On 24/04/2015 03:38, Matthew Ahrens wrote: > I don't think that we should assume that it can't change. In practice it > might > always be the same (for now) but I don't think that was the design intent.
I have opened an issue to keep track of this potential problem: https://www.illumos.org/issues/5867 > On Thu, Apr 23, 2015 at 5:00 AM, Andriy Gapon <[email protected]> > wrote: > > > A semi-theoretical question: can z_root - or, to be correct, ZFS_ROOT_OBJ > - > value change after a filesystem has its content replaced by zfs recv -F > of a > full stream? > It seems that that case wouldn't be handled correctly if the filesystem is > mounted and goes through zfs_suspend_fs() + zfs_resume_fs(). > > The reason I am asking is that I noticed this: > $ zdb -dddd pond/var 1 > .... > casesensitivity = 0 > SA_ATTRS = 4 > normalization = 0 > VERSION = 5 > ROOT = 3 > utf8only = 0 > DELETE_QUEUE = 2 > > $ zdb -dddd pond/var/crash 1 > .... > ROOT = 4 > utf8only = 0 > casesensitivity = 0 > VERSION = 5 > SA_ATTRS = 2 > SHARES = 7 > DELETE_QUEUE = 3 > normalization = 0 > > This may have very well happened because of my playing with the code, but > in > general the root object does not have a fixed ID. > > -- > Andriy Gapon > _______________________________________________ > developer mailing list > [email protected] <mailto:[email protected]> > http://lists.open-zfs.org/mailman/listinfo/developer > > -- Andriy Gapon _______________________________________________ developer mailing list [email protected] http://lists.open-zfs.org/mailman/listinfo/developer
