diff --git a/ledger/src/blockstore_db.rs b/ledger/src/blockstore_db.rs
index f7003112c7a..c9ef037032f 100644
--- a/ledger/src/blockstore_db.rs
+++ b/ledger/src/blockstore_db.rs
@@ -95,9 +95,6 @@ const PROGRAM_COSTS_CF: &str = "program_costs";
 /// Column family for optimistic slots
 const OPTIMISTIC_SLOTS_CF: &str = "optimistic_slots";
 
-// 1 day is chosen for the same reasoning of DEFAULT_COMPACTION_SLOT_INTERVAL
-const PERIODIC_COMPACTION_SECONDS: u64 = 60 * 60 * 24;
-
 #[derive(Error, Debug)]
 pub enum BlockstoreError {
     ShredForIndexExists,
@@ -281,7 +278,7 @@ impl Rocks {
                     path,
                     Self::cf_descriptors(&options, &oldest_slot),
                 )?,
-                access_type: access_type.clone(),
+                access_type,
                 oldest_slot,
                 column_options,
                 write_batch_perf_status: PerfSamplingStatus::default(),
@@ -302,70 +299,13 @@ impl Rocks {
                         &secondary_path,
                         Self::cf_descriptors(&options, &oldest_slot),
                     )?,
-                    access_type: access_type.clone(),
+                    access_type,
                     oldest_slot,
                     column_options,
                     write_batch_perf_status: PerfSamplingStatus::default(),
                 }
             }
         };
-        // This is only needed by solana-validator for LedgerCleanupService so guard with AccessType::Primary
-        if matches!(access_type, AccessType::Primary) {
-            for cf_name in Self::columns() {
-                // these special column families must be excluded from LedgerCleanupService's rocksdb
-                // compactions
-                if should_exclude_from_compaction(cf_name) {
-                    continue;
-                }
-
-                // This is the crux of our write-stall-free storage cleaning strategy with consistent
-                // state view for higher-layers
-                //
-                // For the consistent view, we commit delete_range on pruned slot range by LedgerCleanupService.
-                // simple story here.
-                //
-                // For actual storage cleaning, we employ RocksDB compaction. But default RocksDB compaction
-                // settings don't work well for us. That's because we're using it rather like a really big
-                // (100 GBs) ring-buffer. RocksDB is basically assuming uniform data write over the key space for
-                // efficient compaction, which isn't true for our use as a ring buffer.
-                //
-                // So, we customize the compaction strategy with 2 combined tweaks:
-                // (1) compaction_filter and (2) shortening its periodic cycles.
-                //
-                // Via the compaction_filter, we finally reclaim previously delete_range()-ed storage occupied
-                // by pruned slots. When compaction_filter is set, each SST files are re-compacted periodically
-                // to hunt for keys newly expired by the compaction_filter re-evaluation. But RocksDb's default
-                // `periodic_compaction_seconds` is 30 days, which is too long for our case. So, we
-                // shorten it to a day (24 hours).
-                //
-                // As we write newer SST files over time at rather consistent rate of speed, this
-                // effectively makes each newly-created sets be re-compacted for the filter at
-                // well-dispersed different timings.
-                // As a whole, we rewrite the whole dataset at every PERIODIC_COMPACTION_SECONDS,
-                // slowly over the duration of PERIODIC_COMPACTION_SECONDS. So, this results in
-                // amortization.
-                // So, there is a bit inefficiency here because we'll rewrite not-so-old SST files
-                // too. But longer period would introduce higher variance of ledger storage sizes over
-                // the long period. And it's much better than the daily IO spike caused by compact_range() by
-                // previous implementation.
-                //
-                // `ttl` and `compact_range`(`ManualCompaction`), doesn't work nicely. That's
-                // because its original intention is delete_range()s to reclaim disk space. So it tries to merge
-                // them with N+1 SST files all way down to the bottommost SSTs, often leading to vastly large amount
-                // (= all) of invalidated SST files, when combined with newer writes happening at the opposite
-                // edge of the key space. This causes a long and heavy disk IOs and possible write
-                // stall and ultimately, the deadly Replay/Banking stage stall at higher layers.
-                db.db
-                    .set_options_cf(
-                        db.cf_handle(cf_name),
-                        &[(
-                            "periodic_compaction_seconds",
-                            &format!("{}", PERIODIC_COMPACTION_SECONDS),
-                        )],
-                    )
-                    .unwrap();
-            }
-        }
 
         Ok(db)
     }