scrubbing is buggy "somewhere"

[haer22]

Guys, remember when you helped me in IRC 2-3 weeks ago. Same problem popped up again. When I import the pool which is scrubbing the system just silently reboots after ca 1 minute.

I have upgraded to 1.5.2 and still the same problem.

I finally remembered the sysctl kstat.zfs.darwin.tunable.zfs_no_scrub_io=1 thing which enabled me to import without the reboot. zpool status indicated that scrubbing was still going! But at a very nice rate: 1.5GBps . I stopped it (zpool scrub -s zeus) and now everything is fine again.

I have moved disks around onto various combos of cabinets, eSata, USB but no change, i.e. reboot after 1 minute on all combos.

To eliminate something in my setup, maybe someone could start scrubbing and then reboot the system and see if the problem can be reproduced.

Is there any debugging I can turn on to help you find the problem?

[haer22]

I have started scrubbing, exported and then imported on my other pool (gaia) and that worked fine. No reboot and the scrubbing continues.

So it seems like the problem is somewhere around my pool zeus.

As I can provoke the reboot, is there a debug-flag I can turn on and then you can check the log for problems.

[haer22]

And here is my config:

Code: Select all

[ihecc:~] root# zpool status -L
  pool: gaia
 state: ONLINE
  scan: scrub in progress since Sun May  1 11:38:46 2016
    2,75T scanned out of 17,6T at 274M/s, 15h48m to go
    0 repaired, 15,63% done
config:

   NAME        STATE     READ WRITE CKSUM
   gaia        ONLINE       0     0     0
     raidz1-0  ONLINE       0     0     0
       disk13  ONLINE       0     0     0
       disk14  ONLINE       0     0     0
       disk15  ONLINE       0     0     0
       disk16  ONLINE       0     0     0
   logs
     disk3s2   ONLINE       0     0     0
   cache
     disk3s4   ONLINE       0     0     0

errors: No known data errors

  pool: zeus
 state: ONLINE
  scan: scrub canceled on Sun May  1 11:36:26 2016
config:

   NAME        STATE     READ WRITE CKSUM
   zeus        ONLINE       0     0     0
     raidz1-0  ONLINE       0     0     0
       disk4   ONLINE       0     0     0
       disk6   ONLINE       0     0     0
       disk7   ONLINE       0     0     0
       disk5   ONLINE       0     0     0
     raidz1-1  ONLINE       0     0     0
       disk10  ONLINE       0     0     0
       disk9   ONLINE       0     0     0
       disk11  ONLINE       0     0     0
       disk8   ONLINE       0     0     0
   logs
     disk3s3   ONLINE       0     0     0
   cache
     disk3s5   ONLINE       0     0     0

errors: No known data errors
[ihecc:~] root#


raidz1-0 is 4*3TB whereas radiz1-1 is 4*4TB. Both logs are 8 GB and the caches are 56 GB each