I want my mongodb indexes to not be loaded/used by RAM. Instead I want them to work by reading and writing from Local Storage itself and not RAM.
Welcome to the MongoDB Community Forums @Shubham_Singh4!
The I/O performance of RAM is orders of magnitude faster than local storage, so all databases (not just MongoDB) generally require sufficient RAM to support common workloads for a deployment.
If performance is not a consideration, you could try minimising the amount of secondary indexes you create and limit available resources using containers or
cgroups. However, you may be better served by a different solution optimised for low memory environments.
Can you elaborate on your use case for avoiding using RAM for indexes?
Thanks for your response @Stennie,
I actually want to try out a way in which indexes work without being loaded into RAM. All indexing operations must occur from the local storage only.
I’m not aware of a database which completely avoids loading indexes into RAM as they are used. If limiting available resources (which will force the database server to swap to/from disk more regularly) doesn’t meet your testing scenario you perhaps want to look into an embedded solution or file-based library rather than a database server.
I still don’t understand the use case. Is this for an academic experiment?
Yeah it was for an academic experiment. Thanks for your help. Peace.
This topic was automatically closed 5 days after the last reply. New replies are no longer allowed.