Background: We are trying to troubleshoot a backup problem for a client. They are using dataprotector from HP to back up a couple of domino servers with dom.doc on them. The servers are running with transactional logging in archived mode to allow point in time restores. There are a few hundred MB of changes perday to the databases on 300GB of total data. The servers are 6.0.4 on Windows 2000
Problem :The problem is that the incremental backups are almost always backing up every database on the server and thus taking too long (basically a full backup nightly). We have been asked to confirm the functionality of the NSFIsNewBackupNeeded api query that is used by HP and see what it returns. We wrote an agent using a ComfortSpan of (0) against the data on their servers and our and got this type of data:
from what we understand (have read) nsfisnewbackupneeded will:
"Determine the state of the database or database backup file with respect to the current log state. This function either returns a Boolean (0 or 1 value) for backupNeeded, if a comfortSpan is supplied (specified in K bytes), or returns the actual span of log needed for recovery, if a comfortSpan of 0 is supplied (the only supported configuration for archived logging). The comfortSpan is the amount of log to be remaining before the backup would be made useless by a logging wrap in a 'CIRCULAR' type logged system (see Transactional Logging settings in the Administering the Domino System documentation)."
So we are ending up with a value of 17Gigs as needed to recover the database? The other odd thing is the values are almost identical regardless of database size (they are not 17gigs each) and also they roughly decrease in numerically in alphabetical order.
Does anyone out there have a clue what these numbers might actually represent?