Re: Shutdown Immediate Command

From: <robertgfreeman_at_my-deja.com>
Date: Thu, 18 Jan 2001 23:17:21 GMT
Message-ID: <947thn$3mq$1@nnrp1.deja.com>

Agreed, thats how it's supposed to work. I've seen corruption issues where the client/database thought the redo had been written, but in fact it was corrupted. In one case in particular this was due to controller failure. In this case, we did a shutdown abort. We could not restart the database because of a corrupted online redo log. A normal shutdown or shutdown immediate, because of the associated checkpoint, would have solved this problem becuase the datafiles would have been checkpointed, and thus at an SCN that was at the point of the ABORT (or bloody close to it).

Bottom line is, if you are going to do an ABORT, then checkpoint if at all possible.

Robert

In article <3A673845.23BB_at_btinternet.com>, High.Flight_at_btinternet.com wrote:
> The client won't be sent a 'statement processed' message until LGWR
> indicates that is has finished writing to the online log group.
>
> The client may issue a commit, but it will never hear that the commit
> was ever acknowledged - hence the client knows that the commit was not
> performed.
>
> Easy really.
>
> D.P.
>

Sent via Deja.com
http://www.deja.com/ Received on Thu Jan 18 2001 - 17:17:21 CST