Advice on really stressing a HD

[nimbu]

Sorry long wall of text below, but would appreciate any insight to my buried (bolded) questions below.

 

Background:

 

I am planning to migrate from a Synology off the shelf NAS to an Unraid box.  As part of all this I fired up a test box so that I could explore/familiarize/ learn more about unraid.  Rough spec of the test machine:

 

HP Gen 8 Microserver

8GB Ram

1 x 512GB 840 Pro Cache drive

3 x WD 12TB array single parity (calling them Drive 1, 2 and 3 for below).

 

So a little while back, I awoke to seeing that my unraid array was "degraded" and that Drive 2 had fallen out of the array, no obviously smart issues.  After a little google fu and no understanding of how unraid works, instead of assigning the Drive 2 back to slot 2, I instead assigned it to slot 4.  This increased the size of my array and the original content of Drive 2 was still being emulated.

 

I did post on the forums after the fact and found the error of my ways.  A few days later, I again awoke to errors in Unraid, this time I was seeing a large amount of 

"Seek error rate" errors being reported with this just going up and up.  The server was running but often slowing down and at one point locked up, so I stupidly resorted to a power cycle, which you guessed it corrupted the filesystem on Drive 3.  I couldnt retrieve the super block so wasnt able to mount this disk.  Again just to stress this was not production and all of the data is on my current NAS and is backed up offsite.  I had a busy time with work coming up so I powered down.

 

A few days ago I began to start my build off, I ordered another 12TB from Amazon on prime day, which is currently being precleared (I set up a trial unraid on my spare nuc specifically for the purpose of the preclearing.  Just for reference, the build is "new" everything, mobo, cpu mem case etc, the HP Gen 8 is being retired.

 

I powered the HP box up and strangely no errors on Disk 1, I even got a notification to tell me that Drive 1 had returned to normal operation (array itself screwed).  I had written off Drive 1 at this point however I thought I would have a fiddle.

 

I powered down the box, removed the other 2 x 12 TB, restarted unraid and opted for a new configuration.  I then started off a preclear with pre read and post write disabled completely expecting the drive to fail..... 20 hours later, the drive reports all is well!

 

# Step 1 of 3 - Zeroing the disk: [20:13:19 @ 164 MB/s] SUCCESS #

# Step 2 of 3 - Writing unRAID's Preclear signature: SUCCESS #

# Step 3 of 3 - Verifying unRAID's Preclear signature: SUCCESS #

 

Obviously this wasn't a complete preclear, but before I do that does anyone have any other recommendations of what I could do to really stress the drive out?  As it stands I am a little untrusting of the drive.  (Though I also have a suspicion its the server given that Drive 2 fell out of the array and for some reason it wont detect my 2TB Green drives..).

 

If it were you and it passed another full preclear (maybe 2-3 passes) as well as anything additional that is suggested above, would you trust this drive?  If so what role would you assign it?  I was thinking parity 2 or just as a unassigned device in a non important role.

 

RMA isnt an option as the drive was shucked and I didnt keep the enclosure.

 

Regards,

 

[trurl]

I found the other thread you were talking about, but it was never determined what the problem was since we got involved after you had already done some things and we never asked for more information.

 

Bad connections are far, far more common than bad disks.

 

Go to Tools - Diagnostics and attach the complete Diagnostics ZIP file to your NEXT post in this thread.