[653] in NetBSD-Development
Lossage under -current and libafs1_0A.o
daemon@ATHENA.MIT.EDU (ghudson@MIT.EDU)
Sun Mar 12 23:27:17 1995
From: ghudson@MIT.EDU
To: netbsd-afs@MIT.EDU
Date: Sun, 12 Mar 1995 23:27:05 EST
This problem has been observed on granola (which does not have my
updated libafs1_0A.o) and glacier (which does).
The symptoms I've observed are:
1. (twice) the machine, running X, stops responding or
pinging and has to be power-cycled. Once on glacier,
once on granola.
2. (once) granola hangs in console mode, and does not
respond to ctrl-alt-esc or any other keys. Yes, the
kernel has ddb.
3. (once) A "cvs update" process goes into disk-wait
along with a shell that either was running it or
was run by it. The process locations are:
cvs afs f88c5870
afsd afs f87e76ee
afsd afs f87e76ee
afsd ttyout f8759840
afsd afs f87e76ee
afsd f87e8d7c (yes, it looked like that)
afsd afs f878f230
The first time this happened, on glacier, I was running "make
install" in /usr/src/games. The second, third, and fourth times,
I was running "cvs update" on granola in /afs/sipb/user/ghudson/
pthreads. A ktrace of the cvs process (currently in /u1/build/
afs-crash/ktrace.out) has the last few lines:
263 cvs CALL open(0xf7bfa840,0,0x1b6)
263 cvs NAMI "/afs/sipb.mit.edu/project/pthreads/src/CVS/pthreads/pthreads/fs_sysv.c,v"
263 cvs RET open 3
263 cvs CALL fstat(0x3,0xf7bfa6e0)
263 cvs RET fstat 0
263 cvs CALL read(0x3,0x5c000,0x2000)
I left a copy of netbsd-afs in /u1/build/afs-crash/netbsd-afs.
