how to clean up failed instance's SGA

I am not sure what happened to the database
but I tried to startup the database and received error 27100 whch tells me to clean up the failed instance's SGA

How do I do this ?

Thanks

If just a single instance was running, do the following

1, ipcs --note the shmid and semid from the result

2, ipcrm shm

ipcrm sem

Note: The command syntax may be different in different Unix platform.

Also, if you rae running more instnace on the same machine the procedure is little different

The posting got messed-up.

When do they going to fix the hassle with the "LESS THAN" symbol.

anyway, here it goes again

If just a single instance was running, do the following

1, ipcs --note the shmid and semid from the result

2, ipcrm shm shmid_value from 1

3, ipcrm sem semid_value from 1

Note: The command syntax may be different in different Unix platform.

Also, if you rae running more instnace on the same machine the procedure is little different

None of the instance are running

Are you suggesting to remove the semaphore and shared memory with oracle id ??


[Edited by hoctro on 05-01-2001 at 08:07 PM]

The case I described is when the instance is crashed or aborted. use the "shmid", "semid" given by the ipcs command to remove it, when using ipcrm.

\Prince.

I am not sure who did what to the instance but i need to fix it
So given the error number, is this the correct way to resolve it ?

Doing this way is perfect. If u have anyother instances running you need to find the shm/semid using different way.

yahoo!! you are the man gsprince :)

does this usually happen ?