Graham, December 01, 2016 - 10:18 am UTC
I'd definitely run some crosscheck commands in RMAN to see what that thinks you already have backed up or still have available. You could be unnecessarily backing up multiple copies of the same archivelog if they aren't being managed through an archivelog deletion policy.
Are you backing up to disk first or straight to tape (or somewhere similar) or even over the network to a cross mounted filesystem? Is it purely your archivelog backups taking longer or database backups too (that your level 0 and 1?)? Could be contention if you're asking for tape channels that are being used or requested elsewhere.
You could always have a friendly chat with your server/storage/network guys to see if they've changed any settings. We've occasionally seen dips (and also increases) in performance whenever TSM throughput settings have "mysteriously" been altered. Could be that but if your database backups are fine then perhaps not. Of course, we'd never blame third party vendors...
Could maybe be an RMAN setting? Is backup optimisation on? Sometimes the filesperset parameter in your backup script can be in need of adjustment, we've gradually altered ours throughout the years from 20 incrementally down to 10 or 15 where we've found a nice sweet spot.
December 02, 2016 - 2:32 am UTC
Nice input.
Graham, December 01, 2016 - 10:19 am UTC
I'd definitely run some crosscheck commands in RMAN to see what that thinks you already have backed up or still have available. You could be unnecessarily backing up multiple copies of the same archivelog if they aren't being managed through an archivelog deletion policy.
Are you backing up to disk first or straight to tape (or somewhere similar) or even over the network to a cross mounted filesystem? Is it purely your archivelog backups taking longer or database backups too (that your level 0 and 1?)? Could be contention if you're asking for tape channels that are being used or requested elsewhere.
You could always have a friendly chat with your server/storage/network guys to see if they've changed any settings. We've occasionally seen dips (and also increases) in performance whenever TSM throughput settings have "mysteriously" been altered. Could be that but if your database backups are fine then perhaps not. Of course, we'd never blame third party vendors...
Could maybe be an RMAN setting? Is backup optimisation on? Sometimes the filesperset parameter in your backup script can be in need of adjustment, we've gradually altered ours throughout the years from 20 incrementally down to 10 or 15 where we've found a nice sweet spot.