Github user mridulm commented on a diff in the pull request:
https://github.com/apache/spark/pull/604#discussion_r12257245
--- Diff:
core/src/main/scala/org/apache/spark/storage/DiskBlockManager.scala ---
@@ -71,13 +71,13 @@ private[spark] class DiskBlockManager(shuffleManager:
ShuffleBlockManager, rootD
// Create the subdirectory if it doesn't already exist
var subDir = subDirs(dirId)(subDirId)
- if (subDir == null) {
+ val newDir = new File(localDirs(dirId), "%02x".format(subDirId))
+ if (subDir == null || !newDir.exists) {
subDir = subDirs(dirId).synchronized {
val old = subDirs(dirId)(subDirId)
- if (old != null) {
+ if (old != null && newDir.exists) {
old
} else {
- val newDir = new File(localDirs(dirId), "%02x".format(subDirId))
newDir.mkdir()
--- End diff --
This entire codepath is buggy for other reasons - which I should be
resolving in a different PR anyway.
Btw, this change does not fix things - you are just going to face DCL
issues which making getFile very expensive (and it is a common operation).
---
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 infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
