I have read the previous parallel vs sequential threads. I have 5 drives/locations I keep synchronized. I copy from my C drive to my One Drive, and 3 others. I notice that all the mismatches are computed at once. This means I often need to re-run the batch file again and sometimes again before everything is synchronized. What I would like is to be able to scan for mismatches for one folder pair then synchronize and then go to the next folder pair, scan and synchronize in a loop. Like this C<>1D 1D<>Z Z<>Q Q<>V V<>C. I realize that I will need to run this twice if there is an update on the Z drive it won't be updated on the C drive until I run the batch file again. I update often. I don't use the data base to detect changes. Just time and size.
Right now my alternative is to make a .cmd file with a series of robocopy executions since each gets done one at a time.
More Parallel vs Sequential processing
- Posts: 19
- Joined: 30 Mar 2022
-
- Posts: 4866
- Joined: 11 Jun 2019
You will need (n-1)! folder pairs to do this:
C<>1D
C<>Z
C<>Q
C<>V
1D<>Z
1D<>Q
1D<>V
Z<>Q
Z<>V
Q<>V
This setup pairs every location with every other one, so any change in one location propagates in one operation. The number of pairs can be reduced if some location are backup-only. For example:
If Z, Q, and V are all backups and never used directly, you can change the folder pairs syncing C or 1D to them so they are 'mirror' syncs instead of 'two-way'. You can also entirely eliminate the folder pairs syncing between those three locations, reducing the overall sync by six pairs.
C<>1D
C<>Z
C<>Q
C<>V
1D<>Z
1D<>Q
1D<>V
Z<>Q
Z<>V
Q<>V
This setup pairs every location with every other one, so any change in one location propagates in one operation. The number of pairs can be reduced if some location are backup-only. For example:
If Z, Q, and V are all backups and never used directly, you can change the folder pairs syncing C or 1D to them so they are 'mirror' syncs instead of 'two-way'. You can also entirely eliminate the folder pairs syncing between those three locations, reducing the overall sync by six pairs.
-
- Posts: 2946
- Joined: 22 Aug 2012
In addition to xCSxXenon's suggestions:
If the scope of all above syncs is all the same (the same set of files/folders is to be synced) it might be better to designate one of the above location as your central point of syncing.
I will assume C, but it can be any of the above.
Then create a single sync having multiple left-right pairs.
C<>1D
C<>Q
C<>V
C<>Z
After running the above sync twice in succession, all your drives should be completely in sync (assuming no changes occurred during or in between the two syncs)
Or if (as per xCSxXenon's suggestion) Q, V and Z are only backups, you can have a first sync with just a single pair
C<>1D
and a second sync with the pairs (I use C, but it can also be 1D)
C<>Q
C<>V
C<>Z
You can even automatically launch the second sync from the "Run a command" command-line in the sync settings (F8) of the first sync.
If the scope of all above syncs is all the same (the same set of files/folders is to be synced) it might be better to designate one of the above location as your central point of syncing.
I will assume C, but it can be any of the above.
Then create a single sync having multiple left-right pairs.
C<>1D
C<>Q
C<>V
C<>Z
After running the above sync twice in succession, all your drives should be completely in sync (assuming no changes occurred during or in between the two syncs)
Or if (as per xCSxXenon's suggestion) Q, V and Z are only backups, you can have a first sync with just a single pair
C<>1D
and a second sync with the pairs (I use C, but it can also be 1D)
C<>Q
C<>V
C<>Z
You can even automatically launch the second sync from the "Run a command" command-line in the sync settings (F8) of the first sync.
