You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by GitBox <gi...@apache.org> on 2020/06/10 07:59:19 UTC
[GitHub] [flink] rkhachatryan commented on a change in pull request #12478: [FLINK-17869][task][checkpointing] Fix race condition when calling ChannelStateWriter.abort
rkhachatryan commented on a change in pull request #12478:
URL: https://github.com/apache/flink/pull/12478#discussion_r437933410
##########
File path: flink-runtime/src/main/java/org/apache/flink/runtime/checkpoint/channel/ChannelStateWriterImpl.java
##########
@@ -56,7 +56,7 @@
public class ChannelStateWriterImpl implements ChannelStateWriter {
private static final Logger LOG = LoggerFactory.getLogger(ChannelStateWriterImpl.class);
- private static final int DEFAULT_MAX_CHECKPOINTS = 5; // currently, only single in-flight checkpoint is supported
+ private static final int DEFAULT_MAX_CHECKPOINTS = 1000; // includes max-concurrent-checkpoints + checkpoints to be aborted (scheduled via mailbox)
Review comment:
Me and @AHeise also had these concerns and discussed it [above](https://github.com/apache/flink/pull/12478#discussion_r436502262) (I increased the limit to 1K since then).
I'm not sure if there __is__ a proper solution because in the general case (max-concurrent-checkpoints > 1) we simply don't know which checkpoints to abort. Another issue is thread-safety, which is probably solvable but with some extra complexity which I think doesn't pay off here.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org