zfs discuss - Jul 2007 - Why is "stop scrub" taking so long time???

I have begun a scrub on a 1,5TB pool which has 600GB data, and seeing that it
will take 11h47min I want to stop it. I invoked "zpool scrub -s pool"
and nothing happens. There is no message: "scub stopped" or something
similar. The cursor just sits there on a new line and blinks with no output.

I have a 32bit P4 (which is slow for ZFS - gives me rouglhy 15MB/sec), 1GB ram,
build 68.


What happens if I shutdown the computer, my zpool wont break? scrub doesnt write
anything, just reads?
 
 
This message posted from opensolaris.org

On Jul 31, 2007, at 5:44 AM, Orvar Korvar wrote:
> I have begun a scrub on a 1,5TB pool which has 600GB data, and  
> seeing that it will take 11h47min I want to stop it. I invoked  
> "zpool scrub -s pool" and nothing happens. There is no message:  
> "scub stopped" or something similar. The cursor just sits there
on
> a new line and blinks with no output.
I believe you''ve hit:
6580855 spa_scrub() should be called with spa_namespace_lock held,  
not the config lock

which was introduced in build 68.

I have a fix that''s going through final testing.
>
> I have a 32bit P4 (which is slow for ZFS - gives me rouglhy 15MB/ 
> sec), 1GB ram, build 68.
>
>
> What happens if I shutdown the computer, my zpool wont break? scrub  
> doesnt write anything, just reads?
ZFS guarantees consistency on disk, so rebooting your system will be  
fine.

''zpool scrub'' will do self-healing (which implies writes) if
necessary.

eric