Best strategy for OS upgradation?

Your question is how to remove some of the nodes from the existing cluster right?
Check the below document link and hope this will help.
https://www.vertica.com/docs/9.2.x/HTML/Content/Authoring/AdministratorsGuide/ManageNodes/RemovingNodesFromADatabase.htm

@Nimmi_gupta that can be one of the way removing 9 nodes from existing cluster(24 nodes) and use copy cluster between similar cluster to replicate data in new 15 node cluster with updated os and robust disks.
I was wondering any other alternative without removing nodes migrate all the 20 tb data (10k+ tables) to new 15 node cluster.

Copying data between two cluster with different number of nodes, check the below link
http://vertica.com/kb/Copying-Data-Between-Dissimilar-Vertica-Clusters/Content/BestPractices/Copying-Data-Between-Dissimilar-Vertica-Clusters.htm

Yeah @Nimmi_gupta saw that one earlier.....i guess it not that handy when replicating 10K plus tables.....I mean export...import

In that case only option would be remove the node from 24 node cluster and bring it down to 15 and do the copycluster.

Thanks @Nimmi_gupta

@SK21, If the source Enterprise database version is 10.0 or above,
And the amount of RAM and CPU power on the 24 nodes is equal or above the new 15 nodes,
Then another option one may consider is to use the nodes with large storage to play as MinIO (S3) communal storage
and migrates the Enterprise database to Eon Mode via the meta-function MIGRATE_ENTERPRISE_TO_EON
Advantages:

• Not to reduce the amount of nodes (keep the 24 nodes) while using the new nodes storage.
• The ability to easily add or remove nodes and resize the cluster
• Add more data
• Allocate certain nodes via subclustering
• Achieve workload isolation for different teams with different compute requirements.