That is very bizarre... Its conceivable that that the later
	bootblocks either set something extra up, or clear some memory
	that the older ones leave potentially full of garbage.

	If all is still working after a while would it be possible to test
	installing the 1.4.1 bootblocks to see in which camp they lie?


		David/absolute

On Fri, 3 Dec 1999, Brian Buhrow wrote:

> 	OK.  I have another data point to toss at this problem, though  I admit
> I don't understand what's changed.  As I wrote previously, I've been having
> trouble running NetBSD 1.3.x, 1.4.x on this Sparc 5.  I'd get kernel data
> faults, alignment faults, etc.  1.3.x kernels seemed happier than 1.4.x
> kernels in that they might run for as much as two weeks before failing.
> 1.4.x kernels seemed to be even worse, failing after just a few hours or a
> day.
> 	In an effort to diagnose the problem, I decided to try upgrading to a 
> current kernel.  Using the snapshot of mid-October, I installed a new boot
> block so I could boot ELF kernels, and put on the 1.4l kernel.  Because I
> have 256MB of RAM in the machine, the 1.4l kernel from the snapshot
> wouldn't boot.  So, I rebooted the 1.4.1 kernel, using the newly installed
> boot block, figuring I'd solve the problem of how to get a new kernel to
> boot later.
> 	That was two days ago and the machine has been up and 
> busy that entire time.  While we're not completely out of the woods yet, it
> has a better feel about it already.
> 	What I'm trying to figure out is why changing the boot block
> might have such a dramatic effect on the health of the machine?  The old
> boot block was from NetBSD 1.3.1 or there abouts, and the new one is from
> the snapshot.  Any ideas?
> -Brian
>