Github user chtyim commented on a diff in the pull request: https://github.com/apache/incubator-tephra/pull/23#discussion_r91457499 --- Diff: tephra-core/src/main/java/org/apache/tephra/TransactionManager.java --- @@ -348,27 +348,30 @@ private void cleanupTimedOutTransactions() { } long currentTime = System.currentTimeMillis(); - List<Long> timedOut = Lists.newArrayList(); + Map<Long, InProgressType> timedOut = Maps.newHashMap(); --- End diff -- Unrelated to the change, but it seems like the locking in this method is wrong. First, we are holding the read lock for the entire cleanup method, which seems unnecessary. The inProgress itself is a concurrent tree, so for the first `for-loop`, it shouldn't need the lock at all, as it is just for collecting the in progress tx that are timed out. Then in the `if (!timedOut.isEmpty())`, seems like only the modification to `invalid` and `invalidArray` needs to be protected by lock, but I think it should be a "write lock"?
--- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---