I recently replaced Disk1 (2 TB Western Digital EARS) because it was extremely slow when copying data and stuttered when playing movies (all other drives were fine).  I replaced the drive with a 2TB Seagate drive, it showed up blue, and I rebuilt it.  I went on vacation thinking I'd come back to a completely rebuilt system.  When I checked a week later, it shows a red flashing light by disk 1 (all other drives are solid green).  I tested copying 5 GB to it and it went through quickly.  Before I replaced the drive, it would take HOURS for a single DVD instead of minutes as normal (GBIT network throughout).

 

I hit "Clear Statistics" before copying the data.  After copying a test to it, it only the parity drive showed any writes and no drive showed any reads.

 

I'm running version 4.7.

 

Any thoughts/suggestions?

It sounds like your drive is offline. You are running in a simulated mode using the parity to simulate the missing drive. if another drive fails you will most likely loose data.

 

Post a syslog.

Thanks for the response JohnM.  I've running the pro version for years, but I basically set it and forget it (except for adding discs).

 

I looked for the syslog and I don't see one.  I read http://lime-technology.com/forum/index.php?topic=9880.0 , which said to look in "//tower/log/syslog" , but I don't have that folder/share.  When I look on my tower, I have disk 1 through 10, the 4 shares I created, and the flash folder.  Where do I find the syslog or does it have to be enabled?

 

Also, how do I tell if I'm running in simulated mode?  I read the FAQs and it said to test the cable/connection/port, so here's what I've done so far:

 

1)  Reseated the SATA and power cables -- still red

2)  Swapped to a new SATA cable -- still red

3)  Swapped SATA cable to a new spot on the add on card -- no difference.  Also, other hard drives are on the same card and function properly.

After swapping slots on the add on card, I had to reset the drive (again, like I did before I left for vacation), so I won't be able to troubleshoot for a while.  Initially when I swapped the drive, it showed up blue, but when I enabled the drive as Disk 1, it went red.  So far, it's still red.  I have 16TB space, so the rebuild takes a while.

 

Thanks,

Jon

 

The blinking red light means that a write failed to the drive. Whether that's from a bad connection (e.g. cable or port) or a bad drive remains to be seen.

 

If a drive is disabled, then you are automatically simulating any writes to that drive. Those files get distributed to your other drives, and you're without parity protection. It's really a good idea to minimize writes to an array with a disabled drive.

 

You can do two things to help others determine what's going on:

1. Post a syslog. Type the following command into the console or a telnet/SSH session: cp /var/log/syslog /boot/syslog.txt

That will put a copy of your syslog to the flash share of your flash drive. Then, go to that share from another machine, and upload that txt file to this forum.

 

2. Run a SMART test, and post the results. See here for instructions on how to do this.

 

Thanks Kenoka.  I'm attaching the syslog.txt (zipped due to size).  The light by Disk 1 isn't blinking red, it's solid red.  The SMART test says it could take 2 - 4 hours.  Should I wait to do this until the rebuild it done?  It's at 1.5% and no drives show any errors.

 

I've replaced the drive and the cable and swapped SATA ports, so those can't be bad.  It could be the add on card or the motherboard itself, but 2 other drives are working properly that are attached to it.

syslog.zip

Do the short test first. You can cancel the long test and then start the short.

Okay, after a brief glance at your syslog, it looks like device ata11 (ST2000DL003-9VT166) is the culprit. It doesn't look like it's responding at all to the system. If it's connecting now, then you may have had a loose connection. If it's rebuilding now, then you may be okay, but run the short SMART test to see if there's anything problematic.

Will do once I get home.  Disk 1 is the one that's red, but it's the one I already replaced.  Disk 6 is the one you listed.  It shows green.  Once I run the SMART test and if it comes back bad, if I replace it will I lose any data due to Disk 1 showing red?

 

Thanks again Kenoka.

 

**EDIT** aaah, wait.  Disk 1 and 6 are both ST2000DL003-9VT166, just with different serial numbers and slots.  Disk 1 is host 11, which would match up to ata11, correct?  I don't see any mention of ata on the devices page or on the main page.  Disk 1 is the one I just got in as a replacement for the WD Green that was going bad.  I know drives can come bad in the mail, but it just seems like an odd coincidence.  I'm running the SMART test on Disk 1/host 11 and I'll post later today.

You will have to rebuild disk1 first.  Otherwise if you attempt to replace disk6 you will have 1 missing drive and 1 disabled drive, and will lose data.

I'm running "smartctl -d ata -tshort /dev/sdj (sdj is disk 1/host 11), and it comes back with:

"Smartctl:  Device Red Identity Failed (not an ATA/ATAPI device)"

"A mandatory SMART command failed:  exiting.  To continue, all one or more '-T permissive' options"

 

When I run "smartctl -d ata -tshort /dev/hdd (hdd is disk 6, which is the other ST2000DL003-9VT166 drive), I don't get this error.  It goes into "START OF OFFLINE IMMEDIATE....", sending command, drive command, testing has begun, please wait 1 minutes for test to complete, etc.

 

Both drives are 2TB Seagate and both SATA.  Am I doing something wrong or does this confirm that I received a DOA drive?  I can attach the Disk 6 smart test if it'll help, but I figured it wouldn't matter since it won't run on Disk 1.

 

The other odd thing is that before I put it in the unRAID system, I plugged it into a SATA HDD bay that I have at work just to make sure it showed up (I've seen a few drives shipped DOA in my days) and it showed up as expected.  I didn't format or initialize it to test further however.

Have an open port to try to run the smart test?

Have an open port to try to run the smart test?

The SMART test runs fine on Disk 6 but not Disk 1, so it should be a drive/controller card/cable issue and not an open port issue, right?

 

Just to make sure I'm doing this right, I attached a smarthdd.txt file.  This is from Disk 6, which is named hdd in parenthesis on the Devices tab.  The same commands on sdj (disk 1) give me the errors in my previous post.

 

Please let me know if Disk 6 looks fine, and what I should do about Disk 1.  I'm assuming I need to shut the system down, pull the drive out, test it with Seagate's software, and RMA it (assuming it comes back okay).

smarthdd.txt

 

The SMART test runs fine on Disk 6 but not Disk 1, so it should be a drive/controller card/cable issue and not an open port issue, right?

Correct. What I was asking was could you switch Disk1 to another port to attempt to run the test.

Just to make sure I'm doing this right, I attached a smarthdd.txt file.  This is from Disk 6, which is named hdd in parenthesis on the Devices tab.  The same commands on sdj (disk 1) give me the errors in my previous post.

 

Please let me know if Disk 6 looks fine, and what I should do about Disk 1.  I'm assuming I need to shut the system down, pull the drive out, test it with Seagate's software, and RMA it (assuming it comes back okay).

Disk6 is good.

Correct.

Aaah, okay.  I understand now.  I'm at work and running all this remotely, so I'll have to do that when I get home.  I thought you meant the port option in puTTy.  I'll test that tonight.  Thanks Bryan.

I moved Disk 1 from one controller card to another controller card (by board has 6 slots, all of which are in use).  Disk 1 moved to sdi.  The command worked, so I'm guessing it's the controller card (cable, drive and SATA port have all been changed).  I'm attaching the SMART test.  Everything looked okay except for "UDMA_CRC_Error_Count", but I don't know what's too high considering the more important readings were zero (Reallocated, Current Pending, and Temp).

 

Another question regarding the 3 drive letters in the parenthesis ( -- (sdj), (hdd), etc).  I figured it would be good to run the SMART test on the only drive left on the 2nd controller card to see if it comes back bad.  That would almost certainly mean the controller cards is bad, correct?  Disk 2 is the only drive left on that card, which is labeled (sdj).  SDJ is what disk 1 was called before I swapped slots.  How would that happen?  I thought the 3 digits were based on drive and/or port?

 

The SMART test ran on SDJ.  The results look basically the same as SDI.

 

Also, I ran an extended Western Digital test on the drive that I originally replaced that was too slow (movies stuttered, data copy was crazy slow) and both the regular and extended tests came back fine.

 

What gives?  I don't see where my problem is.  Maybe it's on those specific 2 ports on the 2nd controller card?

smartsdi.txt

smartsdj.txt

The disk assignments (sd_) change on every reboot based upon which disks respond first. In general, when you have a stable configuration, the disks will tend to respond in the same order. But any time you make changes, it will change. It's not anything to worry about.

UDMA_CRC_Error_Counts seems to happen when the SATA interface is having issues. Try new cables and make sure the power supply connection is good.

I replaced the cable and swapped the he'd to a different Controller card, so I'd like to see if at fixes the errors. The PSU is a good 80+ Corsair w 700ish watts, so I'd like to make sure e issue isn't fixed before replacing it.  Is there a way to reset these counters so I know if the errors are new errors?

 

It's still rebuilding without errors so far, which I'd think is a good sign.

It's still rebuilding, but this is the first time it's gone this far (85%) without errors, so it sounds like the controller card is bad.  Is there any way to see what type of cards I have in my system now through unRAID?  I can go back through Newegg, but I buy all my work stuff through there too, so I could be searching for a while.

 

I know when I bought mine, I went off the approved controller thread on these forums, but I don't see it now, so it must have been replaced.  My sata connectors come off the top, so it's not the Adaptec listed here:  http://lime-technology.com/forum/index.php?topic=12404.0

 

The drive labeled hdd is not configured correctly in BIOS. It should have a sdX identifier. Make sure all SATA ports are set to AHCI only. The should not be combined or legacy mode.

dgaschk -- when I make the change to AHCI only, is there any affect on the data?  Typically in windows (I don't know Linux), you don't want to change to/from AHCI unless you're going to format, so I wanted to make sure before I did it.

 

The rebuild came back fine, so it must be that controller card.  I found the cards I currently have in the system (Syba) -- http://www.syba.com/index.php?controller=Product&action=Info&Id=864 .  I'm guessing either of the 4 port controller cards suggested here http://lime-technology.com/forum/index.php?topic=12404.0 will work as a replacement.

dgaschk -- when I make the change to AHCI only, is there any affect on the data?  Typically in windows (I don't know Linux), you don't want to change to/from AHCI unless you're going to format, so I wanted to make sure before I did it.

No effect. Just reassign the drive after the reboot.