Hi,

I seam to get a heap of these in my syslog:

 

Apr 26 20:43:14 Tower kernel: [30110.050251] hdf: dma_intr: status=0x51 { DriveReady SeekComplete Error }

Apr 26 20:43:14 Tower kernel: [30110.050261] hdf: dma_intr: error=0x84 { DriveStatusError BadCRC }

Apr 26 20:43:14 Tower kernel: [30110.050266] ide: failed opcode was: unknown

Apr 26 20:43:14 Tower kernel: [30110.204054] hdf: dma_intr: status=0x51 { DriveReady SeekComplete Error }

Apr 26 20:43:14 Tower kernel: [30110.204064] hdf: dma_intr: error=0x84 { DriveStatusError BadCRC }

Apr 26 20:43:14 Tower kernel: [30110.204069] ide: failed opcode was: unknown

Apr 26 20:43:14 Tower kernel: [30110.357874] hdf: dma_intr: status=0x51 { DriveReady SeekComplete Error }

Apr 26 20:43:14 Tower kernel: [30110.357883] hdf: dma_intr: error=0x84 { DriveStatusError BadCRC }

Apr 26 20:43:14 Tower kernel: [30110.357887] ide: failed opcode was: unknown

Apr 26 20:43:14 Tower kernel: [30110.511659] hdf: dma_intr: status=0x51 { DriveReady SeekComplete Error }

Apr 26 20:43:14 Tower kernel: [30110.511668] hdf: dma_intr: error=0x84 { DriveStatusError BadCRC }

Apr 26 20:43:14 Tower kernel: [30110.511673] ide: failed opcode was: unknown

Apr 26 20:43:14 Tower kernel: [30110.511680] hde: DMA disabled

Apr 26 20:43:14 Tower kernel: [30110.654301] ide2: reset: success

Apr 26 20:47:39 Tower kernel: [30375.075944] hdh: dma_intr: status=0x51 { DriveReady SeekComplete Error }

Apr 26 20:47:39 Tower kernel: [30375.075955] hdh: dma_intr: error=0x84 { DriveStatusError BadCRC }

Apr 26 20:47:39 Tower kernel: [30375.075960] ide: failed opcode was: unknown

Apr 26 20:47:40 Tower kernel: [30376.305116] hdh: dma_intr: status=0x51 { DriveReady SeekComplete Error }

Apr 26 20:47:40 Tower kernel: [30376.305126] hdh: dma_intr: error=0x84 { DriveStatusError BadCRC }

Apr 26 20:47:40 Tower kernel: [30376.305132] ide: failed opcode was: unknown

Apr 26 20:47:41 Tower kernel: [30376.632073] hdh: dma_intr: status=0x51 { DriveReady SeekComplete Error }

Apr 26 20:47:41 Tower kernel: [30376.632083] hdh: dma_intr: error=0x84 { DriveStatusError BadCRC }

Apr 26 20:47:41 Tower kernel: [30376.632087] ide: failed opcode was: unknown

Apr 26 20:47:41 Tower kernel: [30376.785848] hdh: dma_intr: status=0x51 { DriveReady SeekComplete Error }

Apr 26 20:47:41 Tower kernel: [30376.785856] hdh: dma_intr: error=0x84 { DriveStatusError BadCRC }

Apr 26 20:47:41 Tower kernel: [30376.785860] ide: failed opcode was: unknown

Apr 26 20:47:41 Tower kernel: [30376.939619] hdh: dma_intr: status=0x51 { DriveReady SeekComplete Error }

Apr 26 20:47:41 Tower kernel: [30376.939629] hdh: dma_intr: error=0x84 { DriveStatusError BadCRC }

Apr 26 20:47:41 Tower kernel: [30376.939633] ide: failed opcode was: unknown

Apr 26 20:47:41 Tower kernel: [30377.093362] hdh: dma_intr: status=0x51 { DriveReady SeekComplete Error }

Apr 26 20:47:41 Tower kernel: [30377.093371] hdh: dma_intr: error=0x84 { DriveStatusError BadCRC }

Apr 26 20:47:41 Tower kernel: [30377.093374] ide: failed opcode was: unknown

Apr 26 20:47:41 Tower kernel: [30377.093380] hdg: DMA disabled

Apr 26 20:47:41 Tower kernel: [30377.269291] ide3: reset: success

 

My read and write speed drops to basically nothing when this occurs. The only fix to get the speed back is a reboot. I am using 4.0b10 on the Lime Tech supplied Intel board etc.

 

Any suggestions on where to look? My IDE cables are as short as I can make them.

 

Daniel

"BadCRC" is referring to the CRC check done on each packet transfered from the hard drive to the controller.  If you are confident in your cables, the next thing to check is power.

 

Also, in the log "hdf", "hdg", "hdh" are all on same PCI disk controller.  You might try swapping the two PCI disk controllers and see if the errors change to a different set of drives (would point to disk controller being flaky).  If the errors stay on the same set of drives after swapping controllers, then report back & we'll try to think of something else

Any suggestions on where to look? My IDE cables are as short as I can make them.

 

Daniel

are you using flat 80 conductor IDE cables... or older 40 conductor ones.  The older cables would result in errors at higher speeds and might cause the symptoms you described. the drive would fall back to a slower access mode.  The 80 conductor cables often have a blue connector at one end... if both ends of your cables are black you might have a 40 conductor cable and it would be a prime suspect, regardless of how short it is.

 

Joe L.

I have replaced both PCI controller cards with other cards, and the problem continues with either set of cards.

 

It seams to get much worse if I am writing to the drive at the same time as reading, and doesnt seam to be drive specific.

 

I am using shortish (The cant be too short because of the are in a CM Stacker case) 80 pin cables, Not 40 Pin ones.

 

At the time of this log dump, only 4 drives were spin up. I have 2 x 450w PSUs in the case.

 

Daniel

Looking at the syslog, there appears to be a file system corruption.  Please follow the instructions at the end of this thread.

 

If the 'reiserfsck' program detects errors it will recommend how to re-run to fix the problem.  Usually it recommends adding '--fix-fixable' switch.

I have checked all drives, and no corruption was found.

Looks like it is power related.. 14 drives in my Unraid. I have added a third power supply into the mix to power some of the drives, and so far all of my problems have gone.

Most new power supplies have multiple "rails" for +12V.  For example a p/s might have "+12V1" and "+12V2".  The p/s specs might say "36A supported on +12", when in reality each rail is limited to only 18A (for example).  This means you will only have 18A available for peripherals, ie, hard drives.

 

For unRAID, it's best to try to use power supplies with only 1 +12V rail.