Zakelly commented on code in PR #26415:
URL: https://github.com/apache/flink/pull/26415#discussion_r2032354803
##########
flink-state-backends/flink-statebackend-forst/src/main/java/org/apache/flink/state/forst/fs/cache/CachedDataInputStream.java:
##########
@@ -102,62 +102,81 @@ private FSDataInputStream getStream() throws IOException {
if (isFlinkThread()) {
cacheEntry.touch();
}
- FSDataInputStream stream = tryGetCacheStream();
- if (stream != null) {
- fileBasedCache.incHitCounter();
- return stream;
- }
-
- if (streamStatus == StreamStatus.CACHED_CLOSED
- || streamStatus == StreamStatus.CACHED_CLOSING) {
- if (streamStatus == StreamStatus.CACHED_CLOSING) {
- try {
- semaphore.acquire(1);
- } catch (InterruptedException e) {
- throw new RuntimeException(e);
- }
- originalStream.seek(position);
- position = -1;
- LOG.trace(
- "Stream {} status from {} to {}",
- cacheEntry.cachePath,
- streamStatus,
- StreamStatus.CACHED_CLOSED);
- streamStatus = StreamStatus.CACHED_CLOSED;
- }
- // try reopen
- tryReopen();
- stream = tryGetCacheStream();
+ // Repeat until get a stream
+ while (true) {
+ // Firstly, we try to get cache stream
+ FSDataInputStream stream = tryGetCacheStream();
if (stream != null) {
fileBasedCache.incHitCounter();
return stream;
}
- fileBasedCache.incMissCounter();
- return originalStream;
- } else if (streamStatus == StreamStatus.ORIGINAL) {
- fileBasedCache.incMissCounter();
- return originalStream;
- } else {
- if (streamStatus == StreamStatus.CACHED_OPEN) {
- stream = tryGetCacheStream();
+
+ // No cache stream, so is it closing?
+ if (streamStatus == StreamStatus.CACHED_CLOSED
+ || streamStatus == StreamStatus.CACHED_CLOSING) {
+ if (streamStatus == StreamStatus.CACHED_CLOSING) {
+ // if closing, update the position
+ try {
+ semaphore.acquire(1);
+ } catch (InterruptedException e) {
+ throw new RuntimeException(e);
+ }
+ originalStream.seek(position);
+ position = -1;
+ LOG.trace(
+ "Stream {} status from {} to {}",
+ cacheEntry.cachePath,
+ streamStatus,
+ StreamStatus.CACHED_CLOSED);
+ streamStatus = StreamStatus.CACHED_CLOSED;
+ }
+ // try reopen
+ stream = tryReopen();
if (stream != null) {
fileBasedCache.incHitCounter();
return stream;
}
+ fileBasedCache.incMissCounter();
Review Comment:
I suggest not. The `ORIGINAL` means this stream could only be read via
original stream, no cache mechanism enabled. It's a stablized status without
transition. There is an `else if` branch under this to handle the `ORIGINAL`
condition
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscr...@flink.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
