[19686] in Athena Bugs
Knfs server dying on Athena 9.0.14
daemon@ATHENA.MIT.EDU (Tom Fitzgerald)
Wed Aug 22 12:45:37 2001
Message-Id: <200108221645.MAA28957@sligo.mit.edu>
To: bugs@MIT.EDU
cc: tfitz@MIT.EDU, miki@MIT.EDU
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Date: Wed, 22 Aug 2001 12:45:34 -0400
From: Tom Fitzgerald <tfitz@MIT.EDU>
Hi. After upgrading some knfs servers to Athena 9.0.14, the mountd
now dies at random. It seems to die most often on the first NFS mount
after startup - if it survives that, it seems to be fine for days, but
I believe I've seen some cases of mountd dying later as well.
I've tested this on at least 2 servers (CAT2.MIT.EDU and ARCHFILE.MIT.EDU)
and am pretty sure it has happened on at least one more. All of these
systems are Ultra 5.
Truss on the mountd ends with:
14973: poll(0x000CC4F0, 9, -1) (sleeping...)
(mount request received here)
14973: poll(0x000CC4F0, 9, -1) = 1
14973: brk(0x000EC958) = 0
14973: brk(0x000FA958) = 0
14973: lwp_sema_post(0xFED09E30) = 0
14973: lwp_sema_wait(0xFED09E30) = 0
14973: getmsg(6, 0xFEC07C00, 0x000E3FC8, 0xFEC07C2C) = 0
14973: uname(0xFEC07130) = 1
14973: uname(0xFEC070D0) = 1
14973: getuid() = 0 [0]
14973: stat("/etc/krb5.conf", 0xFEC06518) = 0
14973: open("/etc/krb5.conf", O_RDONLY) = 16
14973: access("/etc/krb5.conf", 2) = 0
14973: fstat64(16, 0xFEC06330) = 0
14973: ioctl(16, TCGETA, 0xFEC062BC) Err#25 ENOTTY
14973: read(16, " [ l i b d e f a u l t s".., 8192) = 3508
14973: read(16, 0x000DEEA4, 8192) = 0
14973: llseek(16, 0, SEEK_CUR) = 3508
14973: close(16) = 0
14973: stat("/etc/athena/krb5.conf", 0xFEC06518) Err#2 ENOENT
14973: time() = 998497916
14973: stat("/etc/krb5.conf", 0xFEC06588) = 0
14973: stat("/etc/athena/krb5.conf", 0xFEC06588) Err#2 ENOENT
14973: Incurred fault #6, FLTBOUNDS %pc = 0x0002CCA4
14973: siginfo: SIGSEGV SEGV_MAPERR addr=0x00000014
14973: Received signal #11, SIGSEGV [caught]
14973: siginfo: SIGSEGV SEGV_MAPERR addr=0x00000014
14973: sigprocmask(SIG_SETMASK, 0xFF26F010, 0x00000000) = 0
14973: sigaction(SIGSEGV, 0xFEC06138, 0x00000000) = 0
14973: sigprocmask(SIG_SETMASK, 0xFF27ADE0, 0x00000000) = 0
14973: setcontext(0xFEC05FF8)
14973: Incurred fault #6, FLTBOUNDS %pc = 0x0002CCA4
14973: siginfo: SIGSEGV SEGV_MAPERR addr=0x00000014
14973: Received signal #11, SIGSEGV [default]
14973: siginfo: SIGSEGV SEGV_MAPERR addr=0x00000014
14973: *** process killed ***
