Re: Standby Database: How do you make sure the redolog copied is the good one

Here's what I do:

1. notice a new logfile, then pick the seq# out of the filename
2. do a "select max(sequence#) from v$log where archived='YES'" to get the last archived seq#.
3. compare the last archived # to the filename's # and decide to copy, or wait and try again later.

I do very rarely run into problems applying the logfile on the standby, so I'm not 100% that this is foolproof. But I've never been able to figure out what those rare problems are, so, take this for what it's worth.

> We have a standby database, and we use
> a perl scripts to copy all the redo logs
> to the standby site.
>
> The way we do is after we detect the file
> by ls -l, we wait another 30 seconds to make
> sure the redolog writing is finished. But
> it is not foolproof. Sometimes we got corrupted
> redolog files.
>
> Any better ideas except wait for another 30 seconds?
>
> Thanks
Received on Mon May 07 2001 - 17:58:05 CEST