If oracle online database backup is in-progress and there's a instance restart due to switch-over of a fail-over cluster...do i need to perform a new round of online database backup again or can i use the "old" online backup copy and perhaps do something about it to bring it updated ??
any input will be appreciated ;)
You will have to re-backup the tablespace which was being backed-up and all other tablespace that are yet to be backed-up. You can still use the backup of tablespaces that have been successfully completed before instance shutdown.
if you cannot open database you would have to mount the database first and do
alter database datafile 'xxx' end backup
then alter database open
and backup the affected tablespace/s again
do u mean that if its like the scenario i mentioned..after the oracle instance restarted..database definitely wont be able to open is it ?? so u hv to end the online backup of the tablespace(that was interrupted) first correct???
yes but since you cannot do alter tablespace xxx end backup in mount state you have to do the way I posted
the problem is if you try to start the database most probably Oracle will complain about file syncronization problems and to overcome that (unfreeze datafile header) we have to end datafile backup
Yes, pando thats the way to do it, but from experience you need to backup the whole database again to have a consistent backup and make it easy for you if you end up in a restore/recovery situation. You know how it tends to be when this situation occurs, managers hanging all over the place trying to speed the process up!
well this only happened to me once and what I did is what i posted so if we have to talk about experience on this kind of situation mine is little however from theoretical view i dont see why I should backup whole database again if I was doing backup tablespace after tablespace. If it crashes I simply start backing up my last tablespace again since I have all the archived logs I dont see what can go wrong if I have to recover anything later?
well if you do a backup per tablespace it's OK with your approach, but if you are running a full online consistent backup that backup is not valid any more!
ahh another missunderstanding :D
I never do that kind of backups, generates too much archived logs!