This has been happening over the years but I couldn’t figure out why….finally found this article on My Oracle Support (MOS): Setup Local archiving on standby to Prevent Logs sent to $ORACLE_HOME/dbs (on both GRID and SQLPLUS) (Doc ID 1271095.1)
I previously thought this was due to a misconfigured rman backup configuration/implementation.
Just a note here on using this document…
If you only set two archiving destinations on a standby your database may not start in the following situation. The following sentence is a quote from the article.
“Check for at least one log_archive_dest_<n>, should point to a available location and the VALID_FOR parameter should be either VALID_FOR=(ALL_LOGFILES,ALL_ROLES) or (STANDBY_LOGFILES,STANDBY_ROLE).”
When I set two archiving locations, the database dumped core on restart , not mounting while complaining it had no destination to archive to:
LOG_ARCHIVE_DEST_1=’LOCATION=/arch2/standby/ VALID_FOR=(STANDBY_LOGFILES,STANDBY_ROLE) DB_UNIQUE_NAME=standbydb’
log_archive_dest_2 =’SERVICE=PRIM LGWR ASYNC VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=PRIM
Changing the following log_archive_dest_1 for all roles fixed the issue.
LOG_ARCHIVE_DEST_1=’LOCATION=/arch2/standby/ VALID_FOR=(ALL_LOGFILES,ALL_ROLES) DB_UNIQUE_NAME=standbydb’
I don’t know if this is working as it was designed…or a bug. Something unique about my environment is that the FLASH_RECOVERY_AREA is different on the primary and the standby.