I've gotten myself into a bit of a bind and need some help.

 

When I first started building my server (in November of last year), I started with a couple of WD20EARS drives.  When I first installed them in the server I did NOT install the jumpers.  My memory is a bit sketchy but I believe I completely pre-cleared them without jumpers before realizing that jumpers were needed for unRAID.  I installed the jumpers and pre-cleared them once again.  Despite some error messages in the syslog (I know, I should have paid more attention to them) they added just fine to the array and I've been using them successfully ever since.  From an operational standpoint they've been fine.

 

Here is an example of some of the errors I'm getting in my syslog for one of the drives:

 

Mar  1 18:43:35 Tower kernel: ata10.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x0

Mar  1 18:43:35 Tower kernel: ata10.00: irq_stat 0x40000001

Mar  1 18:43:35 Tower kernel: ata10.00: failed command: READ DMA EXT

Mar  1 18:43:35 Tower kernel: ata10.00: cmd 25/00:01:af:88:e0/00:00:e8:00:00/e0 tag 0 dma 512 in

Mar  1 18:43:35 Tower kernel:          res 51/01:01:af:88:e0/06:00:e8:00:00/e0 Emask 0x1 (device error)

Mar  1 18:43:35 Tower kernel: ata10.00: status: { DRDY ERR }

 

Recently I tried upgrading to 5.0 beta (first 4, then 5, then 6) and in every instance all of my drives showed up fine EXCEPT for these two.  I posted a syslog in the beta 6 thread and limetech kindly pointed out that I had a huge number of errors on those two drives.  Apparently something in 5.0 is pickier about drive errors than 4.x.  Some more research today suggested that it all had to do with me first preclearing these drives without the jumpers installed.

 

So now I'm trying to figure out what my best option is to fix this problem.  I'm currently copying all of the data off one of those drives, and was going to remove the jumper, run preclear 1.7 with the -a option on it (I'm running unRAID 4.7 btw) and then put it back into service.  Repeat on the second drive.

 

Will this solve the problem or am I just asking for more headaches by doing this?  It seems that some people actually go and RMA their drives but that seems a bit drastic to me.

 

Any suggestions are greatly appreciated.  I can't believe what a pain those two little pieces of plastic have caused me...

 

-David

First clear the MBR with

dd if=/dev/zero count=100 of=/dev/sdX

where X is the drive number. Then run preclear with the -A (capital A) for an EARS drive with no jumper.

or run

preclear_disk.sh -z /dev/sdX

which zeros the mbr