I have been in such situation where somehow we lost or (most recently) got corrupt the archive logs from production site and we had to synchronize our standby/DataGaurd site. Anyhow the terrible fact about such situation is that we had to recreate standby database, here is the scenario with solution :)
Scenario: Your database is archiving itself to shared storage (of course in RAC) or it is using local disk (in NON-RAC) and somehow your disk get failed and you lost all/(some of) your archive files, or oracle wrote archive log on a corrupt block and a specific archive log file is lost. In either case you lost archives before transferring to standby/Dataguard database.
Consequences: You d'nt have any other way to regenerate that archive file and you are now to a point where you have to recreate that specific standby database.
Protection: In our office, we have a slight protection against such failure i.e. we have a shell cron script that is running after every 5 min. and coping archive log files to a NFS mounted disk. This is very handy in cases when we lost archive log file on disk but it is found in NFS mounted disk. BUT last week we even did'nt had those missing archive log files in NFS mounted disk.
Solution: You have to undergo with following procedure.
Step 1: Create a new rman based backup, EXCLUDING Archive Log Files, Including control file for standby.
Step 2: Copy all archives to standby site, those who are recent then your last backup. i.e. those archive log file that were generated after you made step 1 backup.
Step 3: recreate your standby database just as normal.
PS: you can take help from my post to accomplish step 3.
Request: if you know any other better way then this, d'nt forget to share with me :p
Cheers
Subscribe to:
Post Comments (Atom)
3 comments:
If its 10g applying an incremental backup to a physical standby would be another option.
I mean 10gR2.
Hi,
to reconstruct standby you would require a full backup! incremental would not serve the cause!
cheers
Post a Comment