Hi guys,

 

I just finished up building my server. Ran the preclear on most of the drives with success. One drive didn't succeed and one had relocated sectors in the SMART report. I ran the preclear on 8 SATA drives simultaneously and I figured that's why it must've failed. Re ran it on the drive that didn't pass and it passed. The drive with the single SMART error is being rechecked at the moment and we'll see what happens.

 

I thought that I should start the array with the known working drives and copy what I can, just not to waste any more time. I have about 4 TB of data on some old drives in my HTPC. Downloaded TeraCopy and attempted to copy it all at once. Left it running and it was stuck on 1.7 TB and some files were showing CRC errors.

 

What's the safest method to copy those files over. So I know they will be working and could delete them safely from the old drives in order to add the old drives to the array and have them precleared/formatted etc...

 

I have an ESATA dock and could copy them this way. Is it better/safer than Gig E ? Should I leave the Parity sync running ? Or should I cancel it and run in manually as soon as the copying is done ?

 

I'm lost and I don't know how to proceed.

 

Thanks in advance...

You could use Teracopy and enable CRC checking.

 

http://www.codesector.com/teracopy.php

 

John

There should not be CRC errors using Terracopy. This indicates a hardware issue. I'd run a memtest overnight. Please post a syslog. What SMART errors are you seeing? Please post the SMART reports.

Hi John, I did use Teracopy. It got stuck and I was getting CRC errors as I mentioned in my OP.

 

Hi dgaschk, I'll post a syslog and SMART reports as soon as I get home. The preclear of that drive should be done too. I'll then stop the array, reboot and run memtest overnight and will post that too. I'm retrying Teracopy at the moment and just copying the content one drive at a time. Should be done within 3 Hours and will let you know how it goes and if there's any errors.

 

I had 2 GB of Crucial ram and the mobo wouldn't even POST on that. I'll have to return it. Which brand of RAM do you suggest I get instead ?

 

Thanks for the help guys.

 

What motherboard are you using?  That will affect memory suggestions.  It would probably help if you could get the part number of your memory as well.  Some boards are more finicky than others when it comes to memory module compatibility.

Hey Rich, The motherboard I'm using is in my sig The Crucial ram part number is CT25664BA1339. The Transcend I'm not exactly sure. I'll have to get back to you on that.

Hi dgaschk, The preclear of the drive that I said hard SMART errors (they were in the email I got the first time I tried to preclear it, it said 1 sector relocated or something like that) I attached the first email I got which has the error and the new email which reported no error during the preclear (as far as I can tell).

 

The syslog wouldn't matter I think since I rebooted the server since. I could post the new one if you want but I haven't had an error since so I don't know if it's helpful.

 

Also Teracopy is working like a charm at the moment since I'm copying from one hard drive. It still haven't finished copying the files over, it's taking longer that I expected. I will verify CRC when it's done and will post if there was errors so we can troubleshoot.

 

The SMART error I was talking about is attached, it was from the preclear result and not from a SMART test. If you still want a SMART report I'll try to attach one for you.

 

P.S. 1 : To make things clearer as I see I haven't been so clear. 1.txt is the drive that did NOT pass preclear originally. 2.txt is the same drive after I have rerun it just to make sure it didn't choke because I was running 8 simultaneous preclears, it passed this time. 3.txt is the drive that had SMART error the first time, it did pass preclear. 4.txt is the rerun of that drive, the letter has changed because I enabled AHCI in the bios but the serial as you'll notice is the same.

 

P.S. 2 : I'm not sure if it's relevant or not. The hard drives were once jumpered but now aren't and I removed the jumpers, precleared them then formatted them since installing them with the new motherboard. I had a Gigabyte motherboard but ditched that based on Joe's advice since that mobo had HPA issues. Removed the HPA correctly and then precleared after installing them in the new mobo. Syslog was reporting HPA but isn't now since I fixed them. I also used the MBR 4K aligned option as you can see from the preclear reports.

 

Sorry for being such a n00b. It's my first time dealing with unRaid, I won't say it was smooth sailing but I'm learning everyday. If there is something that I did wrong please elaborate so I can try and do it correctly.

 

Abdulla

1.txt

2.txt

3.txt

4.txt

Missed your signature before.  Supermicro is the main brand I've heard a lot of reports of being particular about memory.  I'd recommend reseating your module and run a 24 hour session of memtest.  Errors can be very intermittent and require particular conditions to occur.  Memtest will try to expose these errors for you.  If memtest passes, I'd also recommend checking all of your component connections.  If memtest fails, people have recommended Kingston KVR1333D3N9/2G.

 

Edit:  you can also check the Supermicro C2SEE page for memory modules that have been tested with the board.

No problem Rich. I'll do a memtest run to make sure. I'll get the Kingston RAM if it fails.

 

Thanks a lot for your help

I would pre-clear the drive in [1,2].txt a few more times. The first failure was a bad one. I'd pre-clear the drive in [3,4].txt  a couple of more times to make sure the reallocated sectors does not change. The quality of cheap consumer drives is bad. I've started running 4 pre-clear cycles on all new drives. It takes days but I have confidence in the drives I add to my array or provide to a client. I suggest you keep a pre-cleared spare on hand for quick replacement. I used to use my spare as a cache drive but now I have several on hand anyway. If you do use a cache try to keep the original copy for at least 24 hours so that you know the data has been moved into the array (or increase the mover schedule.)