SSTable increase a lot after repair

I'm trying repair in one of the node nodetool repair -full -st start_token -et end_token keyspace table_name I found the SSTable count of this table continusly increase and it's useless on compact commend nodetool compact keyspace table_name Eventually the node shutdown. Issue: 1. It seems the compaction after repair or nodetool compact will fail, there's many error log as bellow > ERROR [CompactionExecutor:3743] 2025-07-15 05:04:33,672 CassandraDaemon.java:119 - Exception in thread Thread[CompactionExecutor:3743,5,main] java.lang.AssertionError: point2 should follow point1 2. When I try to start the node , it stops when initial because there're too many SSTable & metadata file. But the open file parameter is already the limit of my server. > ERROR [SSTableBatchOpen:38] 2025-07-21 14:31:51,170 CassandraDaemon.java:119 - Exception in thread Thread[SSTableBatchOpen:38,5,main] java.lang.RuntimeException: java.nio.file.FileSystemException: /db/cassandra/data/keyspace/table_name-ba0eccc225f811e9ae0c059871c9553d/aa-9273-bti-Partitions.db: Too many open files Thanks for any suggestion.