Subject: Re: RFC: progressbars for fsck_ffs
To: None <tech-userlevel@netbsd.org>
From: Greg A. Woods <woods@weird.com>
List: tech-userlevel
Date: 11/18/2002 13:29:54
[ On Sunday, November 17, 2002 at 23:26:37 (+0000), David Laight wrote: ]
> Subject: Re: RFC: progressbars for fsck_ffs
>
> > but I like the idea I know something was happening rather
> > than nothing being displayed for extended periods.
> 
> If the fsck is being done during system boot, it would actually
> be useful just to output a line indicating that the fsck is in
> progress - not just one when it has finished.
> At least you'd know why there was a pregnant pause.

This is what (and all that) I've wanted to do for a very long time as
well.  I don't need to see any indication of progress -- just an
indication of when each individual filesystem check starts, and of
course the results summary when it completes.

On the other hand I've more or less also thought that simply adding a
suitable prefix to the lines displayed by plain "fsck" (i.e. without the
'-p'), would also suffice.  I always though '-p' stood for "parallel".  :-)

I don't mind seeing the notification of each phase, just so long as each
line of output identifies which filesystem it belongs to (I take lots of
advantage of parallel checks!).

I've not yet got around to actually imlementing this, though I've now
looked at the code, at least in fsck_ffs, several times with this intent
in mind.

Yes, the ability to type a VSTATUS character (and thus send SIGINFO to
the running fsck* processes) is useful, but it requires user interaction
whereas a message printed at the start of the operation is both
immediately visible in a glance and is also persistent in the system
message buffer and thus available to be logged and recorded for
posterity too.

-- 
								Greg A. Woods

+1 416 218-0098;            <g.a.woods@ieee.org>;           <woods@robohack.ca>
Planix, Inc. <woods@planix.com>; VE3TCP; Secrets of the Weird <woods@weird.com>