An old W2K installation now fails with core dump after postmaster.
A new WXP installation fails in the same way.
Cygserver is running as a service and is started
Postmaster core dumps when run.
Running postmaster with debug gives:
bash-3.00$ postmaster -d 5
DEBUG: postmaster: PostmasterMain: initial environ dump:
DEBUG: -----------------------------------------
DEBUG: HOMEPATH=\Documents and Settings\adb
DEBUG: TERM=cygwin
DEBUG: PROCESSOR_IDENTIFIER=x86 Family 15 Model 47 Stepping 2,
AuthenticAMD
DEBUG: WINDIR=C:\WINDOWS
DEBUG: USERDOMAIN=ACONCAGUA
DEBUG: OS=Windows_NT
DEBUG: !::=::\
DEBUG: COMMONPROGRAMFILES=C:\Program Files\Common Files
DEBUG: PROCESSOR_LEVEL=15
DEBUG:
PATH=/usr/bin:/cygdrive/c/WINDOWS/system32:/cygdrive/c/WINDOWS:
/cygdrive/c/WINDOWS/System32/Wbem:
/cygdrive/c/Program Files/ATI Technologies/ATI Control Panel
DEBUG: FP_NO_HOST_CHECK=NO
DEBUG: PWD=/usr/bin
DEBUG: SYSTEMDRIVE=C:
DEBUG: CYGWIN=server ntsec tty
DEBUG: USERPROFILE=C:\Documents and Settings\adb
DEBUG: CLIENTNAME=Console
DEBUG: LOGONSERVER=\\ACONCAGUA
DEBUG: PROCESSOR_ARCHITECTURE=x86
DEBUG: !C:=C:\Documents and Settings\adb
DEBUG: SHLVL=1
DEBUG: HOME=/home/adb
DEBUG: PATHEXT=.COM;.EXE;.BAT;.CMD;.VBS;.VBE;.JS;.JSE;.WSF;.WSH;.sh
DEBUG: HOMEDRIVE=C:
DEBUG: PROMPT=$P$G
DEBUG: COMSPEC=C:\WINDOWS\system32\cmd.exe
DEBUG: TMP=/cygdrive/c/DOCUME~1/adb/LOCALS~1/Temp
DEBUG: SYSTEMROOT=C:\WINDOWS
DEBUG: LOGNAME=adb
DEBUG: PROCESSOR_REVISION=2f02
DEBUG: PGDATA=/var/postgresql/data
DEBUG: PROGRAMFILES=C:\Program Files
DEBUG: NUMBER_OF_PROCESSORS=1
DEBUG: SESSIONNAME=Console
DEBUG: COMPUTERNAME=ACONCAGUA
DEBUG: _=/usr/bin/postmaster
DEBUG: TZ=FLEST-2FLEST-3,M3.5.0/3,M10.5.0/4
<a few other environ vars snipped>
DEBUG: -----------------------------------------
DEBUG: searching PATH for executable
DEBUG: found "/usr/bin/postgres" using PATH
DEBUG: invoking IpcMemoryCreate(size=9469952)
Segmentation fault (core dumped)
The sygserver log shows plenty of semctl ops, but no obvious errors.
I've used the default cygserver.conf, but have since played around with the
settings. By setting all the kern shm and sem settings to the max, the
cygserver fails to start and then becomes unstoppable (as reported by
someone else).
Any ideas?
Antony
