BCSS question

[JohnFLand]

I'm comparing a BCSS file of Drive A to real files on Drive B. On the BCSS side, after the compare, I get a tree of orphan "files". I'd like to select those orphans and snapshot them to a new BCSS file (thus, while the original BCSS file has a list of about 200,000 files, the new BCSS file may only have a list of about 200 files).

However, I can't seem to take a snapshot of only selected files -- the process starts writing out a snapshot of ALL files on the BCSS side. Am I doing something wrong, or can't BC3 snapshot just selected files?


John Land

[Aaron]

Hello,

The Snapshot command does do the entire side, but you can use Edit->Inverse Selection, then RightClick Exclude to filter out the previously unselected items. Then uncheck the Snapshot option "Items excluded by file filters".

[JohnFLand]

Thanks for the work around -- I'll try it. I'm not sure what the speed is going to be: my "small" file set is about 700,000 files, and I'll be excluding over 99% of it, typically. The orginal compare usually takes about 20 minutes (slow processor), and results in a few hundred orphan files; having to filter out all the other files just to take a snapshot of the orphan files is likely to take another 20 minutes, whereas selecting the orphaned files is instantaneous.

I'll like to request as a feature the ability to snapshot selected files only (another difference between filtering and selecting that is useful).

[JohnFLand]

Confirmed: the workaround time is essentially 2N, where N is the original filter time. Snapshoting only selected files would be essentially N+1 second.

[JohnFLand]

I wrote too soon -- the workaround apparently won't work -- I came back to my machine several hours after I determined that filtering would take at least as long as the initial compare to find BC3 still churning away (at 90%+ CPU usage).

I then tried to filter my "tiny" file set of about 156,000 files, selecting about 20 files, inverting the selection (which is fast) and then trying to exclude the files from the inverted selection -- BC3 doesn't appear to be able to do so in any reasonable time.

Have you tried your process with large file sets?

[JohnFLand]

So I did further testing, figuring that maybe the problem was that I'd expanded all folders, selected 20 or so files, inverted the selection, and then tried to exclude the selected inverse -- which means BC3 was trying to apply a filter with 150,000+ entries.

So I collapsed all folders, expanding only one folder, about 5 directories deep, for a test. I selected 20 or so files, inverted the selection, then right clicked and tried to exclude the now selected inverse -- which went really fast (since the filter list was mostly top level directories), BUT the problem now is that my desired 20 or so are all filtered out -- that is, NO files appear at all. Why? Because when you invert the selection, you are then selecting (for exclusion) the ancestor folders that contain the selected files. I consider this a bug (that is, excluding all inverse files and folders should not exclude the files and folders originally selected).

[Aaron]

Thanks for the additional info, John. I'll pass it along to our developers and see if they can think of a better solution for you.

[JohnFLand]

Developers: PLEASE consider allowing snapshotting of only selected files; this would make "sneakernet" syncing of two systems MUCH faster, since only copies of changed files and a snapshot of leftover orphans needed be carried to a remote site to sync it up to a local site.