Hello !
I'm about to install O9i on FreeBSD box.
uname -a:
FreeBSD stat2.scn.ru 5.2.1-RELEASE-p3 FreeBSD 5.2.1-RELEASE-p3 #2:
Fri Apr 23 19:19:43 KRAST 2004
rd@stat2.scn.ru:/usr/src/sys/i386/compile/RDSTAT2-ORACLE i386
uname -a under oracle user [linux shell]:
Linux stat2.scn.ru 2.4.2 FreeBSD 5.2.1-RELEASE-p3 #2: Fri Apr 23
19:19:43 KRAST 2004 i686 unknown
My problem is when I run "startup nomount" command, I get
$ sqlplus "sys/admin as sysdba"
Connected to an idle instance.
SQLstartup nomount
ORA-00445: background process "PMON" did not start after 120 seconds
and I found out that 'oracle' process dies every time I enter "startup
nomount":
bash-2.05b# tail -1 /var/log/messages
Apr 23 21:09:18 stat2 kernel: pid 824 (oracle), uid 71: exited on
signal 11
I spend many days to make it work without results.
I found post at http://www.shadowcom.net/freebsd-oracle9i/
and make sure that I did all step-by-step:
1. kernel options applied
2. Installed all required ports, pkg_info |grep linux
linux-sun-jdk-1.3.1.10
linux_base-7.1_7
linux_devtools-7.1_3
3. created mtab
4. created oracle account, groups also was created
etc..
Also I view http://tomclegg.net/oracle9i-bsd5 ,
http://www.bsdforums.org/forums/show...?threadid=7441
recomendations, but this contain nothing about this problem.
This result I get on Oracle 9.2.0.1.0 and 9.2.0.4.0 (last patchset
applied),
the result is the same, "background process "PMON" did not start after
120 seconds",
"(oracle), uid 71: exited on signal 11".
What I did make incorrect? I tried to play with init-parameters, but
this did not help - somewhere in forum ppls said this can occur when
SGA size exceed physical size (server have 512M of RAM),
but my SGA size small. Last my version of initCLDB.ora was:
db_name=CLDB
instance_name=CLDB
#db_files = 400
#db_block_buffers = 3200
#shared_pool_size = 9000000
#processes = 100
#log_buffer = 32768
db_file_multiblock_read_count = 16
log_checkpoint_interval = 10000
parallel_max_servers = 5
# audit_trail = true # if you want auditing
timed_statistics = true # if you want timed statistics
#max_dump_file_size = 10240 # limit trace file size to 5 Meg
each
# log_archive_dest = disk$rdbms:[oracle.archive]
# log_archive_format = "T%TS%S.ARC"
# transactions = 40
# transactions_per_rollback_segment = 5
# Global Naming -- enforce that a dblink has same name as the db it
connects to
#my## global_names = TRUE
# db_domain = us.acme.com # global database name is
db_name.db_domain
# define two control files by default
control_files = ('/opt/oracle/oradata/u01/ctrl01.ctl',
'/opt/oracle/oradata/u02/ctrl02.ctl')
compatible = 9.2.0.4.0
background_dump_dest = /usr/opt/oracle/admin/CLDB/bdump
user_dump_dest = /usr/opt/oracle/admin/CLDB/udump
core_dump_dest = /usr/opt/oracle/admin/CLDB/cdump
os_authent_prefix = ""
oracle_trace_enable=TRUE
many parameters have default values...
I have properly configured listener, tnsping works. After connecting
to
an idle instance (before enter 'startup nomount') I have some
oracle process:
$ ps ax | grep ora
895 ?? Ss 0:00,03 oracleCLDB
(DESCRIPTION=(LOCAL=YES)(ADDRESS=(PROTOCOL=beq))) (oracl
884 p1 I 0:00,05 /usr/opt/oracle/product/9.2.0/bin/tnslsnr
LISTENER -inherit
I even tries to make ktrace of oracle, here is last lines of this dump
(forked process, I think it is PMON):
6877 oracle RET getpid 6877/0x1add
6877 oracle CALL pathconf(0x3,0xbfbfb81c)
6877 oracle RET pathconf 0
6877 oracle CALL pwrite(0x20,0xbfbfb77c,0,0x8)
6877 oracle RET pwrite 0
6877 oracle CALL pwrite(0x21,0xbfbfb77c,0,0x8)
6877 oracle RET pwrite 0
6877 oracle CALL pwrite(0x22,0xbfbfb77c,0,0x8)
6877 oracle RET pwrite 0
6877 oracle CALL obs_execv(0xbfbfba84,0xbfbfbca8,0xbfbfe7a4)
6877 oracle NAMI
"/compat/linux/usr/opt/oracle/product/9.2.0/bin/oracle"
6877 oracle NAMI "/usr/opt/oracle/product/9.2.0/bin/oracle"
6877 oracle NAMI "/compat/linux/lib/ld-linux.so.2"
6877 oracle RET obs_execv 0
6877 oracle CALL settimeofday(0xbfbfe2a4)
6877 oracle RET settimeofday 0
6877 oracle CALL ktrace(0)
6877 oracle RET ktrace 181866496/0xad71000
6877 oracle CALL semget(0,0x1,0)
6877 oracle RET semget -1 errno -9 Unknown error: -9
6877 oracle CALL open(0x4a4cd30c,0x20000,0xbfbfe26c)
6877 oracle NAMI "/compat/linux/dev/null"
6877 oracle NAMI "/dev/null"
6877 oracle RET open 0
6877 oracle CALL mmap(0,0xbfbfe204,0)
6877 oracle RET mmap 0
6877 oracle PSIG SIGSEGV SIG_DFL
Is it since semget-around problem? Or what is obs_execv?
Please help to point me to where my mistake is?
Thanks a lot.
--
best regards,
Ruslan A Dautkhanov ru*****@scn.ru
