Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
Home -> Community -> Usenet -> c.d.o.server -> Re: Help getting past missing archivelog file
Stan Brown wrote:
> In <3f743a3f$0$32058$afc38c87_at_news.optusnet.com.au> "Howard J. Rogers"
> <howardjr2000_at_yahoo.com.au> writes:
>
>>Alistair Thomson wrote:
>>> Hi Stan >>> >>> You can restore until cancel or until a point in time, check out the >>> syntax below: >>> >>> SVRMGRL> connect internal >>> >>> SVRMGRL> startup mount db_name; >>> >>> SVRMGRL> recover database until time '2003-09-23:14:40:00'; >>> >>> While still in svrmgrl, execute the normal command to open the database, >>> using the RESETLOGS option. This forces the database to reset the redo >>> log sequence number information in the control files and the online redo >>> log files This in turn makes sure that any redo log entry data that >>> followed the "RECOVER DATABASE UNTIL" time will not be applied to the >>> database: >>> >>> SVRMGRL> alter database open resetlogs; >>> >>> Now backup your database. >>> >>> >>> >>> Hope it helps >>> >>> >>> >>> Alistair
>>I have to say that's a dreadful piece of advice! Sorry, but it is so. The >>guy already has stated that the datafile in question (which is all that is >>being recovered here) only contains indexes, and that those indexes can be >>re-created. Your suggestion to do a database-wide incomplete recovery >>means that you will be causing data loss on the entire database, totally >>unnecessarily. >>The correct response to this sort of situation is to alter database >>datafile X offline drop. That will permit him to open the rest of his>>database, drop the tablespace, re-create it, and then re-create all the >>indexes. And no data will have been lost in the meantime.
>>The only slight bummer to this approach is that it is easier said than >>done to get rid of the index tablespace, because some of those indexes >>might be being used to enforce constraints... and all those constraints >>will therefore have to be disabled before the tablespace drop can be >>accomplished... and then re-enabled at the end of it. But that's merely >>just a lot of hard manual work... it's still better than losing data he >>didn't need to lose in the first place!
Can I suggest that you stop posting about the same subject in different threads? Stick to one thread per problem.
I've answered you elsewhere.
Regards
HJR
Received on Fri Sep 26 2003 - 08:45:28 CDT