Fix incorrect validity checks in getInfo() and newDerivativeImageInpu…

…tStream()
cantaloupe-project · May 3, 2021 · 59897e3 · 59897e3
1 parent 000c28c
commit 59897e3
Show file tree

Hide file tree

Showing 2 changed files with 29 additions and 2 deletions.
diff --git a/CHANGES.md b/CHANGES.md
@@ -13,6 +13,7 @@
 * Fixed a floating-point rounding bug that could cause an excessively large
   TIFF pyramid level to be selected, resulting in unnecessary scaling.
 * Fixed a bug in TurboJpegProcessor that could cause incorrect cropping.
+* Fixed a bug in S3Cache that could cause hits against expired content.
 * Fixed a bug that could cause S3 credentials from the configuration file to be
   erroneously used even when not set.
 * Fixed a bug that could cause corrupt image to be written to a derivative

diff --git a/src/main/java/edu/illinois/library/cantaloupe/cache/S3Cache.java b/src/main/java/edu/illinois/library/cantaloupe/cache/S3Cache.java
@@ -279,13 +279,15 @@ public Optional<Info> getInfo(Identifier identifier) throws IOException {
                 .build();
         final Stopwatch watch = new Stopwatch();
         try (ResponseInputStream<GetObjectResponse> is = client.getObject(request)) {
-            if (is != null) {
+            // This extra validity check may be needed with minio server
+            if (is != null && is.response().lastModified().isAfter(earliestValidInstant())) {
                 final Info info = Info.fromJSON(is);
                 LOGGER.debug("getInfo(): read {} from bucket {} in {}",
                         objectKey, bucketName, watch);
                 touchAsync(objectKey);
                 return Optional.of(info);
             } else {
+                consumeStreamAsync(is);
                 LOGGER.debug("{} in bucket {} is invalid; purging asynchronously",
                         objectKey, bucketName);
                 purgeAsync(bucketName, objectKey);
@@ -315,10 +317,12 @@ public InputStream newDerivativeImageInputStream(OperationList opList)
                 .build();
         try {
             ResponseInputStream<GetObjectResponse> is = client.getObject(request);
-            if (is != null) {
+            // This extra validity check may be needed with minio server
+            if (is != null && is.response().lastModified().isAfter(earliestValidInstant())) {
                 touchAsync(objectKey);
                 return is;
             } else {
+                consumeStreamAsync(is);
                 LOGGER.debug("{} in bucket {} is invalid; purging asynchronously",
                         objectKey, bucketName);
                 purgeAsync(bucketName, objectKey);
@@ -501,6 +505,28 @@ public void put(Identifier identifier, Info info) throws IOException {
         }
     }
 
+    /**
+     * The AWS client logs a warning when we close an InputStream without fully
+     * reading it. This method does that and then closes the stream.
+     */
+    private void consumeStreamAsync(InputStream inputStream) {
+        ThreadPool.getInstance().submit(() -> {
+            try {
+                inputStream.readAllBytes();
+            } catch (IOException e) {
+                LOGGER.warn("consumeStreamAsync(): failed to consume the stream: {}",
+                        e.getMessage());
+            } finally {
+                try {
+                    inputStream.close();
+                } catch (IOException e) {
+                    LOGGER.warn("consumeStreamAsync(): failed to close the stream: {}",
+                            e.getMessage());
+                }
+            }
+        });
+    }
+
     /**
      * Updates an object's "last-accessed time." Since S3 doesn't support a
      * last-accessed time and S3 objects are immutable, this method copies the