I just put together the pieces of the budget box from the wiki (thanks for that raj). I have 4.5.4 basic running on a 2gb verbatim stor_n_go flassh drive. I threw in one of the wd 2tb green drives I bought (jumpered 7- to start moving data over. I moved about 40gb of music over without issue. When I started in on my movies I started to get errors. Now I am having issues seeing the files over the network all the time (explorer.exe will hang and have to me killed) and the files I get to transfer are very slow (500k-1000k range). I also got unmenu installed (thanks Joe L) and in the "MyMain" "Smart" view it shows a big red X in the overall status of the drive. I have another identical drive sitting here I can test to see if that might be the issue, but I wanted to ask the experts before I RMA'ed anything. Thanks.

 

bux

 

ps. system log attached, if needed

syslog.txt

I just put together the pieces of the budget box from the wiki (thanks for that raj). I have 4.5.4 basic running on a 2gb verbatim stor_n_go flassh drive. I threw in one of the wd 2tb green drives I bought (jumpered 7- to start moving data over. I moved about 40gb of music over without issue. When I started in on my movies I started to get errors. Now I am having issues seeing the files over the network all the time (explorer.exe will hang and have to me killed) and the files I get to transfer are very slow (500k-1000k range). I also got unmenu installed (thanks Joe L) and in the "MyMain" "Smart" view it shows a big red X in the overall status of the drive. I have another identical drive sitting here I can test to see if that might be the issue, but I wanted to ask the experts before I RMA'ed anything. Thanks.

 

bux

 

ps. system log attached, if needed

Get a "smart" report on the drive with the "X" and post it.

 

From your syslog, I'm guessing you have not yet added a parity drive?  Or have you?, and that is the one with the red-X?

SMART report attached. No parity drive added yet, was planning on doing that after I moved most of the data over, from what I have read it will make things faster that way. Thanks.

 

bux

smart.txt

SMART report attached. No parity drive added yet, was planning on doing that after I moved most of the data over, from what I have read it will make things faster that way. Thanks.

 

bux

That drive has indeed failed.  The current normalized value for re-allocated sectors is 81, the threshold for failure is 141, and there are currently 949 re-allocated sectors.  What is WAY worse is that there are 65524 sectors pending re-allocation.  (I don't know of ANY drive with that many spare sectors)

 

Forget using that drive.  It is just about the worst I've seen other than a drive that is just plain dead.

 

I hope you  have copies of what you were loading on that drive, because it was never written to the defective sectors and a HUGE number of your sectors are defective...  If you had put a parity drive in place you could have reconstructed what you were attempting to write to the disk, but without one, let's just say you were writing to a nowhere... and thinking you were writing to the disk.

 

Joe L.

 

SMART overall-health self-assessment test result: FAILED!

 

  5 Reallocated_Sector_Ct  0x0033  081  081  140    Pre-fail  Always  FAILING_NOW 949

197 Current_Pending_Sector  0x0032  001  001  000    Old_age  Always      -      65524

Yes, I have all the data still on the old hdd's that I was moving it from (the reason I didn't add a parity disk yet). Time to RMA this bad boy. Just sucks to get drives that fail so fast. Thanks for the help Joe L.

 

bux

One last question: is there anything quick and easy I can do to completely wipe this drive before I send it back to newegg? I also moved some personal files over to it (tax returns, etc) that don't need to become public knowledge. Thanks.

 

bux

You could run the preclear script on it or use 'dd' to write full zeros to it.

fastest way to wipe it is:

 

dd if=/dev/urandom of=/dev/sda bs=8192

(that is not a misprint, the device used as an input is /dev/urandom spelled with a leading "u")

Writing at 100MB/s will take 10 seconds to write 1 Gigabyte.  For a 2TB drive (2000 Gigabytes) that would be 20,000 seconds or 5.55 hours.

 

Most drives can be written to at between 50MB/s and 100MB/s, so somewhere between 5.5 and 11 hours will clear the drive.

Type the command on the system console, or else you'll need to leave the PC you connected via telnet logged in for that period of time.

 

Just make absolutely certain that you type the device name of the correct drive.  Writing random data to the drive is not something that can be undone.

 

Joe L.

 

 

 

 

dd cmd running, thanks Joe. Once that is done I'll pull that hdd from the box and put in the other 2tb I bought for a parity drive. Let's hope that one has a longer life. Thanks again everybody.

 

bux

dd cmd running, thanks Joe. Once that is done I'll pull that hdd from the box and put in the other 2tb I bought for a parity drive. Let's hope that one has a longer life. Thanks again everybody.

 

bux

After you physically install the other 2TB drive, and BEFORE you assign it to the array, please consider using the preclear_disk.sh script on it.  It is specifically designed to exercise a disk and uncover early problems before you start trusting it with your data.  In fact, use the same script on EVERY drive before you add it to the array.  On a 2TB drive it will take the better part of a day, but that time is small compared to the time it takes to deal with a defective drive.

 

Find it here: http://lime-technology.com/forum/index.php?topic=2817.0

 

Joe L.

 

 

Will do Joe. I found out about it too late for the first drive, I guess I was just in a hurry to get the parts together and get it going.

 

Should I expect any output in telnet from that dd cmd? I've let it run over night, but still nothing. Thanks.

 

bux

Will do Joe. I found out about it too late for the first drive, I guess I was just in a hurry to get the parts together and get it going.

 

Should I expect any output in telnet from that dd cmd? I've let it run over night, but still nothing. Thanks.

 

bux

It will print nothing until it is done...  However, you can open a second telnet session and in it type:

killall -USR1 dd

 

This sends a  request to the running "dd" process to print its status in its session.  You can use the killall command to send the USR1 signal as frequently as you like to track how the command is progressing.

 

Joe L

 

Edit: corrected spelling of signal.

Thanks for the quick reply. I get this when trying that command:

 

root@Tower:~# killall -USER1 dd

USER1: unknown signal; killall -l lists signals.

root@Tower:~# killall -l         

HUP INT QUIT ILL TRAP ABRT IOT BUS FPE KILL USR1 SEGV USR2 PIPE ALRM TERM

STKFLT CHLD CONT STOP TSTP TTIN TTOU URG XCPU XFSZ VTALRM PROF WINCH IO PWR SYS

UNUSED

Sorry...  my typo

try:

killall -USR1 dd

 

Actually, I spelled the word right, but the command uses the abbreviated USR1, not USER1

 

Joe L.

 

 

That worked perfectly, thanks Joe. Shows 184gb done.... this is going to be awhile.

 

bux

That worked perfectly, thanks Joe. Shows 184gb done.... this is going to be awhile.

 

bux

You did originally say transfer rate to the drive slowed as it was being written to...  Apparently, it has not sped up... 

 

At least you'll be able to monitor its progress.  Did it report a data rate at which it was writing?

379060887552 bytes (379 GB) copied, 76216.7 s, 5.0 MB/s

 

is the line I'm getting out of it. It has been running for around 24 hours if I had to guess. At this rate I hope it's done in time to beat the 30 day RMA at newegg.

 

bux

It might be quicker to find an industrial magnet / degausser to use on it

379060887552 bytes (379 GB) copied, 76216.7 s, 5.0 MB/s

 

is the line I'm getting out of it. It has been running for around 24 hours if I had to guess. At this rate I hope it's done in time to beat the 30 day RMA at newegg.

 

bux

Actually, 76216.7 seconds=21.17111 hours.

 

or 18.75 GB per hour or 450 GB per day.  Looks like it would take a bit over 3 more days.

I was kidding about the 30 day thing. I can wait 3 days. I don't want to, but I can wait. Thanks again for all the help guys.

 

bux

So if I restart the pc I did the dd cmd on (this one), I'll have to start the process over from square 1, won't I? Thanks.

 

 

bux

So if I restart the pc I did the dd cmd on (this one), I'll have to start the process over from square 1, won't I? Thanks.

 

 

bux

You don't need to do your taxes again. 

But any files you had copied to the disk that had failed need to be copied again to its replacement.

 

I encourage you to pre-clear the replacement, just to weed out the disks that go defective in their first few hours of use.

 

Joe L.

So if I restart the pc I did the dd cmd on (this one), I'll have to start the process over from square 1, won't I? Thanks.

 

 

bux

I think I mis-read your last post.

 

If you know how far the dd command had progressed, you can start it again but use an additional argument requesting it start writing at a spot in the disk other than at the first sector.

 

You'll need to do the math on how many sectors to skip, but you can use something like:

dd if=/dev/urandom of=/dev/sda bs=8192 seek=100000000

The above command would seek to position 10000000*8192 or 819,200,000,000 where it will then start writing. 

( about 820 Gigabytes into it )  You can adjust the "seek" parameter to where you currently are in the writing of random data to the disk.

 

Joe L.

Once again, thank you very much. I had something hang up bad on this computer and I had to restart it. The thought of having to rerun those 2 days worth of cleaning wasn't pleasant at all.

 

bux

So I used dd if=/dev/urandom of=/dev/sda bs=8192 seek=120000000 to start the dd cmd back again around the 1tb mark on the drive. It has been running since then and shows 965001494528 bytes (965 GB) copied, 190214 s, 5.1 MB/s currently. To me this means that it should be done with the entire drive now (2tb shows up as 1.81tb formatted). Should I expect some output when the drive finishes or will it simply keep running? Thanks.

 

bux