You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Bug the first: `S3BlobContainer#delete` spuriously attempts to delete a
blob named after the container, as if it were a directory in a
filesystem. This makes sense for filesystem repositories but in S3 the
blob key is an opaque string, may be a prefix of other blob keys, and
can legitimately end in a `/`. We never create such a blob in the first
place, but this was hidden because AWS S3 silently ignores these
deletion requests, and also because...
Bug the second: `S3HttpHandler` would delete extant blobs but ignore
nonexistent blobs when processing a multi-object delete request. This is
apparently how S3 behaves in practice but it's not documented as such so
we cannot rely on it and must be stricter in our tests. Fixing this
exposed...
Bug the third: `S3BlobContainer#deleteBlobsIgnoringIfNotExists` wasn't
actually ignoring `NoSuchKey` errors should any arise, and the S3
reference documentation does not proscribe this behaviour, so we must
handle it properly.
Copy file name to clipboardExpand all lines: test/framework/src/main/java/org/elasticsearch/repositories/blobstore/ESBlobStoreRepositoryIntegTestCase.java
+1-2Lines changed: 1 addition & 2 deletions
Original file line number
Diff line number
Diff line change
@@ -230,8 +230,7 @@ public void testDeleteBlobs() throws IOException {
0 commit comments