Hi, I'm having trouble copying my data to my unraid server the last few days.

 

Teracopy keeps reporting "specified network name no longer available". After looking at the syslog I noticed line after line of handle_stripe read errors that happen at the same time as the network errors.

 

I have to cancel the transfer, delete the incomplete file, then it always works again straight away and transfers succesfully. Then an hour or so later I'll get a transfer error again.

 

At first I thought it might be my new gigabit switch I'm using but then thought it must be the drive I added last which is giving the errors. 1.5Tb Samsung F2.

 

Is the drive dieing and would I be advised to replace it? Also will all the data I've been copying to it be OK? Teracopy has never flagged up any CRC errors yet but I don't have it CRC test everything as it takes twice as long.

 

I've checked the SATA cable and it's in properly. Once I've finished copying data I'll power down the server and try a different SATA cable.

 

I'm using unraid pro 5.0-beta8b and Windows 7 from main PC.

 

I've included the smart report from the drive that's giving me problems and as it's too big to upload the text file here's my syslog: http://pastebin.com/ba8dFgEL EDIT: didn't think to zip syslog text so done that now too and included if that's the prefered method.

 

Thanks for any help.

SMART_report.txt

syslog.zip

Don't worry about your network. Definitely, something's up with sde (SAMSUNG_HD154UI_S1XWJ1KS908640). The integrity of copies will be suspect as long as the drive is throwing errors. Did you have a chance to preclear the drive?

No, I've just been clearing them through the web GUI as it's a lot faster than preclearing and I've been leaving it doing it over night as I've been transfering all my data drives from main PC so not bothered about the downtime.

 

Is it advisable to pre-clear drives instead of taking the array offline and doing it through the web GUI?

 

Lines that look worrying from the SMART report are:

ID# ATTRIBUTE_NAME          FLAG     VALUE WORST THRESH TYPE      UPDATED  WHEN_FAILED 	RAW_VALUE
  1 Raw_Read_Error_Rate     0x000f   099   011   051    Pre-fail  Always   In_the_past 	574
  5 Reallocated_Sector_Ct   0x0033   090   090   010    Pre-fail  Always       -        431
12 Power_Cycle_Count       0x0032   100   100   000    Old_age   Always       -        182
13 Read_Soft_Error_Rate    0x000e   099   012   000    Old_age   Always       -        573
187 Reported_Uncorrect      0x0032   100   100   000    Old_age   Always       -        43426
195 Hardware_ECC_Recovered  0x001a   100   100   000    Old_age   Always       -        565004116
196 Reallocated_Event_Count 0x0032   090   090   000    Old_age   Always       -        431
201 Soft_Read_Error_Rate    0x000a   099   098   000    Old_age   Always       -        41

 

That's a lot of errors!

 

Anyone have any advice on if I should just replace with a new drive or is it just a case of those sectors being unusable and the rest of the drive will be ok? (wishful thinking )

Don't worry about the raw values, especially raw hardware ECC values. All reads involve error correction so you'll see that value move all over the place. Value, worst, and threshold are where you should look to see how the drive manufacturer interprets it.

 

At this point we don't know if the problem is with the drive, data cable, power cable/splitter, SATA port, or whatever.

 

Preclearing is actually the best way to minimize array downtime when adding drives. Drives that are precleared can be added to the array after a simple format, instead of the lengthy overnight operation you mentioned. Of course, you still have to take the system down to add the drive, then let it preclear for a day, but when that's finished, adding the drive is quick & simple. Besides that, you get a more exhaustive test so you can have confidence the drive won't be a waste of your time. If these errors had come up during preclearing your array and copies wouldn't have been affected.

Thanks for your help, I'll have a read up on what those SMART values mean.

Surprisingly the drive gets a green tick in the SMART health report menu of unmenu with the following reports:

» reallocated_sector_ct=432

» reported_uncorrect=43450

» reallocated_event_count=432

 

Now, even more worrying is an old 300GB drive that has a red square instead of green tick and says:

» udma_crc_error_count=3

» ata_error_count=2

 

Don't think that was in red last time I checked. Suppose it's worth swapping that for a nice new 2TB drive incase both dodgy drives die together!

Both of those drives are bad. More than a handful of reallocated sectors normally (always from what I've seen) means that the drive is on a downward spiral of more and more reallocated sectors, until there are no more sectors to reallocate. The red X in myMain is telling you the drive's smart system is saying the drive is bad - a sure sign that the drive is not reliable.

@bjp999, I would have said the 300GB drive could be a cabling problem. The other one looks cooked.

Thanks for the replies. The 300GB drive now has a green tick again but I've ordered a new 2TB drive to replace the 300GB one just to be safe which should come tomorrow. I'll try a different cable in the mean time.

 

As for the 1.5TB I'd rather not have to throw it away but do have one more 2TB drive I'm about to move from my main PC now I've transfered it all to a drive in my unraid server. Would I be living too dangerously if I left it in and kept using it as if it did die it's just a case of replacing it with a new drive and parity will rebuild it? Or is that a bad idea and I should just replace it and lose the 1.5TB of space?

 

What other SMART tests can I run to check my drives?

 

Please could you have a look at the unmenu SMART report as other drives also have what look like minor problems and give some advice.

 

Also I haven't done a parity check yet. Now that I've transfered all the data and drives from main PC into my unraid server should I run a parity check and should I keep 'correct parity' selected?

 

Uploaded with ImageShack.us

 

Thanks again.

Also, not sure why it says last parity check 15174 days ago when I only build the server about 18 days ago. Ran the parity check last night and it found 0 errors.  

 

Any advice on the questions in the previous post? I currently have a brand new 2TB drive preclearing to replace the 300GB one (disk 9) and have an old 2TB drive from a USB2 enclosure to add so not sure if to use it to replace the 1.5TB drive (disk 4) and lose all that lovely disk space or use it as an extra 2TB to my array. Please help.