A little background on this:

 

I recently got a red ball next to one of my drives.  I ran a SMART report and it looked like based on Reallocated Sector Ct that my drive was failing.  I submitted an RMA with WD and got a replacement drive in the mail.  Let me back up.  After doing a SMART report, I did a quick Google search to see how to interpret this.  I took the first advice I saw without really understanding it, and did an RMA.

 

I got the drive, installed it, and booted up.  Everything seemed OK, rebuild was originally on-pace to finish in about 24 hours.  Then I checked a few hours later and speed had slowed to under 1MB/sec.  Rebuild was now scheduled for 22 days out.  Fast forward about a half hour, the web GUI had become completely unresponsive.

 

Like the bozo I am that knows a lot about computers but not enough to be an expert, I figured a shutdown -r now would reboot the system.  After about 10 minutes, I saw that was going exactly nowhere and hard PO/PO'ed the tower.

 

The server didn't come back immediately.  For some reason the BIOS stopped booting from my thumb drive at this point and decided it would try to boot from what I'm guessing was my parity drive (/dev/sda).  I fixed this issue and started up again, and that's where I am today.  The web GUI is up, but "Starting..." for a few hours now.

 

Running unRAID 4.7 if that's not apparent.

SMART.txt

syslog.txt

disk5, /dev/hdd has a lot of problems.  The OS keeps resetting the disk controller, and it has even dropped out of DMA mode (translation .. it is now in polling mode, and that is realllllllllly sloooooooooooooow )

 

Check the cabling to disk5.  If it is not an IDE drive, change the disk controller to use native SATA ACHI mode and not a legacy emulation mode.

 

Mar 30 22:10:59 dr-unRAID kernel: hdd: ide_dma_sff_timer_expiry: DMA status (0x60)

Mar 30 22:10:59 dr-unRAID kernel: hdd: DMA timeout retry

Mar 30 22:10:59 dr-unRAID kernel: hdd: timeout waiting for DMA

Mar 30 22:10:59 dr-unRAID kernel: hdd: DMA timeout: status=0x50 { DriveReady SeekComplete }

Mar 30 22:10:59 dr-unRAID kernel: hdd: possibly failed opcode: 0x25

Mar 30 22:10:59 dr-unRAID kernel: hdd: task_pio_intr: status=0x51 { DriveReady SeekComplete Error }

Mar 30 22:10:59 dr-unRAID kernel: hdd: task_pio_intr: error=0x04 { DriveStatusError }

Mar 30 22:10:59 dr-unRAID kernel: hdd: possibly failed opcode: 0x29

Mar 30 22:10:59 dr-unRAID kernel: hdd: task_pio_intr: status=0x51 { DriveReady SeekComplete Error }

Mar 30 22:10:59 dr-unRAID kernel: hdd: task_pio_intr: error=0x04 { DriveStatusError }

Mar 30 22:10:59 dr-unRAID kernel: hdd: possibly failed opcode: 0x29

Mar 30 22:10:59 dr-unRAID kernel: hdd: task_pio_intr: status=0x51 { DriveReady SeekComplete Error }

Mar 30 22:10:59 dr-unRAID kernel: hdd: task_pio_intr: error=0x04 { DriveStatusError }

Mar 30 22:10:59 dr-unRAID kernel: hdd: possibly failed opcode: 0x29

Mar 30 22:10:59 dr-unRAID kernel: hdd: task_pio_intr: status=0x51 { DriveReady SeekComplete Error }

Mar 30 22:10:59 dr-unRAID kernel: hdd: task_pio_intr: error=0x04 { DriveStatusError }

Mar 30 22:10:59 dr-unRAID kernel: hdd: possibly failed opcode: 0x29

Mar 30 22:10:59 dr-unRAID kernel: hdc: DMA disabled

Mar 30 22:10:59 dr-unRAID kernel: ide1: reset: success

Mar 30 22:11:05 dr-unRAID kernel: hdd: status timeout: status=0x80 { Busy }

Mar 30 22:11:05 dr-unRAID kernel: hdd: possibly failed opcode: 0x29

Mar 30 22:11:05 dr-unRAID kernel: hdd: DMA disabled

Mar 30 22:11:05 dr-unRAID kernel: hdd: drive not ready for command

Mar 30 22:11:09 dr-unRAID kernel: ide1: reset: success

 

Joe L.

Ok.  I'll shut down and check the cables - but it's a SATA cable right from the mobo to the drive.  No disk controller in between.  I'm inexperienced with hard disk controllers, I'm not sure that I know what you mean by taking the disk out of legacy emulation mode.  I never set that up... but I'll take a stroll through the BIOS and see what I can find.

I powered down (had to do a power button shutdown), and pulled/reconnected the SATA cable.  Powered back up, and now at least I'm not stuck in Starting...

 

Now at least the web GUI shows my drive is in Data-Rebuild again.  Looks like about 25MB/sec... not great, but not <1, either.  At this rate it will take about a day to rebuild the drive.  I find that acceptable - but I'd like to give things a few hours to see if the rebuild slows to a crawl again.  In the meantime I can try getting important data off the array in case I have to rebuild the whole thing.

 

The ball next to disk 3 (/dev/sde) is still red.  Is that normal?

 

Updated syslog attached.

syslog.txt

You somehow think that things magically repair themselves... and a reboot fixes everything. 

(or must have used Microsoft help-desk in the past)

 

Fix the BIOS option.  Set it correctly.

 

The "red" indicator is because a write to that disk failed.  It will not go away on its own, then disk must be re-constructed AFTER you discover and correct the reason a "write" to it failed.

 

If a "write" to /dev/sde has failed, you have multiple issues.  A failed "write" to a drive, AND a drive in emulation mode wanting to barely work.   

 

Tread carefully... you are on the brink of a two drive failure.  You might be able to solve one of them with the proper setting of your BIOS option to have it not emulate an IDE drive and run so slowly.

 

Then, you need to tackle the other one.  The drive that needs to be re-constructed since a "write" to it failed for some reason.

 

Joe L.

Thanks Joe.

 

I had to re-read your original post to catch on... disk5 has had nothing written to it yet, so it was completely off my radar, and I misunderstood.  You are educating me.  I wasn't sure why the last two drives were showing up as IDE before - now I understand a little more about this.  I do thank you for that, sir.

 

/dev/sde is the new device that is currently being rebuilt.  Does that change anything?

 

Updated syslog attached.

syslog.txt

Fixing the BIOS option absolutely worked, and the drive has been rebuilt.  Rebuild speed was around 55MB/sec.  All greens now and the performance is much better.