GPT partitions on whole disk gmirror - questions about the metadata issue

[Matthias Petermann]

Hello,

in the meantime I did some further research and found out that newfs has 
a parameter to reserve some space at the end of the to be formatted 
device. The manpage newfs(8) states:

"-r reserved
     The size, in sectors, of reserved space at the end of the
     partition specified in special.  This space will not be occupied
     by the file system; it can be used by other consumers such as
     geom(4).  Defaults to 0.
"

Anyway, I did not see anyone using the -r parameter using for formatting 
a filesystem in a gmirror based device[1], and wondering at the same 
time as the default of "0" would mean to me that at some point - maybe 
only after the filesystem has filled up - it would touch the very last 
block on the disc and overwrite the gmirror meta data. Especially in the 
case of the illustrated workaround for GPT partitioned devices:

+---------------------------------------------------------------------+
|+-------------------------------------------------------------------+|
||+-----------------------------------------------------------------+||
|||+---------------------------------------------------+-----------+|||
||||  UFS Filesystem                                   |  gmirror  ||||
||||                                                   |  meta     ||||
||||                                                   |  data     ||||
|||+---------------------------------------------------+-----------+|||
|||                          /dev/mirror/data                       |||
||+-----------------------------------------------------------------+||
||                           /dev/gpt/data1                          ||
|+-------------------------------------------------------------------+|
|                            /dev/ada0                                |
+---------------------------------------------------------------------+

So what is the general advice to deal with that? Is it safe to use newfs 
without -r and why?

Kind regards,
Matthias

[1] http://www.wonkity.com/~wblock/docs/html/gmirror.html
[2] 
https://forums.freebsd.org/threads/freebsd-10-2-install-with-gpt-and-gmirror.55400/
[3] 
https://www.ateamsystems.com/tech-blog/installing-freebsd-9-gmirror-gpt-partitions-raid-1/

Am 21.11.2018 um 05:47 schrieb Matthias Petermann:
> Hello,
> 
> in section 18.3 of the FreeBSD handbook[1] there is a warning regarding 
> using whole disc mirroring with gmirror together with GPT:
> 
> "gmirror(8) stores one block of metadata at the end of the disk. Because 
> GPT partition schemes also store metadata at the end of the disk, 
> mirroring entire GPT disks with gmirror(8) is not recommended. MBR 
> partitioning is used here because it only stores a partition table at 
> the start of the disk and does not conflict with the mirror metadata."
> 
> Why is it that gmirror does not represent the mirrored device for the 
> levels above it in a way that does not allow access to the last block 
> containing the metadata? Would it be enough to simply mimic a smaller 
> device, one block less than the underlying providers?
> 
> In the case mentioned above, the workaround is to first partition both 
> providers using GPT and then form gmirrors from two GPT partitions each. 
> In this case, the metadata problem should not be critical because 
> gmirror exists within the partition itself. Here is my further question: 
> how does the file system (UFS) ensure that e.g. newfs does not overwrite 
> the last block of a gmirror in this setting?
> 
> Best regards,
> Matthias
> 
> [1] https://www.freebsd.org/doc/handbook/geom-mirror.html
> 

-- 
Matthias Petermann <matthias at petermann-it.de> | www.petermann-it.de
GnuPG: 0x5C3E6D75 | 5930 86EF 7965 2BBA 6572  C3D7 7B1D A3C3 5C3E 6D75