3.2C crash caused by NFS

From: Jason Van Patten <jvp_at_visix.com>
Date: Mon, 11 Mar 1996 10:11:00 -0500

Hi folks -

        I'm in some desperate need of help here with my 3.2C box. Recently,
I've added some memory to it, bumping it up to 80M. (Alpha AXP 150 PC)
After that, I started doing some software builds, and my machine has crashed
twice. I want to avoid blaming the new memory just yet, because I _have_
successfully built the software in question quite a few times, with no
problem (and this build will definitely stress-test the physical RAM.)

        Anyhow, a lot of NFS reads are performed when I do a build, and I
_think_ that's what's causing the problem (just a guess.) In the crash-data
file, I find these messages:


_dump_begin:
> 0 boot(0x0, 0x0, 0xfffffc000052e4f0, 0x0, 0xfffffc0000316ea0) ["../../../../s
rc/kernel/arch/alpha/machdep.c":1735, 0xfffffc000044139c]

   1 panic(s = 0xfffffc000052e4f0 = "ubc_invalidate: NULL object")
["../../../..
/src/kernel/bsd/subr_prf.c":757, 0xfffffc0000400954]
pcpu = (nil)
i = -1981416088
bootopt = 13856256
mycpu = 20055047
spl = 0
prevcc = 18446739675695000392
nextcc = 18446739675668823848
timer = 4
limit = 1280

   2 ubc_invalidate(0xfffffc0004722000, 0x0, 0x0, 0x0, 0xfffffc00002b6d40)
["../
../../../src/kernel/vfs/vfs_ubc.c":2674, 0xfffffc000029ebc0]

   3 nfs_purge_caches(0xfffffc0004722000, 0xfffffc00047220b0,
0xfffffc00002a48cc
, 0x0, 0x1002a45d8) ["../../../../src/kernel/nfs/nfs_client.c":176,
0xfffffc0000
2a1fc8]

   4 nfs_cache_check(0xfffffc0004722000, 0x3141202f, 0x3141202f,
0xfffffc00035e8
b80, 0x10290) ["../../../../src/kernel/nfs/nfs_client.c":219,
0xfffffc00002a20d4
]

   5 makenfsnode(fh = 0x30, attr = 0xfffffc0003fe7224, mp = 0xfffffc00048aa800)
["../../../../src/kernel/nfs/nfs_subr.c":1325, 0xfffffc00002a52b4]
rp = 0xfffffc00047220b0
vp = 0xfffffc0004722000
newnode = '^_at_'
nfh = struct {
    fh_len = 32
    fh_buf = ",M-^P*!^F^CM-^OM-QM-WOM-<"
}
trp = 0x3

   6 nfs_create(0xfffffc0000000007, 0xffffffff000001b4, 0xffffffff000000fa,
0xff
ffffffffffffff, 0xffffffffffffffff)
["../../../../src/kernel/nfs/nfs_vnodeops.c"
:1847, 0xfffffc00002ab610]

   7 vn_open(0xffffffff89e5f720, 0x1602, 0x1b4, 0x4fd7, 0xfffffc000044f244)
["..
/../../../src/kernel/vfs/vfs_vnops.c":490, 0xfffffc0000298350]

   8 copen(0xfffffc00035e8210, 0xffffffff89e5f8c8, 0xffffffff89e5f8b8, 0x0,
0xff
fffc000044e608) ["../../../../src/kernel/vfs/vfs_syscalls.c":1824,
0xfffffc00004
16538]

   9 open(0xffffffff89e5f8b8, 0x0, 0xfffffc000044e608, 0xfffffc000044ec74,
0xfff
ffc000044e078) ["../../../../src/kernel/vfs/vfs_syscalls.c":1781,
0xfffffc000041
645c]

  10 syscall(0x107b0, 0x1, 0x1, 0x0, 0x2d)
["../../../../src/kernel/arch/alpha/s
yscall_trap.c":519, 0xfffffc000044e074]

  11 _Xsyscall(0x8, 0x3ff800e7a18, 0x3ffc0091560, 0x107b0, 0x601)
["../../../../
src/kernel/arch/alpha/locore.s":1094, 0xfffffc000043e294]

_dump_end:


        I'm sorry for all the spam, but I figured it might help diagnose the
problem. The machines that I've mounted drives off of are mostly Sparcs
running SunOS 4.1.4. One crash happened while mounting a Linux drive as
well.

        It doesn't happen all the time, only on certain occassions (planets
have to be aligned just right, etc.) But it is a big problem nonetheless.
Does anyone out there have any ideas what _might_ be causing it, and whether
or not there's a way to fix it?

        Thanks.

                                                Jason
-- 
Jason Van Patten                  | Yes it is 12" long, but I don't use |
Visix Software, Inc.   Reston, VA | it as a rule.                       |
jvp_at_visix.com                     |                     - Benny Hill    |
      Opinions expressed here are not those of Visix Software.
Received on Mon Mar 11 1996 - 16:59:33 NZDT

This archive was generated by hypermail 2.4.0 : Wed Nov 08 2023 - 11:53:46 NZDT