Re: Insufficient Privileges / Oracle Not Available

Adley wrote:
> Peace EdStevens,
> The SID should be ELLTST. I've confirmed it with my senior, and it was
> ELLTST until before this happens.
>
> Peace Jeroen,
> I'm sorry if I'm not making myself clear enough. Yes, I wanted to
> startup by issuing "startup" command to an idle instance. The problem
> is, I couldn't even use SYS AS SYSDBA to connect due to insufficient
> privilege, much less starting the instance up. :D
>
> Peace Jon Fife,
> I am using the user 'oracle' and I don't see any problem with this user
> as the primary group is already 'dba' as can be seen below:
>
> ORACLESID=elltst> id
> uid=203(oracle) gid=203(dba) groups=202(oinstall),204(mincomrt)
>
> ORACLESID=elltst> env | grep ORACLE
> ORACLE_SID=elltst
> PS1=ORACLESID=$ORACLE_SID>
> ORACLE_HOME=/usr/oracle/product/8.1.7
>
> ORACLESID=elltst> cat /etc/oratab | grep elltst
> elltst:/usr/oracle/product/8.1.7:Y
>
> No trailing "/" either on the ORACLE_HOME. :D
>
> Peace JG,
> This is my oratab file:
>
> elltst:/usr/oracle/product/8.1.7:Y
> ellacme:/usr/oracle/product/8.1.7:Y
> *:/usr/oracle/product/8.1.7:N
> ellgpd:/usr/oracle/product/8.1.7:N
> ellip:/usr/oracle/product/8.1.7:N
> ellmig:/usr/oracle/product/8.1.7:Y
> elltrn:/usr/oracle/product/8.1.7:N
> ellnew:/usr/oracle/product/8.1.7:N
> elldev:/usr/oracle/product/8.1.7:N
> ellptba:/usr/oracle/product/8.1.7:N
> elltry:/usr/oracle/product/8.1.7:N
>
> I might try rebooting the machine and see if that helps. But I have to
> wait until tonight.
> Regarding ipcs, this is what I got:
>
> IPC status from /dev/mem as of Wed Jul 19 09:05:46 WIB 2006
> T ID KEY MODE OWNER GROUP
> Message Queues:
> q 2097152 0x4107001c -Rrw-rw---- root printq
> q 2 0x0b033801 -Rrw-rw-rw- root system
> Shared Memory:
> m 786432 0x58002097 --rw-rw-rw- root system
> m 1 0x470010ca --rw-r--r-- imnadm imnadm
> m 2 0x580010ca --rw-r--r-- imnadm imnadm
> m 3 0x4d0010ca --rw-r--r-- imnadm imnadm
> m 4 0x490010ca --rw-r--r-- imnadm imnadm
> m 5 0x500010ca --rw-r--r-- imnadm imnadm
> m 6 0x450010d0 --rw-rw-rw- imnadm imnadm
> m 7 0x430010d0 --rw-rw-rw- imnadm imnadm
> m 8 0x420010d0 --rw-rw-rw- imnadm imnadm
> m 9 0x410010d0 --rw-rw-rw- imnadm imnadm
> m 10 0x440010d0 --rw-rw-rw- imnadm imnadm
> m 11 0x435dce60 --rw-rw-rw- root system
> m 12 0x0d002a4a --rw-rw-rw- root system
> m 131086 0x15033801 --rw-rw-rw- root system
> m 131087 0x20033801 --rw-rw-rw- root system
> m 131088 0x16033801 --rw-rw-rw- root system
> m 131092 0x741cc1a0 --rw-rw-rw- root system
> m 21 0x741cc1a1 --rw-rw-rw- root system
> m 22 0x741cc1a2 --rw-rw-rw- root system
> m 23 0x741cc1a3 --rw-rw-rw- root system
> Semaphores:
> s 1048576 0x58002097 --ra-ra-ra- root system
> s 786433 0x44002097 --ra-ra-ra- root system
> s 131074 00000000 --ra-ra-ra- imnadm imnadm
> s 3 0x6200281c --ra-r--r-- root system
> s 4 00000000 --ra-ra-ra- imnadm imnadm
> s 5 00000000 --ra-ra-ra- imnadm imnadm
> s 6 00000000 --ra-ra-ra- imnadm imnadm
> s 7 00000000 --ra-ra-ra- imnadm imnadm
> s 8 00000000 --ra-ra-ra- imnadm imnadm
> s 9 00000000 --ra-ra-ra- imnadm imnadm
> s 10 0x450010d0 --ra-ra-ra- imnadm imnadm
> s 11 00000000 --ra-ra-ra- imnadm imnadm
> s 12 00000000 --ra-ra-ra- imnadm imnadm
> s 13 0x01002885 --ra------- root system
> s 393231 0x0101c6e8 --ra-ra-ra- root system
>
> I'm not sure what to look here. Do you see any problems?

What we'd be looking for is something owned by oracle. In general, there would be one semaphore and one message area per instance (although that can be changed and also varies by platform and configuration). On some unix platforms, if you crash an instance these areas can remain in use, not letting you get new ones to bring an instance up. It is also possible to have two instances hash to the same ID, giving the same sort of realm error as you got. It is also possible (but bizzarro, and probably nothing to do with your situation) to have ownership issues by changing oracle ownership or the executable protection while the instance is up.

Are you sure you didn't change unix kernel configurations when you did the "AIX maintenance?" If you changed the shared memory settings, that could give your symptoms. There just mightn't be enough shared memory for oracle to start. Doublecheck your install guide and https://metalink.oracle.com/metalink/plsql/f?p=130:3:4291818653361394759::::p3_database_id,p3_docid,p3_show_header,p3_show_help,p3_black_frame,p3_font:NOT,169706.1,1,1,1,helvetica#AIX (on other unix there are specific manual kernel settings, on AIX I believe it is implicit in the patching - are you getting any coredumps or other symptoms?).

On a more ridiculous note, I've seen admins mess up the groups, so the oracle dba group is another number than the one everything is installed under...

jg

--
@home.com is bogus.
http://home.pacific.net.au/~turner23/2002/hippy-peace.jpg