I'm currently building yet another ATLAS copy. I have everything except for my SAS cables. So currently, I am using a couple of the drives in RDM mode under ESXi using the breakout cable. Right now, I am using 2 3TB seagates and one WD EARS 2TB. I have not been able to get SMART reports from these drives with my current setup. I followed the instructions in Johnm's guide with the only change being -z instead of -r when setting up the RDM files.

 

I have decided to try preclearing the drives while waiting for my cables. The initial SMART command failed. Smartctl did not work either. Is there any way to get SMART working? Am I accomplishing anything by running a few preclear cycles without it? If there are no errors after 3 cycles, would it be okay to start my array up with these drives?

1) You could boot another OS besides ESX.

2) You could test on another machine/server.

 

Getting a baseline smart report is a good idea. Running any tool which writes every block will get sector reallocations done. badblocks -cc does double duty since it reads, then writes, does read verify, and finally re-writes the original value. It's also good to include some "randomized" movement to verify tracking. thrash does that.

I decided to let the preclear keep running until it finished and i woke up to this:

 

Jul  7 07:36:08 unRAID kernel: sd 3:0:2:0: [sdd] Unhandled error code (Errors)

Jul  7 07:36:08 unRAID kernel: sd 3:0:2:0: [sdd]  Result: hostbyte=0x01 driverbyte=0x00 (System)

Jul  7 07:36:08 unRAID kernel: sd 3:0:2:0: [sdd] CDB: cdb[0]=0x28: 28 00 6d 99 44 e0 00 01 00 00 (Drive related)

Jul  7 07:36:08 unRAID kernel: end_request: I/O error, dev sdd, sector 1838761184 (Errors)

Jul  7 07:36:08 unRAID kernel: Buffer I/O error on device sdd, logical block 229845148 (Errors)

Jul  7 07:36:08 unRAID kernel: Buffer I/O error on device sdd, logical block 229845149 (Errors)

Jul  7 07:36:08 unRAID kernel: Buffer I/O error on device sdd, logical block 229845150 (Errors)

Jul  7 07:36:08 unRAID kernel: Buffer I/O error on device sdd, logical block 229845151 (Errors)

Jul  7 07:36:08 unRAID kernel: Buffer I/O error on device sdd, logical block 229845152 (Errors)

Jul  7 07:36:08 unRAID kernel: Buffer I/O error on device sdd, logical block 229845153 (Errors)

Jul  7 07:36:08 unRAID kernel: Buffer I/O error on device sdd, logical block 229845154 (Errors)

Jul  7 07:36:08 unRAID kernel: Buffer I/O error on device sdd, logical block 229845155 (Errors)

Jul  7 07:36:08 unRAID kernel: Buffer I/O error on device sdd, logical block 229845156 (Errors)

Jul  7 07:36:08 unRAID kernel: Buffer I/O error on device sdd, logical block 229845157 (Errors)

Jul  7 07:36:08 unRAID kernel: sd 3:0:2:0: [sdd] killing request (Minor Issues)

Jul  7 07:36:08 unRAID kernel: sd 3:0:2:0: [sdd] Unhandled error code (Errors)

Jul  7 07:36:08 unRAID kernel: sd 3:0:2:0: [sdd]  Result: hostbyte=0x01 driverbyte=0x00 (System)

Jul  7 07:36:08 unRAID kernel: sd 3:0:2:0: [sdd] CDB: cdb[0]=0x28: 28 00 6d 99 45 e0 00 01 00 00 (Drive related)

Jul  7 07:36:08 unRAID kernel: end_request: I/O error, dev sdd, sector 1838761440 (Errors)

 

 

Does this mean that the drive is toast? I had my suspicions when my FreeNAS server kept on kicking it out of the RAID. However, it did pass 2 previous pre-clears.

 

Once my 3TB drives are done pre-clearing, I am going to boot straight from the unRAID stick and run badblocks. I don't see much about it on the forums here. Is it a new thing? Any reason that preclear seems to be the standard vs badblocks?

 

I'm a linux noob, so would this be a proper way to run it?

 

badblocks -wvsb 4096 /dev/sdX

 

Thanks.

badblocks -wvsb 4096 /dev/sdX

 

Yes, that will do a write test using 4k blocks, showing the progress and verbose messages.