ALL DB = Primary + DR databases.
-> Implement checkfile concept so script will not page for planned maitenance.
(1) Connect and update all Primary db for hang check.
(2) Check for pmon for all db.
(3) Check Alert log for ORA error for Primary database.
(4) Purge Alert log of Primary and keep backup for last 15 days.
(5) Nullify listener.log of Primary every Month/15days/week.
(6) Switch logfile for each Primary for every 15 mins.
(7) Purge old archlog file that is 1 week old. ( This needs to check for the logs must be applied on all DR boxes, Else Page Immediately)
(8) Purge old acrhfile based on threshold. ( This needs to check for the logs must be applied on all DR boxes, Else Page Immediately)
(9) Check for space on all segments in the databases.
(10) Check for MAXEXTENTS reached in all databases.
(11) Copy the listener.ora / tnsnames.ora from Primary to its DRs whenever any changes done.
(12) Do tnsping on all database link from Primary and its DRs.
(13) Run Autorecovery for all DRs.
(14) Check recovery for all DRs.
(15) Purge all tracefiles for every week.
(16) Check all .err files in $SCRIPTS/log folder for any kind of Errors and page if that box is production.
(17) Copy arch files from Primary to DR.
(18) Copy all crontabs for all production server.
(19) Implement checkfile concept. So this scripts will not page when there will be a planned maintenance.
(20) Preserve the backup of all monitoring scripts. (Weekly or manually).
(21) Check for consistent locks in the databases and email it periodically of locks are held for some threshold time.
(22) Query on vmstat output to see if there is any CPY hike that crossed certain thereshold.
(23) For JDBC based applications, monitor open_cursors count from V$sysstat for databases.
(24) Validate standby for any NOLOGGING operations done on Priduction.
For statistics gathering:
(1) Run vmstat on each server to collect stats for every 5 mins.
(2) Check statspack/awr is running for 9i/10g instances.
Other periodic tasks
(1) Analyze weekly.
Thanks & Regards
Paresh Mehta