postmaster does not want to restart on Redhat 5.1

Since we had an unexpected power failure on Friday, I have had numerous
problems starting postmaster. Problems which did not exist before.
Friday morning power failure.
Friday night i notice postmaster isn't running. I managed to start the
server with the following command :
postmaster -d 3 -D /var/lib/pgsql > server.log 2>&1 &

which seemed to work. Unfortunately i had to stop the service on Sunday
because I had forgotten about the TCP/IP socket option (-i). After adding
the -i option it refused to work.
I looked for traces of lock files, removed them. The PostgreSQL-Linux FAQ
said that I should recompile the kernel with IPC. However the database
system seemed to be running fine before that, so I don't think it could have
been a problem with IPC not being in the kernel.

I don't know what to try next but I would appreciate any suggestions.

---Debug info---

System : Redhat 5.1
kernel : 2.0.35
PostgreSQL : version 6.3 (rpm)

---command to start postmaster (located in /etc/rc.d/init.d/postgresql)

/usr/bin/postmaster -i -a nounauth -S -D/var/lib/pgsql

---output of postmaster

[root(at)hive init.d]# postgresql start
Starting postgresql service: IpcMemoryCreate: memKey=5432010 , size=24588 ,
perm
ission=448IpcMemoryCreate: shmget(..., create, ...) failed: Identifier
removed
This type of error is usually caused by improper
shared memory or System V IPC semaphore configuration.
See the FAQ for more detailed information
IpcMemoryIdGet: memKey=5432010 , size=24588 , permission=0IpcMemoryIdGet:
shmge
t() failed: Identifier removed
This type of error is usually caused by improper
shared memory or System V IPC semaphore configuration.
See the FAQ for more detailed information
IpcMemoryAttach: shmat() failed: Invalid argument
This type of error is usually caused by improper
shared memory or System V IPC semaphore configuration.
See the FAQ for more detailed information
FATAL 1: AttachSLockMemory: could not attach segment
postmaster []