Bug 643368

Also seeing this problem on a fresh RHEL5.5 system and tried a few things (interactive vs answer file, update the system, disable firewall, change SELinux to permissive) without any luck.

This is the error from the install:
[root@dali cdrom]# ./install.pl --answer-file=/tmp/answers.txt  
* Starting the Red Hat Network Satellite installer.
* Loading answer file: /tmp/answers.txt.
* Performing pre-install checks.
* Pre-install checks complete.  Beginning installation.
* RHN Registration.
** Registration: System is already registered with RHN.  Not re-registering.
* Checking for uninstalled prerequisites.
* Applying updates.
* Installing RHN packages.
* Now running spacewalk-setup.
* Loading answer file: /tmp/answers.txt.
* Setting up Oracle environment.
* Setting up database.
** Database: Installing the database:
** Database: This is a long process that is logged in:
** Database:   /var/log/rhn/install_db.log
*** Progress: #########################################
Could not install database.
[root@dali cdrom]# 







This is the end of /var/log/rhn/install_db.log

SQL> SQL> SQL> 
SYSDATE   MESSAGE
--------- --------------------
15-NOV-10 Recompiling packages

1 row selected.

SQL> BEGIN utl_recomp.recomp_serial(); END;

*
ERROR at line 1:
ORA-04031: unable to allocate 2872 bytes of shared memory ("shared
pool","UTL_RECOMP_SEQSYS","sga heap(1,0)","call")
ORA-06512: at "SYS.UTL_RECOMP", line 616
ORA-06512: at "SYS.UTL_RECOMP", line 658
ORA-06512: at line 1


Disconnected from Oracle Database 10g Enterprise Edition Release 10.2.0.4.0 - 64bit Production
With the Partitioning, OLAP, Data Mining and Real Application Testing options

Hi Steve, 
please attach output of
# free
# grep --text -e size -e target /opt/apps/oracle/config/10.2.0/spfilerhnsat.ora
from the satellite server.

Thanks,
Michael

[root@dali cdrom]# free
             total       used       free     shared    buffers     cached
Mem:     132072280    8506492  123565788          0      94192    7862276
-/+ buffers/cache:     550024  131522256
Swap:     20971512          0   20971512
[root@dali cdrom]# grep --text -e size -e target /opt/apps/oracle/config/10.2.0/spfilerhnsat.ora
*.db_block_size=8192
*.max_dump_file_size='5M'# satellite adjusted value
*.pga_aggregate_target=134217728# satellite adjusted value
*.sga_target=268435456# satellite adjusted value
[root@dali cdrom]#

*** Bug 656212 has been marked as a duplicate of this bug. ***

Fixed in spacewalk git by
commit 3cf0ddc8fea44a0ece39df243b6339d884079664
    added correct units to memory params
commit a9079e153b0668f4b539d7cb0e668584c53360e2
    643368 - size oracle sga dynamicaly
commit 3d80d9f859984ed2c97252c886a432834ef6ae49
    643368 - compute sga size dynamicaly

Hit bug when running on 64cores, 256g of ram and installing SAT5.4 on RHEL5.5.  Got around it with - maxcpus=2, finished install, ran script:

--------------------------------------------

#!/bin/sh

# system memory and cpus
cpus=$(grep ^processor /proc/cpuinfo | wc -l)
total=$(free -m -o | awk '/Mem:/ {print $2}')

# all sizes are in MB
sgamin=256
sgamin_for_cpus=$((10*$cpus))
if [ $sgamin -lt $sgamin_for_cpus ] ; then
        sgamin=$sgamin_for_cpus
fi
sgamax=2048
if [ "$(uname -i)" = "i386" ] ; then
        sgamax=1500
fi
pgamin=128
pgamax=1024

oracle_mem=$(($total * 40/100 - 64))

sga_target=$(($oracle_mem * 75/100))
pga_target=$(($oracle_mem * 25/100))

if [ $sga_target -lt $sgamin ] ; then
        sga_target=$sgamin
fi
if [ $sga_target -gt $sgamax ] ; then
        sga_target=$sgamax
fi

if [ $pga_target -lt $pgamin ] ; then
        pga_target=$pgamin
fi
if [ $pga_target -gt $pgamax ] ; then
        pga_target=$pgamax
fi

echo sga_target=${sga_target}M
echo pga_aggregate_target=${pga_target}M

--------------------------------------------


and then updated rhnsat orcale SGA/PGA

ORACLE_SID=rhnsat sqlplus / as sysdba
SQL> alter system set sga_target=XXXM scope=spfile;
SQL> alter system set pga_aggregate_target=yyyM scope-spfile;


Rebooted without maxcpus set, now system is up and running with 64cores.

Backported to SATELLITE-5.4 as
commit aedd1dcf2bb14f839e43c3e5a8e68cbe59d8f170
    added correct units to memory params
    (cherry picked from commit 3cf0ddc8fea44a0ece39df243b6339d884079664)
commit ef852351f518b7228563da8a37c7c9eb0b633642
    643368 - size oracle sga dynamicaly
    (cherry picked from commit a9079e153b0668f4b539d7cb0e668584c53360e2)
commit 03b8550f8411a9cfb85f60caa3819a1570d44bd9
    643368 - compute sga size dynamicaly
    (cherry picked from commit 3d80d9f859984ed2c97252c886a432834ef6ae49)

Taking QA contact.

An advisory has been issued which should help the problem
described in this bug report. This report is therefore being
closed with a resolution of ERRATA. For more information
on therefore solution and/or where to find the updated files,
please follow the link below. You may reopen this bug report
if the solution does not work for you.

https://rhn.redhat.com/errata/RHEA-2011-0875.html