On 27 April 2010 21:30, Chris Barnes <compuguruchrisbarnes@xxxxxxxxxxx> wrote:
I have just initialized a database with no errors, used the postgresql.conf file from another system running 8.4.2.
Attempted to start and got the fatal error below.
I than ran pgtune and got the same error.
I am not sure what the problem is? Is there more detailed logging than below, or can be turned on?
CentOS release 5.4 (Final) (installed and working)
CentOS release 5.2 (Final) (failing)
after pgtune
olap.rownum_name = 'default'
maintenance_work_mem = 1GB # pg_generate_conf wizard 2010-04-27
effective_cache_size = 22GB # pg_generate_conf wizard 2010-04-27
work_mem = 96MB # pg_generate_conf wizard 2010-04-27
shared_buffers = 7680MB # pg_generate_conf wizard 2010-04-27
~
~
[root@rc-rec-five pgtune-0.9.0]# /etc/init.d/postgresql stop
Stopping postgresql service: [FAILED]
[root@rc-rec-five pgtune-0.9.0]# /etc/init.d/postgresql start
Starting postgresql service: [FAILED]
[root@rc-rec-five pgtune-0.9.0]# cat pgstartup.log
cat: pgstartup.log: No such file or directory
[root@rc-rec-five pgtune-0.9.0]# cat /data/pgsql/data/pgstartup.log
2010-04-27 16:19:17 EDT FATAL: requested shared memory size overflows size_t
2010-04-27 16:22:27 EDT FATAL: requested shared memory size overflows size_t
/var/log/messages/
Apr 27 13:57:56 rc-rec5 ntpd[2990]: synchronized to 206.248.171.198, stratum 1
Apr 27 16:13:45 rc-rec5 postgres[5717]: [1-1] 2010-04-27 16:13:45 EDT FATAL: requested shared memory size overflows size_t
Apr 27 16:19:17 rc-rec5 postgres[7736]: [1-1] 2010-04-27 16:19:17 EDT FATAL: requested shared memory size overflows size_t
Apr 27 16:22:27 rc-rec5 postgres[9648]: [1-1] 2010-04-27 16:22:27 EDT FATAL: requested shared memory size overflows size_t
This might sound like a strange question, but are the architectures of the servers different? Specifically, is the failing server 32-bit and the working server 64-bit? And what architecture is PostgreSQL built against on the failing server?
Thom
