Hi there,
I am currently running an M50 with about 1GB collection size, 50 GB indices and 897GB disk usage. Note: collection size was much more before thats why I needed M50, but online archived most of the data now.
Based on the collection size, data should fit into M30 and even M10. However, I am struggling with down scaling to M30 with 100GB. I am receiving the following error message:
“The selected disk size of 100.0 GB is smaller than the amount of data currently used (897.3117332458496 GB)”.
Are the 897 GB just allocated disk space that cannot be down scaled or actual data used besides the collection/indices?
I believe that although data was deleted during archiving process it is still allocated on the disk space.
You should compact your secondaries using the compact command conneting to each secondary one by one. Finally when both secondaries dropped the disk size failover the primary using “Test Failover” and compact it.
When all nodes are below the 100GB mark downgrade the instances to 150GB.
I would not recommend going to M10 directly and keep the database on M30 as a minimum workload demanding instance.