issue-628 Handled failures which might occur during write-behind#3127
issue-628 Handled failures which might occur during write-behind#3127chrisdennis merged 1 commit intoehcache:masterfrom
Conversation
|
Failing tests need investigated |
chrisdennis
left a comment
There was a problem hiding this comment.
This needs test coverage to prove it all works. Doing that should also expose some of the existing issues that need fixing.
...ava/org/ehcache/clustered/client/internal/loaderwriter/writebehind/ClusteredWriteBehind.java
Outdated
Show resolved
Hide resolved
...org/ehcache/clustered/client/internal/loaderwriter/writebehind/ClusteredWriteBehindTest.java
Outdated
Show resolved
Hide resolved
ehcache-api/src/main/java/org/ehcache/spi/loaderwriter/CacheLoaderWriter.java
Outdated
Show resolved
Hide resolved
...va/org/ehcache/impl/internal/loaderwriter/writebehind/BatchingLocalHeapWriteBehindQueue.java
Outdated
Show resolved
Hide resolved
...va/org/ehcache/impl/internal/loaderwriter/writebehind/BatchingLocalHeapWriteBehindQueue.java
Outdated
Show resolved
Hide resolved
...org/ehcache/impl/internal/loaderwriter/writebehind/NonBatchingLocalHeapWriteBehindQueue.java
Outdated
Show resolved
Hide resolved
...org/ehcache/impl/internal/loaderwriter/writebehind/NonBatchingLocalHeapWriteBehindQueue.java
Show resolved
Hide resolved
.../main/java/org/ehcache/impl/internal/loaderwriter/writebehind/operations/WriteOperation.java
Outdated
Show resolved
Hide resolved
.../java/org/ehcache/impl/internal/loaderwriter/writebehind/WriteBehindProviderFactoryTest.java
Outdated
Show resolved
Hide resolved
chrisdennis
left a comment
There was a problem hiding this comment.
We still need test coverage here to show that the correct clean up is happening.
...va/org/ehcache/impl/internal/loaderwriter/writebehind/BatchingLocalHeapWriteBehindQueue.java
Outdated
Show resolved
Hide resolved
...va/org/ehcache/impl/internal/loaderwriter/writebehind/BatchingLocalHeapWriteBehindQueue.java
Outdated
Show resolved
Hide resolved
ehcache-api/src/main/java/org/ehcache/spi/loaderwriter/CacheLoaderWriter.java
Outdated
Show resolved
Hide resolved
...n/java/org/ehcache/impl/internal/loaderwriter/writebehind/operations/DeleteAllOperation.java
Outdated
Show resolved
Hide resolved
|
Since dead lock was happening in earlier solution for batching write-behind, changed to place to run the cleanup for failures. |
chrisdennis
left a comment
There was a problem hiding this comment.
The tests here are still failing, but they do it sporadically. You should go look at the AbstractWriteBehind class and the implications of the implementation of that class and the order in which the cache and the latest map held in both of the concrete write behind implementation are cleaned up.
...org/ehcache/impl/internal/loaderwriter/writebehind/NonBatchingLocalHeapWriteBehindQueue.java
Outdated
Show resolved
Hide resolved
...va/org/ehcache/impl/internal/loaderwriter/writebehind/BatchingLocalHeapWriteBehindQueue.java
Outdated
Show resolved
Hide resolved
ehcache-api/src/main/java/org/ehcache/spi/loaderwriter/CacheLoaderWriter.java
Outdated
Show resolved
Hide resolved
chrisdennis
left a comment
There was a problem hiding this comment.
This needs rebasing against latest HEAD of main and the subsequent commits should probably be squashed to form a more compact and logical commit stream (potentially down to a single commit).
…ng write-behind for issue#628
4f83bcd to
6f38cbd
Compare
|
Re-based with master |
Handle the failure which may occur during following write-behind scenarios and removed failed entries from cache.