diff --git a/src/Hash2Pub/DHTProtocol.hs b/src/Hash2Pub/DHTProtocol.hs
index f431da9..a2dd676 100644
--- a/src/Hash2Pub/DHTProtocol.hs
+++ b/src/Hash2Pub/DHTProtocol.hs
@@ -501,11 +501,15 @@ requestQueryID :: LocalNodeState        -- ^ NodeState of the querying node
 -- TODO: deal with lookup failures
 requestQueryID ns targetID = do
     firstCacheSnapshot <- readTVarIO . nodeCacheSTM $ ns
-    queryIdLookupLoop firstCacheSnapshot ns targetID
+    -- TODO: make maxAttempts configurable
+    queryIdLookupLoop firstCacheSnapshot ns 50 targetID
 
 -- | like 'requestQueryID, but allows passing of a custom cache, e.g. for joining
-queryIdLookupLoop :: NodeCache -> LocalNodeState -> NodeID -> IO RemoteNodeState
-queryIdLookupLoop cacheSnapshot ns targetID = do
+queryIdLookupLoop :: NodeCache -> LocalNodeState -> Int -> NodeID -> IO RemoteNodeState
+-- return node itself as default fallback value against infinite recursion.
+-- TODO: consider using an Either instead of a default value
+queryIdLookupLoop _ ns 0 _ = pure $ toRemoteNodeState ns
+queryIdLookupLoop cacheSnapshot ns maxAttempts targetID = do
     let localResult = queryLocalCache ns cacheSnapshot (lNumBestNodes ns) targetID
     -- FOUND can only be returned if targetID is owned by local node
     case localResult of
@@ -522,8 +526,7 @@ queryIdLookupLoop cacheSnapshot ns targetID = do
                     addCacheEntryPure now
                               ) cacheSnapshot entrySet
               in
-          -- TODO: this could lead to infinite recursion on an empty cache. Consider returning the node itself as default value
-              queryIdLookupLoop newLCache ns targetID
+              queryIdLookupLoop newLCache ns (maxAttempts - 1) targetID
 
 
 sendQueryIdMessages :: (Integral i)
diff --git a/src/Hash2Pub/FediChord.hs b/src/Hash2Pub/FediChord.hs
index 152abd9..8337b06 100644
--- a/src/Hash2Pub/FediChord.hs
+++ b/src/Hash2Pub/FediChord.hs
@@ -165,7 +165,7 @@ fediChordJoin cacheSnapshot nsSTM = do
     ns <- readTVarIO nsSTM
     -- get routed to the currently responsible node, based on the response
     -- from the bootstrapping node
-    currentlyResponsible <- queryIdLookupLoop cacheSnapshot ns $ getNid ns
+    currentlyResponsible <- queryIdLookupLoop cacheSnapshot ns 50 $ getNid ns
     -- 2. then send a join to the currently responsible node
     joinResult <- requestJoin currentlyResponsible nsSTM
     case joinResult of