oke wang wrote:
> Hi,
>
> Primary Node
> OS: RAS3 Oracle: 9204
> log_archive_dest:
> log_archive_dest_1:location=/opt/oracle/oradata/orcl/archive
> log_archive_dest_2:service=standby arch sync noaffirm
> archive_lag_target:300
> log_archive_min_succeed_dest:1
> fal_client:
> fal_sever:
> 　standby_archive_dest：?/dbs/arch
> 　standby_file_management：MANUAL
> For testing,I disconnected standby node from the network,then
> repeated the sql 'alter system switch logfile', when I run the sql
the
> third time,the sql lasts about 17 minute.(the primary database has
> three online redo groups).after that the sql can run fast
again.Why?17
> minute is too long for a production Environment.
> But everything is fine when I just stop the listener of standby
node
> instead of disconnecting it from the network.
We encountered a similar problem in our DG setup. It was caused by
network timetouts when Oracle tried again and again to reach the downed
node. In our case we have solved this by:
MAX_FAILURE=200 REOPEN=180 NET_TIMEOUT=30
So it is not trying to reach the downed node too often and infinitely
anymore. After applying of this everything is OK.
--
Dusan Bolek
Received on Wed Apr 13 2005 - 02:29:22 CDT
