This message is to explain why the ftp and sup server has been less
than 100% reliable in the last few days, and what we're doing about
it.

While we were readying netbsd.warped.com to take over sup and ftp,
sun-lamp lost a disk which forced us to rush the schedule.  There
were a number of mistakes made during the transition which caused
major problems in the end.  For instance:

      * We checked out a new set of files from pain for sup service,
	which caused sup to think everything was out of date and to
	download every file.

      * We did not have the "private" sup port set up until yesterday.

      * We set up the update to regenerate tar files every day instead
        of every week.

      * We enabled too many sup/ftp (~32) daemons, thinking that the
	larger number would clear out the backlog,  without realizing
	that the machine had only 40MB swap.  (And we all know NetBSD's
	mode of failure when swap fills...)

      * We had installed a P24T processor, but failed to realize that
        that the manual had an incorrect jumper switch diagram, which
	caused erratic behavior.

We have been working to address these problems over the last week.  At
one time, when we suspected the disc controller, I ran out and bought
a new one.

And just when we had everything stable, and all of you folks were
beating on it get your updated sources.  It's 9GB Micropolis hard
drive gave up the ghost.  

We've replaced the 9GB drive with a 2GB unit until it can be replaced,
and hope to have both sup and ftp running as soon as we can restore
the lost data.  We've added more swap space in the process, so we
should be able to handle an greatly increased load when we come back
on line.

I'd like to extend my thanks to all of the folks who have stepped up
to the call of duty to get our ftp and sup servers up and running.
I'd like to also thank all of the users who have been patient during
this frustrating transition.

	--jtc

--
J.T. Conklin
The NetBSD Project