> Is it expected behavior for format to perform so many
> operations on all luns when the command is evoked
> with the "-d" switch?

I don't know if I'd expect that based on my first reading of the man page,
but as I look closer, I find it to be quite reasonable.  The man page says:
"If no disk-list is specified, format lists all the disks present in the system
that can be administered by format."

Given the usage in the synopsis at the top of the man page, mentioning
the disk to select with -d disk-name does not constitute supplying a
disk-list.  So the man page is consistent with the behavior you observe.

To only probe one disk, just list one disk on the command line (but not
as an argument to -d).  Thus, instead of

format -f /tmp/label -d cxtxdx

use

format -f /tmp/label cxtxdx

Since under the -d disk-name description, the man page says
"This can also be accomplished by specifying a single disk in the disk list",
that should work just fine.

Or if you like to type more than necessary (but there's no advantage to it!),

format -f /tmp/label -d cxtxdx cxtxdx

should be equivalent (if a few microseconds less efficient).

Of course, with the new in-use checking, it will still take a couple of
seconds, but at least not a couple of seconds per disk on the system.

There's a way to turn that off too, but it's undocumented as far as I know,
and thus could stop working with any patch or update.  That would be
something like

env NOINUSE_CHECK=true format -f /tmp/label cxtxdx

(Any value of the environment variable would work, the check is only if
NOINUSE_CHECK exists.)  That saves that last couple of seconds; like that,
it comes up instantly.

That's what you get when you read both the man page and the source code,
and try to match them up for the matter of concern. :-)
 
 
This message posted from opensolaris.org
_______________________________________________
storage-discuss mailing list
[EMAIL PROTECTED]
http://mail.opensolaris.org/mailman/listinfo/storage-discuss

Reply via email to