almost 2 hours ago.

 

Both drives are connected to eSata ports that are extensions from internal MB SATA II ports (I actually just realized those were II's and not III's).

 

Last night I had some difficulty getting the drives connected to start the pre-clears. 

• I have one other external drive that is used for a backup, IE Not part of the array.  It is in an external case and connected to a separate SATA extension card with an eSata port. 
• When I first connected the two drives externally, only one showed up under UnAssigned drives along with my warm spare.  My backup external drive didnt show up at all and the 2nd new drive wasn't there either. 
• I went through several stages of trouble shooting trying to find a cabling issue before realizing all 4 drives were showing up under "unassigned" on the Dashboard page.  Clicking the fresh button for UnAssigned Drives on the Main page resulted in all appearing there.
• I've never had to hit refresh before to make the main page show everything but rightly or wrongly I assumed this was a non-issue and went on.  
• That said those cables have been checked and re-checked last night.  The cables that go from the eSata ports to the MB internal ports are barely long enough and have a 180 bend at the end.  They are locking / latching cables so shouldn't come loose.  The cables plugged into the drives are also latching.

 

So what now?

1. How could this be happening simultaneously to both drives? Does this sound like a bad drive issue?  
2. Would you stop the preclear, check cables, and restart? or let it play out? 
3. I'm not sure how I could be getting errors on two drives at exactly the same time if it were a bad drive issue or a cabling issue.  Both seem unlikely. 

 

Any thoughts or insight is appreciated. 

 

UDMA errors are errors where the drive momentarily drops it's connection.  Usually due to bad / poor cabling (or power delivery)  The odd one here and there is nothing to worry about, but when if you get thousands (or continually) then something needs to be done

30 minutes ago, Squid said:

UDMA errors are errors where the drive momentarily drops it's connection.  Usually due to bad / poor cabling (or power delivery)  The odd one here and there is nothing to worry about, but when if you get thousands (or continually) then something needs to be done

 

Thanks and I thought that is what this error was.  I've cleared at least 8 drives using this same method / cabling configuration and never had a single issue, including just recently I cleared two of the same model drive number in the same method and had no issues.

I've been trying to figure what changed that would affect both drives at the same virtual moment.  One thought is that the power is delivered via a single IDE to 2 - SATA power cables.  When the first pre-clear finishes I'll probably stop it and check the power cables.

I do have more drives connected at the same time than I ever have, but most of them were spun down when the errors occurred.  I have a Corsair HX850 platinum PSU and no GPU's so I don't think I'm short on power.

 

I don't like leaving "unexplained" issues.  While 6 is not a lot of errors, they happened in a string over a 4 hour period with seemingly no change or cause.  Maybe the power cable will turn up something.

 

 

22 hours ago, Squid said:

UDMA errors are errors where the drive momentarily drops it's connection.  Usually due to bad / poor cabling (or power delivery)  The odd one here and there is nothing to worry about, but when if you get thousands (or continually) then something needs to be done

 

So The rest of the 1st cycle was un-eventful.  The 2nd cycle was fine until it started zeroing and then the UDMA errors came back.

Both drives errored simultaneously and then the same drive single drive as yesterday has had a few other errors.

 

Does the fact that the UDMA issues are happening during zero'ing but not pre/post read tell us something?  It seems zeroing would possibly use more power?

 

Nothing on the power side looks loose or not fully engaged.  I can't imagine how data would affect both at exactly the same time.  The external IDE power connection for those two drives is the only thing on that socket from the PSU.  Most if not all the other drives are spun down during this operation so it isn't taxing the PSU.  

 

Ideally I would do drive replacement using the same interface and then open the case to swap the drives.  I'd like to troubleshoot and then try and get a full pre-clear to run prior to changing the two drives.

 

Any thoughts on how to troubleshoot?  I've cleared a number of drives this way and never had an error before.

 

So I'm certain the UDMA issue is only happening during the "writing zeros" operation.  I think its connected with the power cable not the data cables.

Power cable replaced I am now just writing zeros with preclear to both drives at once so we'll see what happens.