pageout and boot madness in 4.0.1?

[rws at expo.lcs.mit.edu]

[I don't read this list, so I would appreciate manual dexterity in routing
replies back to me.  Thanks.]

I have a diskless 3/60 with 16Mb of memory just recently upgraded to
4.0.1.  I've noticed an annoying tendancy for the system to wedge.
'vmstat 5' shows a pageout rate of between 1000 and 5000, although all
user processes are essentially idle, and none are especially large (an X
server at 1-3Mb is usually the largest).  Have others seen this?

Also, my 3/60 manages to boot about 1 in 20 attempts.  I almost always get
a number of "le: missed packet" messages, and usually end up with 

    whoami RPC call failed with status 10
    panic: vfs_mountroot: cannot mount root

or at times even a failure to download vmunix.  Moreover, the download
time (for both boot and vmunix) often seems like an order of magnitude
slower than it was under 3.4.  Has anyone else experienced this problem,
or should I start blaming hardware?

Finally, at first login after a boot (but not thereafter), I always seem to get
something like:

NFS write error 13 on host expo.lcs.mit.edu fh 317 1 80000 10016bb3 f7dd0000
80000 251a1 fdbf0000

for some NFS-mounted partition.  Can someone point me at the documentation to
parse this error message?