reuvenlax commented on code in PR #38264:
URL: https://github.com/apache/beam/pull/38264#discussion_r3155980812
##########
sdks/java/io/google-cloud-platform/src/main/java/org/apache/beam/sdk/io/gcp/bigquery/StorageApiWritesShardedRecords.java:
##########
@@ -282,7 +242,6 @@ static class AppendRowsContext<DestinationT>
extends RetryManager.Operation.Context<AppendRowsResponse> {
final ShardedKey<DestinationT> key;
String streamName = "";
- @Nullable StreamAppendClient client = null;
Review Comment:
There's a comment in front of the try-with-resources block explaining this.
Basically pins are ref counts. Since all calls happen in the context of
processElements, all we really need is one pin for the entire function - there
was never a need for per-RPC pins which was hard to follow. Instead we add one
pin for the duration of the function, managed via a try-with-resources block.
Since the cache holds onto objects longer, the cache also manages one
additional pin per object (though that can be invalidated at any time)
--
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: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
