Andrew Wong has posted comments on this change. ( http://gerrit.cloudera.org:8080/16952 )
Change subject: KUDU-2612: background task to commit transaction ...................................................................... Patch Set 15: (8 comments) http://gerrit.cloudera.org:8080/#/c/16952/14/src/kudu/transactions/txn_status_manager.h File src/kudu/transactions/txn_status_manager.h: http://gerrit.cloudera.org:8080/#/c/16952/14/src/kudu/transactions/txn_status_manager.h@92 PS14, Line 92: BeginCommitAsyn > nit: maybe name it BeginCommitAsyncTask if it is asynchronously as well. Sa Done http://gerrit.cloudera.org:8080/#/c/16952/14/src/kudu/transactions/txn_status_manager.h@102 PS14, Line 102: commit_timestamp > nit: can you comment on how the timestamp will be used? Done http://gerrit.cloudera.org:8080/#/c/16952/14/src/kudu/transactions/txn_status_manager.h@321 PS14, Line 321: std::lock_guard<simple_spinlock> l(lock_); > nit: can you update the comment to add that we will also load commits in fl Done http://gerrit.cloudera.org:8080/#/c/16952/14/src/kudu/transactions/txn_status_manager.cc File src/kudu/transactions/txn_status_manager.cc: http://gerrit.cloudera.org:8080/#/c/16952/14/src/kudu/transactions/txn_status_manager.cc@123 PS14, Line 123: > nit: space. Done http://gerrit.cloudera.org:8080/#/c/16952/14/src/kudu/transactions/txn_status_manager.cc@215 PS14, Line 215: urn; : } > nit: remove? Done http://gerrit.cloudera.org:8080/#/c/16952/14/src/kudu/transactions/txn_status_manager.cc@257 PS14, Line 257: participant_ids_[participant_idx], : std::move(op_pb), : MonoDelta::FromMilliseconds(FLAGS_txn_background_rpc_timeout_ms), : std::move(participated_cb), : &begin_commit_timestamps_[participant_idx]); : } > nit: do you think this block of code worth writing to a method of own? As I It's not identical everywhere, but sure. http://gerrit.cloudera.org:8080/#/c/16952/14/src/kudu/transactions/txn_status_manager.cc@268 PS14, Line 268: pated_cb = [this, scoped_this = std::move(scoped_this), : participant_idx, commit_timestamp] (const Status& s) { : if (IsShuttingDownCleanupIfLastOp()) { : > So we proceed to ScheduleFinalizeCommitWrite even we encounter error here? Yeah, added a comment to make that clear that it's intentional. http://gerrit.cloudera.org:8080/#/c/16952/14/src/kudu/transactions/txn_status_manager.cc@787 PS14, Line 787: : std::lock_guar > This happens when there is no ops in the transaction? What does invalid tim There is always at least the BEGIN_COMMIT op timestamp, regardless of any write ops. Invalid timestamp is just a dummy timestamp. -- To view, visit http://gerrit.cloudera.org:8080/16952 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: kudu Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: Ie2258dded3ab3d527cb5d0abdc7d5e7deb4da15e Gerrit-Change-Number: 16952 Gerrit-PatchSet: 15 Gerrit-Owner: Andrew Wong <aw...@cloudera.com> Gerrit-Reviewer: Alexey Serbin <aser...@cloudera.com> Gerrit-Reviewer: Andrew Wong <aw...@cloudera.com> Gerrit-Reviewer: Hao Hao <hao....@cloudera.com> Gerrit-Reviewer: Kudu Jenkins (120) Gerrit-Comment-Date: Fri, 05 Feb 2021 05:28:33 +0000 Gerrit-HasComments: Yes
