LuciferYang commented on code in PR #39226:
URL: https://github.com/apache/spark/pull/39226#discussion_r1060249072


##########
core/src/main/scala/org/apache/spark/status/AppStatusStore.scala:
##########
@@ -733,6 +734,15 @@ private[spark] class AppStatusStore(
 
   def close(): Unit = {
     store.close()
+    cleanUpStorePath()
+  }
+
+  private def cleanUpStorePath(): Unit = {
+    storePath.foreach { p =>
+      if (p.exists()) {
+        p.listFiles().foreach(Utils.deleteRecursively)
+      }
+    }

Review Comment:
   The current code uses a strategy similar to `DiskBlockManager` to generate a 
temporary directory for RocksDB (without exposing `spark.ui.store.path` to 
users).
   
   However, I think it is reasonable to expose `spark.ui.store.path` to users: 
when multiple long running Spark Apps(client mode) are running on same machine  
and RockDB is used to store the Live UI, the default use of `/tmp`  to store 
data may cause `no space`, because `/tmp` directory is generally limited (about 
5~7G space)
   
   WDYT @gengliangwang @mridulm ? If `spark.ui.store.path` needs to be exposed 
to users, I need to revert the last commit.
   
   



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to