pre-clear fails on one port but not another

[CaptainTivo]

I just shucked a WD100EMAZ and put it in my server to run pre-clear.  The pre-clear failed during the initial read so I put the drive back into the case and attached to PC to run the WD Data Lifeguard diagnostics on it.  I ran the long test (12 hours) and it passed.  I put it back into the server and ran pre-clear but it again failed during the pre-read.  I then shut down the server and put the drive into a different bay with a different SATA port and ran pre-clear again.  It's not finished yet but it's 80% into the the post-clear read and looks like it will pass.

Here's the thing: the first two times, I inserted the drive into a hot-plug bay while the array was stopped but the machine was up and running.  I have done this before to mount unassigned disks to do backups and it worked fine.  Maybe there is something special about pre-clear that causes it to fail when I insert the disk hot-plugged?  Attached is the syslog.  It looks like the pre-read is working fine for a while and then the disk just drops offline!

 

Log file:

Things seem fine until here:

Aug 30 11:38:15 Tower kernel: sd 7:0:12:0: [sdg] Unaligned partial completion (resid=12284, sector_sz=512)

 

then a string of read errors and then:

Aug 30 11:43:12 Tower preclear_disk_1EGEK2VZ[3582]: /usr/local/emhttp/plugins/preclear.disk/script/preclear_disk.sh: line 1267: echo: write error: No space left on device

 

Then this:

Aug 30 11:38:53 Tower unassigned.devices: Disk with serial 'WDC_WD100EMAZ-00WJTA0_1EGEK2VZ', mountpoint 'easystore' is not set to auto mount and will not be mounted.

which seems to say the disk was umounted.  Then a bunch more read errors and finally:

 

Aug 30 11:43:12 Tower preclear_disk_1EGEK2VZ[3582]: /usr/local/emhttp/plugins/preclear.disk/script/preclear_disk.sh: line 1267: echo: write error: No space left on device

Which is odd since it was supposed to be reading the disk.  "no space left on device"?

 

Anyway, I guess I need to know if I should not use the hot-plug method to mount my external backup disks.  Or if the SATA port that this particular one is on is bad in some way.

 

Thanks for any help.

 

 

 

 

 

tower-syslog-20200831-0315.zip

[Frank1940]

I was under the impression that Unraid did not support 'hot-plugging'.  There is some stuff that has to happen at the BIOS and hardware level for hot-plugging to work.  It could well be that this is occurring successfully on your system but Unraid is not properly mounting the disk for actual use.  I would suggest that you shut the server down (after the present pre-clear cycle finishes) and try it again in that hot plug bay from a cold start. 

[JorgeB]

Possibly unrelated but you should upgrade the LSI's firmware to latest (20.00.07.00)