You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@kudu.apache.org by al...@apache.org on 2021/02/16 19:19:22 UTC

[kudu] branch master updated (a1ea45e -> 0bda51a)

This is an automated email from the ASF dual-hosted git repository.

alexey pushed a change to branch master
in repository https://gitbox.apache.org/repos/asf/kudu.git.


    from a1ea45e  [txn_commit-itest] use hash partitioned table for tests
     new 67018be  [test] fix flake in TestConcurrentRepeatedCommitCalls
     new 0bda51a  [tests] small cleanup on TxnCommitITest

The 2 revisions listed above as "new" are entirely new to this
repository and will be described in separate emails.  The revisions
listed as "add" were already present in the repository and have only
been added to this reference.


Summary of changes:
 src/kudu/integration-tests/txn_commit-itest.cc | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)


[kudu] 02/02: [tests] small cleanup on TxnCommitITest

Posted by al...@apache.org.
This is an automated email from the ASF dual-hosted git repository.

alexey pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/kudu.git

commit 0bda51a9b851f0692f757d743548a823f9f1888b
Author: Alexey Serbin <al...@apache.org>
AuthorDate: Mon Feb 15 23:39:33 2021 -0800

    [tests] small cleanup on TxnCommitITest
    
    * removed unused variable in the TestRestartingWhileCommitting scenario
    * use non-empty transaction (i.e. insert several rows) for the
      TestRestartingWhileCommittingAndDeleting scenario
    
    The latter is important to keep the scenario successfully running
    when txn participants are registered automatically by tablet servers:
    that functionality comes in a follow-up patch.
    
    Change-Id: I47668ef244598edbdd7fcaef88f350f36deabede
    Reviewed-on: http://gerrit.cloudera.org:8080/17070
    Tested-by: Kudu Jenkins
    Reviewed-by: Andrew Wong <aw...@cloudera.com>
---
 src/kudu/integration-tests/txn_commit-itest.cc | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/src/kudu/integration-tests/txn_commit-itest.cc b/src/kudu/integration-tests/txn_commit-itest.cc
index 920d47e..c43fc1d 100644
--- a/src/kudu/integration-tests/txn_commit-itest.cc
+++ b/src/kudu/integration-tests/txn_commit-itest.cc
@@ -515,8 +515,6 @@ TEST_F(TxnCommitITest, TestRestartingWhileCommitting) {
 // assigned commit timestamps across participants.
 TEST_F(TxnCommitITest, TestRestartingWhileCommittingAndDeleting) {
   // First, create another table that we'll delete later on.
-  unordered_set<string> first_table_tablet_ids(
-      participant_ids_.begin(), participant_ids_.end());
   const string kSecondTableName = "default.second_table";
   TestWorkload w(cluster_.get());
   w.set_num_replicas(1);
@@ -540,6 +538,7 @@ TEST_F(TxnCommitITest, TestRestartingWhileCommittingAndDeleting) {
   shared_ptr<KuduTransaction> txn;
   shared_ptr<KuduSession> txn_session;
   ASSERT_OK(BeginTransaction(both_tables_participant_ids, &txn, &txn_session));
+  ASSERT_OK(InsertToSession(txn_session, initial_row_count_, kNumRowsPerTxn));
   FLAGS_txn_status_manager_inject_latency_finalize_commit_ms = 2000;
   ASSERT_OK(txn->Commit(/*wait*/false));
 


[kudu] 01/02: [test] fix flake in TestConcurrentRepeatedCommitCalls

Posted by al...@apache.org.
This is an automated email from the ASF dual-hosted git repository.

alexey pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/kudu.git

commit 67018be8ba27480b050c11504df8a732f6a52daf
Author: Alexey Serbin <al...@apache.org>
AuthorDate: Mon Feb 15 23:27:15 2021 -0800

    [test] fix flake in TestConcurrentRepeatedCommitCalls
    
    Before this patch, the TestConcurrentRepeatedCommitCalls scenario
    of the TxnCommitITest sometimes failed with an error like below:
    
      Bad status: Illegal state: Failed to write to server: 206a8320b5ea4e4c80360fd3de898ab5 (127.0.93.65:36189): transaction ID 0 is not open: state: COMMITTED user: "slave"
    
    Before this patch, I saw such an error in about 1/8 of all DEBUG build
    runs if specifying --stress_cpu_threads=16.  After this patch, I saw
    none after running the scenario with the same flags hundreds of times.
    
    Change-Id: I39118b2b33503a18b77e4e4106979f141b5e5486
    Reviewed-on: http://gerrit.cloudera.org:8080/17069
    Tested-by: Alexey Serbin <as...@cloudera.com>
    Reviewed-by: Andrew Wong <aw...@cloudera.com>
---
 src/kudu/integration-tests/txn_commit-itest.cc | 1 +
 1 file changed, 1 insertion(+)

diff --git a/src/kudu/integration-tests/txn_commit-itest.cc b/src/kudu/integration-tests/txn_commit-itest.cc
index cf3bf8f..920d47e 100644
--- a/src/kudu/integration-tests/txn_commit-itest.cc
+++ b/src/kudu/integration-tests/txn_commit-itest.cc
@@ -683,6 +683,7 @@ TEST_F(TxnCommitITest, TestConcurrentCommitCalls) {
 // Test that committing the same transaction concurrently doesn't lead to any
 // issues.
 TEST_F(TxnCommitITest, TestConcurrentRepeatedCommitCalls) {
+  FLAGS_txn_status_manager_inject_latency_finalize_commit_ms = 1000;
   shared_ptr<KuduTransaction> txn;
   shared_ptr<KuduSession> txn_session;
   ASSERT_OK(BeginTransaction(participant_ids_, &txn, &txn_session));