This may happen if database was killed during long running transaction which generated a lot of undo. Any messages in alert file about smon activity ? Michael On Fri, Sep 12, 2008 at 7:39 AM, Yaping Chen <hangzhoumaster@xxxxxxxxx>wrote: > Hi, > > You may use tools such as truss/strace etc(depend on OS) to check which > steps are slow. > > 2008/9/12 raja rao <raja4list@xxxxxxxxx> > > eventhough the db is big around 1 TB, all datafiles are in single >> mountpoint. >> >> but i am really looking to start debugging (dont knwo if oradebug or >> someother utility will help what is happening after mount and before open). >> >> >> --- On *Fri, 9/12/08, Prasad <p4cldba@xxxxxxxxx>* wrote: >> >> From: Prasad <p4cldba@xxxxxxxxx> >> Subject: Re: startup takes long time on 10gR2. >> To: raja4list@xxxxxxxxx >> Cc: oracle-l@xxxxxxxxxxxxx >> Date: Friday, September 12, 2008, 9:08 AM >> >> >> how about checking the datafiles placements. probably it will be a good >> way to start debugging is by starting with making datafiles offline for some >> non-system tablespaces and see if it helps. >> >> On Thu, Sep 11, 2008 at 8:26 PM, raja rao <raja4list@xxxxxxxxx> wrote: >> >>> One of our test instances, the startup process is taking more time. >>> >>> >>> sga is ~2gb. >>> >>> >>> Total System Global Area 2130706432 bytes >>> Fixed Size 2031448 bytes >>> Variable Size 1140850856 bytes >>> Database Buffers 973078528 bytes >>> Redo Buffers 14745600 bytes >>> >>> especially, the open phase takes long time.(around 10 mins) >>> instance allocation and mount phase are quick. >>> >>> >>> Can someone help me, how to start research and reduce the time. >>> >>> Thanks, >>> Raj >>> >>> >>> >> >> > > > -- > Regards, > Yaping Chen > > http://yaping123.wordpress.com > -- Best Regards Michael Elkin