@brian_pond I know you are not a fan of attempting to keep a bunch of parallel ERPNext servers in sync, but please review this idea and give me your thoughts.
(I have such servers already because of how I run backups.)
Okay… Here is another possible approach.
I am not sure if this would work for everyone, but it might work for me in particular because I maintain multiple, identical servers for each of my clients. I do this in order to make use of my Poor Man’s Backup System v2
In my poor mans backup, I use server images to replicate the server multiple times under sub-domains and then send hourly backups to these servers of the entire live database from the production server.
So… Having these backup servers with very complete databases on them, maybe another approach could be done like this:
- create another server using the client server image
- port the latest backup to the new server
- run the OPPOSITE of a data archive and delete all NEW records after a particular historical date (like the fiscal year transition).
- Once all NEW records are gone, make another backup of the now smaller database
Okay… so why do this you might ask?
This approach would give me database backup of the old data I was hoping to archive off using the old data first method that I started this thread on. By deleting all of the newest records and leaving only the old records in the database that I wanted to archive anyway, I now have a valid (and restorable) database of the archive old data that can be loaded up on any of the backup servers to be used for historical research.
To make this method work one would have to delete all of the records that followed the closing of the first fiscal year of the system data. That would give you year one of your archive.
To get year two, you would do the same process except you delete all of year one old data, then also delete all data following the close of fiscal year two. Now you have a restorable database of the year two data.
Do the same for all subsequent years and you eventually wind up with a restorable database for each archived fiscal year. When you have all of the years archived in the order you want them, the final step would be to take the entire database one last time and delete all of the old data (that now exists as yearly restorable databases).
In my best guess, this should give you a set of archive databases for each fiscal year, plus a smaller live system because you have moved all of that old data out of your way. With all of this setup you should then be able to simply run through the first step again at the close of the next fiscal year to create the next in sequence archive database of an old fiscal year.
This is all THEORY !!!
It may very well blow up somewhere, but I gotta ask…
Do you think this might work?
Again, This is my current theoretical approach that would work for ME because I already have this multiple parallel server thing going on in order to keep my poor mans backup system running. If you are using a similar system then maybe it could work for you as well.
I need someone with great database knowledge and skills to let me know if this might work. Any such people in the community here?
This is still an exercise in progress. Have not decided on a good approach yet, but investigating everything.
~BKM