Github user andrewor14 commented on a diff in the pull request:
https://github.com/apache/spark/pull/11546#discussion_r62278386
--- Diff: core/src/main/scala/org/apache/spark/storage/BlockManager.scala
---
@@ -1108,11 +1118,14 @@ private[spark] class BlockManager(
def removeBlock(blockId: BlockId, tellMaster: Boolean = true): Unit = {
logDebug(s"Removing block $blockId")
val info = blockInfo.get(blockId).orNull
- if (info != null) {
+ if (info != null && !pendingToRemove.containsKey(blockId)) {
+ pendingToRemove.put(blockId, currentTaskAttemptId)
info.synchronized {
+ val level = info.level
// Removals are idempotent in disk store and memory store. At
worst, we get a warning.
- val removedFromMemory = memoryStore.remove(blockId)
- val removedFromDisk = diskStore.remove(blockId)
+ val removedFromMemory = if (level.useMemory)
memoryStore.remove(blockId) else false
+ pendingToRemove.remove(blockId)
--- End diff --
we should wrap this in a try finally in case an exception is thrown.
Otherwise we may never release the "pending" lock:
```
if (info != null && !pendingToRemove.containsKey(blockId)) {
pendingToRemove.put(blockId, currentTaskAttemptId)
try {
info.synchronized {
... // remove the block from memoryStore
}
} finally {
pendingToRemove.remove(blockId)
}
}
```
Here and a few other places.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
