Re: Standby database: Skipping logs?

There is a parameter MIN_ARCHIVE_SUCCEED_DEST (or something very similar) which controls how many of the archive destinations a redo log has to be written to before oracle considers the write successful. I believe the default for this parameter is one.

HTH

--
Niall Litchfield
Oracle DBA
Audit Commission UK
"Vikas Agnihotri" <onlyforposting_at_yahoo.com> wrote in message
news:77e87b58.0107020646.4eb5faf7_at_posting.google.com...


> Oracle 8.1.7/Solaris 2.7

>

> We were using a home-grown method to manually 'rcp' archivelogs to a

> remote site and apply them 'recover database using backup

> controlfile'. Works fine, but is hell to fix when it breaks. I just

> wanted to modernize this stuff and so I decided to go in for Oracle's

> standby database feature.

>

> Tested it on a couple of dev/QA databases, and it worked fine.

>

> Set it all up in parallel with our own method and was monitoring it

> during times of heavy db activity.

>

> To my horror, I saw that Oracle was skipping redologs. i.e. the

> standby machine was just not getting some logs.

>

> Even when I set the log_archive_dest_2='SERVICE=sby MANDATORY

> REOPEN=5', it was still skipping.

>

> I had 2 Unix windows open. The primary's alert.log said 'Completed

> archiving log sequence #1234', but the standby's standby_archive_dest

> did not have the file. There were ABSOLUTELY no errors on either the

> primary or standby alert log. In other words, Oracle was silently

> skipping archivelogs!

>

> This is totally unacceptable.

>

> Anyone else experienced this? Couldnt find anything on Metalink.

>

> I would hate to think that Oracle's much touted standby database

> feature is a bunch of hooey when faced with real-life situations.

>

> For now, I am back to my old-fashioned (but works!), rcp/apply method.

>

> Thanks...