> > I had the flush params commented out = set to defaults. > > Now trying with: > > MaxUnflushedWrites = 1000 > > MaxUnflushedWriteTime = 300 > > > > I assume that should imply that a delay of 300 seconds (5 min) after > > NBACKUP -L should be sufficient. Will try with ten minutes. > > > > Didn't work. I still got a modified file error on copy. > > Now trying with these settings > MaxUnflushedWrites = 100 > MaxUnflushedWriteTime = 60 > > And also 100 select transactions in an isql script between lock and > copy. If this works as expected, I would assume that each transaction > results in at least one disk write (header page?), which should force a > flush. But perhaps the MaxUnflushesWrites param doesn't mean number of > write operations to any page, but rather number of pages that have been > written to? In that case I'm not sure my dummy SQL will have any effect. MaxUnflushesWrites is not about every single page write. It is sooner about "write process" as a whole. Every write (not read-only) transaction finishes with such "write process": when all pages modified by this (and probably some others) transaction is writen to disk.
> I'm reluctant to use a script/mode/setting that essentially ignores the > flush and copies anyway. I assume this would imply a risk of a corrupt > copy. If not, please explain why. There is no writes to the souce file contents and copy can't be corrupted. In any case - you asks for opinions and mine is above. This is just an opinion. Decision is up to you ;) Regards, Vlad
