diff --git a/src/index/base.h b/src/index/base.h
--- a/src/index/base.h
+++ b/src/index/base.h
@@ -30,7 +30,7 @@
         bool ReadBestBlock(CBlockLocator &locator) const;
 
         /// Write block locator of the chain that the txindex is in sync with.
-        bool WriteBestBlock(const CBlockLocator &locator);
+        void WriteBestBlock(CDBBatch &batch, const CBlockLocator &locator);
     };
 
 private:
@@ -52,8 +52,17 @@
     /// over and the sync thread exits.
     void ThreadSync();
 
-    /// Write the current chain block locator to the DB.
-    bool WriteBestBlock(const CBlockIndex *block_index);
+    /// Write the current index state (eg. chain block locator and
+    /// subclass-specific items) to disk.
+    ///
+    /// Recommendations for error handling:
+    /// If called on a successor of the previous committed best block in the
+    /// index, the index can continue processing without risk of corruption,
+    /// though the index state will need to catch up from further behind on
+    /// reboot. If the new state is not a successor of the previous state (due
+    /// to a chain reorganization), the index must halt until Commit succeeds or
+    /// else it could end up getting corrupted.
+    bool Commit();
 
 protected:
     void
@@ -71,6 +80,10 @@
         return true;
     }
 
+    /// Virtual method called internally by Commit that can be overridden to
+    /// atomically commit more index state.
+    virtual bool CommitInternal(CDBBatch &batch);
+
     virtual DB &GetDB() const = 0;
 
     /// Get the name of the index for display in logs.
diff --git a/src/index/base.cpp b/src/index/base.cpp
--- a/src/index/base.cpp
+++ b/src/index/base.cpp
@@ -41,8 +41,9 @@
     return success;
 }
 
-bool BaseIndex::DB::WriteBestBlock(const CBlockLocator &locator) {
-    return Write(DB_BEST_BLOCK, locator);
+void BaseIndex::DB::WriteBestBlock(CDBBatch &batch,
+                                   const CBlockLocator &locator) {
+    batch.Write(DB_BEST_BLOCK, locator);
 }
 
 BaseIndex::~BaseIndex() {
@@ -91,7 +92,12 @@
         int64_t last_locator_write_time = 0;
         while (true) {
             if (m_interrupt) {
-                WriteBestBlock(pindex);
+                m_best_block_index = pindex;
+                // No need to handle errors in Commit. If it fails, the error
+                // will be already be logged. The best way to recover is to
+                // continue, as index cannot be corrupted by a missed commit to
+                // disk for an advanced index state.
+                Commit();
                 return;
             }
 
@@ -99,9 +105,10 @@
                 LOCK(cs_main);
                 const CBlockIndex *pindex_next = NextSyncBlock(pindex);
                 if (!pindex_next) {
-                    WriteBestBlock(pindex);
                     m_best_block_index = pindex;
                     m_synced = true;
+                    // No need to handle errors in Commit. See rationale above.
+                    Commit();
                     break;
                 }
                 pindex = pindex_next;
@@ -116,8 +123,10 @@
 
             if (last_locator_write_time + SYNC_LOCATOR_WRITE_INTERVAL <
                 current_time) {
-                WriteBestBlock(pindex);
+                m_best_block_index = pindex;
                 last_locator_write_time = current_time;
+                // No need to handle errors in Commit. See rationale above.
+                Commit();
             }
 
             CBlock block;
@@ -141,14 +150,22 @@
     }
 }
 
-bool BaseIndex::WriteBestBlock(const CBlockIndex *block_index) {
-    LOCK(cs_main);
-    if (!GetDB().WriteBestBlock(::ChainActive().GetLocator(block_index))) {
-        return error("%s: Failed to write locator to disk", __func__);
+bool BaseIndex::Commit() {
+    CDBBatch batch(GetDB());
+    if (!CommitInternal(batch) || !GetDB().WriteBatch(batch)) {
+        return error("%s: Failed to commit latest %s state", __func__,
+                     GetName());
     }
     return true;
 }
 
+bool BaseIndex::CommitInternal(CDBBatch &batch) {
+    LOCK(cs_main);
+    GetDB().WriteBestBlock(batch,
+                           ::ChainActive().GetLocator(m_best_block_index));
+    return true;
+}
+
 void BaseIndex::BlockConnected(
     const std::shared_ptr<const CBlock> &block, const CBlockIndex *pindex,
     const std::vector<CTransactionRef> &txn_conflicted) {
@@ -225,9 +242,11 @@
         return;
     }
 
-    if (!GetDB().WriteBestBlock(locator)) {
-        error("%s: Failed to write locator to disk", __func__);
-    }
+    // No need to handle errors in Commit. If it fails, the error will be
+    // already be logged. The best way to recover is to continue, as index
+    // cannot be corrupted by a missed commit to disk for an advanced index
+    // state.
+    Commit();
 }
 
 bool BaseIndex::BlockUntilSyncedToCurrentChain() {