zfs discuss - May 2006 - Question about /etc/default/fs for a zfs based zone

I have a zone named foo on a snv_38 thus :

# zonecfg -z foo info
zonename: foo
zonepath: /opt/zone/foo
autoboot: true
pool:
limitpriv:
inherit-pkg-dir:
        dir: /lib
inherit-pkg-dir:
        dir: /platform
inherit-pkg-dir:
        dir: /sbin
inherit-pkg-dir:
        dir: /usr
net:
        address: 192.168.35.210
        physical: qfe3


The zone contents actually reside entirely on a ZFS filesystem :

# zfs list zfs0/zone/foo
NAME                   USED  AVAIL  REFER  MOUNTPOINT
zfs0/zone/foo          250M  52.0G   250M  /opt/zone/foo

# zfs get -o property,value,source all zfs0/zone/foo
PROPERTY       VALUE                      SOURCE
type           filesystem                 -
creation       Sat May 13  9:38 2006      -
used           250M                       -
available      52.0G                      -
referenced     250M                       -
compressratio  1.87x                      -
mounted        yes                        -
quota          none                       default
reservation    none                       default
recordsize     128K                       default
mountpoint     /opt/zone/foo              local
sharenfs       off                        default
checksum       on                         default
compression    on                         inherited from zfs0
atime          on                         default
devices        on                         default
exec           on                         default
setuid         on                         default
readonly       off                        default
zoned          off                        default
snapdir        hidden                     default
aclmode        groupmask                  default
aclinherit     secure                     default

I see the option "zoned" up there and am not too sure what to do with
that yet.

Inside the foo zone I see that /etc/default/fs says :

LOCAL=ufs


Shouldn''t this be LOCAL=zfs ?   Or will that cause problems ?

-- 
Dennis Clarke

{ responding to myself .. very tacky }

I see this section in the zfs manpage that sheds some light on the
"zoned"
property.  It does not say anything about /etc/default/fs however.

  Zones
     A ZFS file system can be added to a non-global zone by using
     zonecfg''s  "add  fs"  subcommand.  A ZFS file system
that is
     added to a non-global zone must have its mountpoint property
     set to legacy.

     The physical properties of an added  file  system  are  con-
     trolled  by  the  global  administrator.  However,  the zone
     administrator can create, modify, or  destroy  files  within
     the  added  file system, depending on how the file system is
     mounted.

     A dataset can also be delegated  to  a  non-global  zone  by
     using   zonecfg''s   "add  dataset"  subcommand.  You 
cannot
     delegate a dataset to one zone and the children of the  same
     dataset  to  another zone. The zone administrator can change
     properties of the dataset or any of its  children.  However,
     the "quota" property is controlled by global administrator.

     A ZFS emulated volume can be added as a  device  to  a  non-
     global zone by using zonecfg''s "add device" subcommand.
How-
     ever, its physical properties can only be  modified  by  the
     global administrator.

     For more information about zonecfg syntax, see zonecfg(1M).

     Once a dataset  is  delegated  to  a  non-global  zone,  the
     "zoned"  property  is automatically set. A zoned file system
     cannot be  mounted  in  the  global  zone,  since  the  zone
     administrator might have set the mount point to an unaccept-
     able value.

     The global administrator can forcibly clear the "zoned" pro-
     perty,  though  this  should  be done with extreme care. The
     global administrator should verify that all the mount points
     are acceptable before clearing the property.



-- 
Dennis Clarke

On Sat, May 13, 2006 at 03:59:00PM -0400, Dennis Clarke
wrote:
> 
> Inside the foo zone I see that /etc/default/fs says :
> 
> LOCAL=ufs
> 
> 
> Shouldn''t this be LOCAL=zfs ?   Or will that cause problems ?
All this determines is the default filesystem type for mount(1M), fsck(1M),
etc.  Given that you use zfs(1M) for all that kind of manipulation,
it seems like this is not a huge deal.

Cheers,
- jonathan

-- 
Jonathan Adams, Solaris Kernel Development

On Mon, May 15, 2006 at 10:13:51AM -0700, Jonathan Adams
wrote:
> 
> All this determines is the default filesystem type for mount(1M), fsck(1M),
> etc.  Given that you use zfs(1M) for all that kind of manipulation,
> it seems like this is not a huge deal.
> 
Note that if you are a fan of legacy mounts, there''s an open RFE to
have
mount(1M) identify a ZFS pool/dataset path without ''-F'',
rather than
trying to interpret it as a NFS mount:

6365048 legacy mount should recognise ZFS filesystems without needing -F flag

- Eric

--
Eric Schrock, Solaris Kernel Development       http://blogs.sun.com/eschrock