[comp.unix.xenix] Mount warning on SCO XENIX

meyer@mimsy.UUCP (John R. Meyer) (01/11/88)

Hello --

	I am using SCO XENIX 2.2 to mount some filesystems on
empty directories.  However, when I try to mount on directories
that are not one level down from root (that is, /usr/src instead
of /src), I get the message

	mount: WARNING!! - mounting: <src> as <usr/s>

The mount works fine, but the message is annoying.  Does anyone
know why the message is issued?  Is there some reason I should
not mount in a subdirectory?  Can I get rid of this message
on boot-up somehow?


				Thanks,

				John
				meyer@mimsy.umd.edu
-- 
John R. Meyer					Domain: meyer@mimsy.umd.edu
10208C Ashbrooke Ct.				Path:   uunet.uu.net!mimsy!meyer
Oakton, VA  22124				Phone:  (703) 644-3944 (O)
Disclaimer:  The views expressed are my own.		(703) 281-5157 (H)

stephm@sco.COM (Stephen P. Marr) (01/28/88)

In article <284@sysco> chapman (Brian Chapman Mx321) writes:
>In article <10120@mimsy.UUCP> meyer@mimsy.UUCP (John R. Meyer) writes:
>< Hello --
>< 	I am using SCO XENIX 2.2 to mount some filesystems on
>< empty directories.  However, when I try to mount on directories
>< that are not one level down from root (that is, /usr/src instead
>< of /src), I get the message
>< 
>< 	mount: WARNING!! - mounting: <src> as <usr/s>
>< 
>The super user can re-assign the name field
>of a file system with fsname(M).  (Please read the manual
>before invoking new utilities at your filesystems).
>
>"mkdev fs" put the original name there.  I guess it
>put "src" instead of "usr/src" there.


What is so is that the space allowed for the filesystem name is
six bytes (including the null).  mkdev fs uses the device name
to presume the name of the filesystem, in this case, there was
a /dev/src, and mkdev fs placed src in the name slot, mount
expects that the name of the filesystem will match the directory
(stripping the leading /) where it is mounted.  In order to
properly use this feature, the pathname of the mount point must
not exceed 5 chars excluding the leading /.  To annihilate the
thing altogether, you can change the name of an unmounted filesystem
by incanting fsname -s "" /dev/block_filesystem_device  e.g
fsname -s "" /dev/src

Have at it.
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Stephen Marr        Computer Services Supervisor       The Santa Cruz Operation
Internet: stephm@sco.COM         USENET: ...!{ihnp4,uunet,amd,ucscc}!sco!stephm
-=- The opinions expressed do not relate to reality and don't belong to SCO -=-
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

domo@riddle.UUCP (Dominic Dunlop) (02/03/88)

In article <164@scovert> stephm@sco.COM (Stephen P. Marr) writes:
>In article <284@sysco> chapman (Brian Chapman) writes:
>>In article <10120@mimsy.UUCP> meyer@mimsy.UUCP (John R. Meyer) writes:
>>< 	I am using SCO XENIX 2.2 to mount some filesystems on
>>< empty directories.  However, when I try to mount on directories
>>< that are not one level down from root (that is, /usr/src instead
>>< of /src), I get the message
>>< 
>>< 	mount: WARNING!! - mounting: <src> as <usr/s>
>>< 
>>The super user can re-assign the name field
>>of a file system with fsname(M).  (Please read the manual
>>before invoking new utilities at your filesystems).

The corresponding System V command is  labelit.  In addition to the six-
character name, it can set a six-character ``volume number''.  This has
some significance to the dcopy program, I seem to recall.  I tend to use it
for the date I built the filesystem (eg 880202).

>>
>>"mkdev fs" put the original name there.  I guess it
>>put "src" instead of "usr/src" there.
>
>What is so is that the space allowed for the filesystem name is
>six bytes (including the null).  mkdev fs uses the device name
>to presume the name of the filesystem...  In order to
>properly use this feature, the pathname of the mount point must
>not exceed 5 chars excluding the leading /.  To annihilate the
>thing altogether, you can change the name of an unmounted filesystem
>by incanting fsname -s "" /dev/block_filesystem_device  e.g
>fsname -s "" /dev/src

Yes.  The six character limit is DUMB.  System V, release 3 mount makes
life a bit more livable by checking the basename of the mount point
against the filesystem name on the disk pack/partition, and not
squawking provided that the first six characters match.  This means
that I can now mount a partition called news on /usr/spool/news without
complaint.  However, I also get no complaint if I mount a partition
called invent on /usr/inventions when I should have been mounting it on
/usr/inventory...

I guess that the merging of XENIX and V.3 will mean that XENIX mount
will soon inherit this still imperfect but less anti-social behaviour
from AT&T.
-- 
Dominic Dunlop
domo@sphinx.co.uk  domo@riddle.uucp

) (02/08/88)

From article <284@sysco>, by chapman@sco.COM (Brian Chapman Mx321):
} In article <10120@mimsy.UUCP> meyer@mimsy.UUCP (John R. Meyer) writes:
} < Hello --
} < 	I am using SCO XENIX 2.2 to mount some filesystems on
} < empty directories.  However, when I try to mount on directories
} < that are not one level down from root (that is, /usr/src instead
} < of /src), I get the message
} < 
} < 	mount: WARNING!! - mounting: <src> as <usr/s>
} < 
} < The mount works fine, but the message is annoying.  Does anyone
} 

I have a similar problem but the message is more annoying, mine says:

	mount: WARNING!! - mounting: <usr} as <usr>

Its been doing this ever since I moved /usr off the root file system.
The error message makes no sense to me.

} The super user can re-assign the name field
} of a file system with fsname(M).  (Please read the manual
} before invoking new utilities at your filesystems).

Yes, I tried this and fsname(M) does change the the name field, but I
still get the error message when I mount /dev/usr /usr.

-- 
Dennis Roth @ Home 8817 Sterling Street, Ardmore Village, Maryland, 20785
{decuac, grebyn, netsys}!macom1!coldbeer!roth
Unix is the OS of the future, and always will be.  Those that cannot 
program in a segmented memory enviroment, are damned to PL/I.