On Sun, 11 Feb 2001, Paul Apprich wrote:

# > On my system trace from DDB gave good data - basically the system
# > is healthy and sitting in the idle loop, no interrupts masked, pingable on
# > the network etc.
#
# Seems that one of my machines is now doing this too.  Yesterday rdists
# to it would mysteriously hang and eventually get aborted by the
# sending side.  I could still log into that machine via ssh, but that
# rdist was stuck.
#
# Just to make sure that it wasn't an ffs corruption bug I went single
# user, mounted an NFS /usr and newfs-ed the local one.  Then I did a
# tar copy from the NFS-ed /usr to the clean local partition.  The tar
# also hung after a few 10's of megs.  The writing tar is waiting on
# "biowait" flags 0x4006 according to DDB.  (This is even with
# "genfs_vnops.c,v 1.29 2001/02/18")
#
# Does this ring a bell for anyone?

Yes.  Mine didn't fail on the install, but when I tried to do a pkg_add
of the perl binary set (no compiler on the box -- it's a router), it
would routinely hang in the middle of the extraction when gzip was involved
in the pipeline.  It seemed to get a bit farther if I used pax as opposed
to tar, but it still hung dead.

				--*greywolf;
--
"Hello, I've just upgraded my OS to Windows 2000..."
"Yes...?"
"Well, my computer doesn't work now."
"Yes, you just said that a second ago."