server process was terminated by signal 11 (easy to duplicate) - BIS

Hello,

I'm experiencing the same problem as Philip Walker and others, on two
different computers (Pentium II and III), both Windows-updated and
CygWin-fromscratch-setup-reinstalled yesterday (i.e. XP-Pro SP1 and Cygwin
1.5.9-1, Cygipc 2.03-2 and PostgreSQL 7.4.1-3; see below 'cygcheck truncated
output').

Whenever I launch a task which seems to need more memory, the server process
gets terminated by signal 11 (i.e. based on my experience : when creating
many tables at ones within a transaction or vacuuming/analyzing a large
database ; I haven't been able to go any further, for my project-database is
unusable on Cygwin while behaving just fine on our Linux/RedHat-based
production machines).

I tried :
kill ALL unnecessary process on my machine (antivirus, vpn server, etc.),
keeping just the bare minimum - no difference
lower shared buffers from 1000 to 20 - no difference
lower/extend vacuum mem from 1024 to 16384 - no difference
extend wal_buffers - no difference
turn fsync on and off - no difference
launch postmaster with -d 5 debugging-level - no further information (DEBUG
: PorcessUtility ; DEBUG: reaping dead process ; LOG : server process (PID
xxxx) was terminated by signal 11)
launch ipc-daemon2 with -d debugging-level - no output

I was able to go further in my project by :
suppressing the transaction blocks - suceeded to create the entire database
schema
vacuum NOT analyzing the database - suceeded to vacuum the database

But...
still not able to vacuum-analyze database
and I supect that as soon as I will launch any other memory-greedy
operation, result will be the same

I also tested :
create a different schema (i.e. stripping the unrequired parts until the
.sql script which crashed my original schema : the tables creation script,
originally transaction-wrapped) - no difference

Other symptoms/side-effects:
when crashing because of the "too"-large transaction blocks, the database
would get corrupted and would prevent server to reinitialize until a new
'initdb' is done (LOG: Exception: STATUS_ACCESS_VIOLATION) ; reinitialize ok
if crash occurs because of vacuum

I will try a 'strace' but I doubt my ability to interpret the results... and
much more fix the problem. ;-)

WHAT ELSE CAN I DO ?

Please help,

Cedric

POSTGRESQL LOG (exerpt)
=====================

[....]
NOTICE: CREATE TABLE / UNIQUE will create implicit index "uq_zone_loc" for
table "tb_zone_loc"
NOTICE: CREATE TABLE / PRIMARY KEY will create implicit index
"pk_zone_loc_note" for table "tb_zone_loc_note"
NOTICE: CREATE TABLE / PRIMARY KEY will create implicit index
"pk_zone_jrnl" for table "tb_zone_jrnl"
NOTICE: CREATE TABLE / PRIMARY KEY will create implicit index
"pk_photographer" for table "tb_photographer"
NOTICE: CREATE TABLE / UNIQUE will create implicit index "uq_photographer"
for table "tb_photographer"
LOG: server process (PID 3284) was terminated by signal 11
LOG: terminating any other active server processes
[...]

CYGCHECK (truncated)
==================

Cygwin Win95/NT Configuration Diagnostics
Current System Time: Sat Apr 24 08:07:03 2004

Windows XP Professional Ver 5.1 Build 2600 Service Pack 1

Cygwin DLL version info:
DLL version: 1.5.9
DLL epoch: 19
DLL bad signal mask: 19005
DLL old termios: 5
DLL malloc env: 28
API major: 0
API minor: 112
Shared data: 4
DLL identifier: cygwin1
Mount registry: 2
Cygnus registry name: Cygnus Solutions
Cygwin registry name: Cygwin
Program options name: Program Options
Cygwin mount registry name: mounts v2
Cygdrive flags: cygdrive flags
Cygdrive prefix: cygdrive prefix
Cygdrive default prefix:
Build date: Thu Mar 18 23:05:18 EST 2004
Shared id: cygwin1S4

Cygwin Package Information
Package Version
_update-info-dir 00227-1
cygipc 2.03-2
cygrunsrv 1.0-1
cygwin 1.5.9-1
postgresql 7.4.1-3