[kudu-CR] exactly once writes-itest for faulty disks

[David Ribeiro Alves (Code Review)]

David Ribeiro Alves has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/8388 )

Change subject: exactly_once_writes-itest for faulty disks
..


Patch Set 8:

(7 comments)

http://gerrit.cloudera.org:8080/#/c/8388/8/src/kudu/integration-tests/exactly_once_writes-itest.cc
File src/kudu/integration-tests/exactly_once_writes-itest.cc:

http://gerrit.cloudera.org:8080/#/c/8388/8/src/kudu/integration-tests/exactly_once_writes-itest.cc@77
PS8, Line 77: using cluster::ExternalTabletServer;
nit: move before the two above


http://gerrit.cloudera.org:8080/#/c/8388/8/src/kudu/integration-tests/exactly_once_writes-itest.cc@83
PS8, Line 83: int64_t
xs nit: why is this the only int64_t ? doesn't seem like it would be feasible 
to have over 2B attempts


http://gerrit.cloudera.org:8080/#/c/8388/8/src/kudu/integration-tests/exactly_once_writes-itest.cc@110
PS8, Line 110: If 'allow_crashes' is set, crashed servers are restarted.
maybe add that if 'allow_crashes' is false the test is caused to fail (that is 
what happens, right?)


http://gerrit.cloudera.org:8080/#/c/8388/8/src/kudu/integration-tests/exactly_once_writes-itest.cc@118
PS8, Line 118: matching commit indexes.
not sure what this means...


http://gerrit.cloudera.org:8080/#/c/8388/8/src/kudu/integration-tests/exactly_once_writes-itest.cc@234
PS8, Line 234: {
 :   std::lock_guard l(lock);
 :   if (num_complete == num_threads) break;
 : }
 : if (allow_crashes) {
 :   RETURN_IF_FATALS(RestartAnyCrashedTabletServers(),
 :   "Failed to restart crashed tablet servers");
 : } else {
 :   RETURN_IF_FATALS(AssertNoTabletServersCrashed(),
 :   "Some tablet servers crashed");
 : }
 : SleepFor(MonoDelta::FromMilliseconds(10));
nit: might be slightly cleaner to use a countdown latch (avoid the extra 
synchronization around the shared var and has a timeout function that replaces 
the sleep)


http://gerrit.cloudera.org:8080/#/c/8388/8/src/kudu/integration-tests/exactly_once_writes-itest.cc@303
PS8, Line 303: LOG(INFO)
debug? same below


http://gerrit.cloudera.org:8080/#/c/8388/8/src/kudu/util/test_macros.h
File src/kudu/util/test_macros.h:

http://gerrit.cloudera.org:8080/#/c/8388/8/src/kudu/util/test_macros.h@27
PS8, Line 27: // Sometimes it's useful to return a status, so we're not forced 
to write void
: // functions to check for fatals
explain this better, can't figure it out just from reading this without much 
more context



--
To view, visit http://gerrit.cloudera.org:8080/8388
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: kudu
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I966b41b71e39827f8f78fcca59a2208f13f71e53
Gerrit-Change-Number: 8388
Gerrit-PatchSet: 8
Gerrit-Owner: Andrew Wong 
Gerrit-Reviewer: Andrew Wong 
Gerrit-Reviewer: David Ribeiro Alves 
Gerrit-Reviewer: Kudu Jenkins
Gerrit-Reviewer: Tidy Bot
Gerrit-Comment-Date: Tue, 24 Jul 2018 21:14:31 +
Gerrit-HasComments: Yes

[kudu-CR] exactly once writes-itest for faulty disks

[Andrew Wong (Code Review)]

Andrew Wong has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/8388 )

Change subject: exactly_once_writes-itest for faulty disks
..


Patch Set 8: Verified+1

Unrelated flake


--
To view, visit http://gerrit.cloudera.org:8080/8388
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: kudu
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I966b41b71e39827f8f78fcca59a2208f13f71e53
Gerrit-Change-Number: 8388
Gerrit-PatchSet: 8
Gerrit-Owner: Andrew Wong 
Gerrit-Reviewer: Andrew Wong 
Gerrit-Reviewer: David Ribeiro Alves 
Gerrit-Reviewer: Kudu Jenkins
Gerrit-Reviewer: Tidy Bot
Gerrit-Comment-Date: Tue, 24 Jul 2018 04:52:06 +
Gerrit-HasComments: No

[kudu-CR] exactly once writes-itest for faulty disks

[Andrew Wong (Code Review)]

Andrew Wong has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/8388 )

Change subject: exactly_once_writes-itest for faulty disks
..


Patch Set 8:

It still runs and isn't too flaky (ran a thousand runs in release and TSAN)! 
Just changed some minor things


--
To view, visit http://gerrit.cloudera.org:8080/8388
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: kudu
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I966b41b71e39827f8f78fcca59a2208f13f71e53
Gerrit-Change-Number: 8388
Gerrit-PatchSet: 8
Gerrit-Owner: Andrew Wong 
Gerrit-Reviewer: Andrew Wong 
Gerrit-Reviewer: David Ribeiro Alves 
Gerrit-Reviewer: Kudu Jenkins
Gerrit-Reviewer: Tidy Bot
Gerrit-Comment-Date: Tue, 24 Jul 2018 03:28:57 +
Gerrit-HasComments: No

[kudu-CR] exactly once writes-itest for faulty disks

[Andrew Wong (Code Review)]

Hello Tidy Bot, David Ribeiro Alves, Kudu Jenkins,

I'd like you to reexamine a change. Please visit

http://gerrit.cloudera.org:8080/8388

to look at the new patch set (#8).

Change subject: exactly_once_writes-itest for faulty disks
..

exactly_once_writes-itest for faulty disks

This patch adds an ExactlyOnceWrites test that exercises intermittently
crashy disks. This serves to not only test disk failures during writes,
but also test disk failures during tablet copies, all while maintaining
exactly-once semantics.

exactly_once_writes-itest is also updated to use std::thread instead of
kudu::Thread, and to favor returning Status instead of void.

Change-Id: I966b41b71e39827f8f78fcca59a2208f13f71e53
---
M src/kudu/integration-tests/exactly_once_writes-itest.cc
M src/kudu/util/test_macros.h
2 files changed, 241 insertions(+), 121 deletions(-)


  git pull ssh://gerrit.cloudera.org:29418/kudu refs/changes/88/8388/8
--
To view, visit http://gerrit.cloudera.org:8080/8388
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: kudu
Gerrit-Branch: master
Gerrit-MessageType: newpatchset
Gerrit-Change-Id: I966b41b71e39827f8f78fcca59a2208f13f71e53
Gerrit-Change-Number: 8388
Gerrit-PatchSet: 8
Gerrit-Owner: Andrew Wong 
Gerrit-Reviewer: Andrew Wong 
Gerrit-Reviewer: David Ribeiro Alves 
Gerrit-Reviewer: Kudu Jenkins
Gerrit-Reviewer: Tidy Bot

[kudu-CR] exactly once writes-itest for faulty disks

[David Ribeiro Alves (Code Review)]

David Ribeiro Alves has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/8388 )

Change subject: exactly_once_writes-itest for faulty disks
..


Patch Set 7:

this still worth merging?


--
To view, visit http://gerrit.cloudera.org:8080/8388
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: kudu
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I966b41b71e39827f8f78fcca59a2208f13f71e53
Gerrit-Change-Number: 8388
Gerrit-PatchSet: 7
Gerrit-Owner: Andrew Wong 
Gerrit-Reviewer: Andrew Wong 
Gerrit-Reviewer: David Ribeiro Alves 
Gerrit-Reviewer: Kudu Jenkins
Gerrit-Reviewer: Tidy Bot
Gerrit-Comment-Date: Mon, 23 Jul 2018 22:46:18 +
Gerrit-HasComments: No

[kudu-CR] exactly once writes-itest for faulty disks

[Andrew Wong (Code Review)]

Andrew Wong has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/8388 )

Change subject: exactly_once_writes-itest for faulty disks
..


Patch Set 7:

Retriggered. Failed due to clock sync issues.


--
To view, visit http://gerrit.cloudera.org:8080/8388
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: kudu
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I966b41b71e39827f8f78fcca59a2208f13f71e53
Gerrit-Change-Number: 8388
Gerrit-PatchSet: 7
Gerrit-Owner: Andrew Wong 
Gerrit-Reviewer: Andrew Wong 
Gerrit-Reviewer: David Ribeiro Alves 
Gerrit-Reviewer: Kudu Jenkins
Gerrit-Reviewer: Tidy Bot
Gerrit-Comment-Date: Thu, 30 Nov 2017 00:27:39 +
Gerrit-HasComments: No

[kudu-CR] exactly once writes-itest for faulty disks

[Andrew Wong (Code Review)]

Hello Tidy Bot, David Ribeiro Alves, Kudu Jenkins,

I'd like you to reexamine a change. Please visit

http://gerrit.cloudera.org:8080/8388

to look at the new patch set (#7).

Change subject: exactly_once_writes-itest for faulty disks
..

exactly_once_writes-itest for faulty disks

This patch adds an ExactlyOnceWrites test that exercises intermittently
crashy disks. This serves to not only test disk failures during writes,
but also test disk failures during tablet copies, all while maintaining
exactly-once semantics.

exactly_once_writes-itest is also updated to use std::thread instead of
kudu::Thread, and to favor returning Status instead of void.

Change-Id: I966b41b71e39827f8f78fcca59a2208f13f71e53
---
M src/kudu/integration-tests/exactly_once_writes-itest.cc
M src/kudu/util/test_macros.h
2 files changed, 229 insertions(+), 115 deletions(-)


  git pull ssh://gerrit.cloudera.org:29418/kudu refs/changes/88/8388/7
--
To view, visit http://gerrit.cloudera.org:8080/8388
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: kudu
Gerrit-Branch: master
Gerrit-MessageType: newpatchset
Gerrit-Change-Id: I966b41b71e39827f8f78fcca59a2208f13f71e53
Gerrit-Change-Number: 8388
Gerrit-PatchSet: 7
Gerrit-Owner: Andrew Wong 
Gerrit-Reviewer: Andrew Wong 
Gerrit-Reviewer: David Ribeiro Alves 
Gerrit-Reviewer: Kudu Jenkins
Gerrit-Reviewer: Tidy Bot

[kudu-CR] exactly once writes-itest for faulty disks

[Andrew Wong (Code Review)]

Andrew Wong has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/8388 )

Change subject: exactly_once_writes-itest for faulty disks
..


Patch Set 6:

The build failed due to a timeout (it just took too long), so I adjusted the 
numbers a bit to speed it up.


--
To view, visit http://gerrit.cloudera.org:8080/8388
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: kudu
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I966b41b71e39827f8f78fcca59a2208f13f71e53
Gerrit-Change-Number: 8388
Gerrit-PatchSet: 6
Gerrit-Owner: Andrew Wong 
Gerrit-Reviewer: Andrew Wong 
Gerrit-Reviewer: David Ribeiro Alves 
Gerrit-Reviewer: Kudu Jenkins
Gerrit-Reviewer: Tidy Bot
Gerrit-Comment-Date: Wed, 29 Nov 2017 04:01:35 +
Gerrit-HasComments: No

[kudu-CR] exactly once writes-itest for faulty disks

[Andrew Wong (Code Review)]

Hello Tidy Bot, David Ribeiro Alves, Kudu Jenkins,

I'd like you to reexamine a change. Please visit

http://gerrit.cloudera.org:8080/8388

to look at the new patch set (#6).

Change subject: exactly_once_writes-itest for faulty disks
..

exactly_once_writes-itest for faulty disks

This patch adds an ExactlyOnceWrites test that exercises intermittently
crashy disks. This serves to not only test disk failures during writes,
but also test disk failures during tablet copies, all while maintaining
exactly-once semantics.

exactly_once_writes-itest is also updated to use std::thread instead of
kudu::Thread, and to favor returning Status instead of void.

Change-Id: I966b41b71e39827f8f78fcca59a2208f13f71e53
---
M src/kudu/integration-tests/exactly_once_writes-itest.cc
M src/kudu/util/test_macros.h
2 files changed, 228 insertions(+), 115 deletions(-)


  git pull ssh://gerrit.cloudera.org:29418/kudu refs/changes/88/8388/6
--
To view, visit http://gerrit.cloudera.org:8080/8388
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: kudu
Gerrit-Branch: master
Gerrit-MessageType: newpatchset
Gerrit-Change-Id: I966b41b71e39827f8f78fcca59a2208f13f71e53
Gerrit-Change-Number: 8388
Gerrit-PatchSet: 6
Gerrit-Owner: Andrew Wong 
Gerrit-Reviewer: Andrew Wong 
Gerrit-Reviewer: David Ribeiro Alves 
Gerrit-Reviewer: Kudu Jenkins
Gerrit-Reviewer: Tidy Bot

[kudu-CR] exactly once writes-itest for faulty disks

[Andrew Wong (Code Review)]

Hello Tidy Bot, David Ribeiro Alves, Kudu Jenkins,

I'd like you to reexamine a change. Please visit

http://gerrit.cloudera.org:8080/8388

to look at the new patch set (#5).

Change subject: exactly_once_writes-itest for faulty disks
..

exactly_once_writes-itest for faulty disks

This patch adds an ExactlyOnceWrites test that exercises intermittently
crashy disks. This serves to not only test disk failures during writes,
but also test disk failures during tablet copies, all while maintaining
exactly-once semantics.

exactly_once_writes-itest is also updated to use std::thread instead of
kudu::Thread, and to favor returning Status instead of void.

Change-Id: I966b41b71e39827f8f78fcca59a2208f13f71e53
---
M src/kudu/integration-tests/exactly_once_writes-itest.cc
M src/kudu/tablet/transactions/transaction_driver.cc
M src/kudu/util/test_macros.h
3 files changed, 233 insertions(+), 117 deletions(-)


  git pull ssh://gerrit.cloudera.org:29418/kudu refs/changes/88/8388/5
--
To view, visit http://gerrit.cloudera.org:8080/8388
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: kudu
Gerrit-Branch: master
Gerrit-MessageType: newpatchset
Gerrit-Change-Id: I966b41b71e39827f8f78fcca59a2208f13f71e53
Gerrit-Change-Number: 8388
Gerrit-PatchSet: 5
Gerrit-Owner: Andrew Wong 
Gerrit-Reviewer: Andrew Wong 
Gerrit-Reviewer: David Ribeiro Alves 
Gerrit-Reviewer: Kudu Jenkins
Gerrit-Reviewer: Tidy Bot

[kudu-CR] exactly once writes-itest for faulty disks

[Andrew Wong (Code Review)]

Andrew Wong has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/8388 )

Change subject: exactly_once_writes-itest for faulty disks
..


Patch Set 4:

(7 comments)

http://gerrit.cloudera.org:8080/#/c/8388/3/src/kudu/integration-tests/exactly_once_writes-itest.cc
File src/kudu/integration-tests/exactly_once_writes-itest.cc:

http://gerrit.cloudera.org:8080/#/c/8388/3/src/kudu/integration-tests/exactly_once_writes-itest.cc@115
PS3, Line 115:   Status DoTestWritesWithExactlyOnceSemantics(int num_batches,
 :   bool 
allow_crashes);
 :
 :   // Validates that the servers all have matchin
> docs
Done


http://gerrit.cloudera.org:8080/#/c/8388/3/src/kudu/integration-tests/exactly_once_writes-itest.cc@128
PS3, Line 128: };
> warning: redundant get() call on smart pointer [google-readability-redundan
Done


http://gerrit.cloudera.org:8080/#/c/8388/3/src/kudu/integration-tests/exactly_once_writes-itest.cc@247
PS3, Line 247:   for (auto& thread : threads) {
> warning: redundant get() call on smart pointer [google-readability-redundan
Done


http://gerrit.cloudera.org:8080/#/c/8388/3/src/kudu/integration-tests/exactly_once_writes-itest.cc@264
PS3, Line 264:   // Make sure we received the same responses for the same 
operations across all
> warning: redundant get() call on smart pointer [google-readability-redundan
Done


http://gerrit.cloudera.org:8080/#/c/8388/3/src/kudu/integration-tests/exactly_once_writes-itest.cc@269
PS3, Line 269:   string expected_ts = Substitute("T:$0 TSidx:$1 TSuuid:$2", 
j, j % FLAGS_num_replicas,
> warning: redundant get() call on smart pointer [google-readability-redundan
Done


http://gerrit.cloudera.org:8080/#/c/8388/3/src/kudu/integration-tests/exactly_once_writes-itest.cc@342
PS3, Line 342:
> nit: TestWritesWithExactlyOnceSemanticsWithCrashyNodes
Done


http://gerrit.cloudera.org:8080/#/c/8388/3/src/kudu/integration-tests/exactly_once_writes-itest.cc@382
PS3, Line 382:   NO_FATALS(CreateClient(_));
 :   NO_FATALS(CreateTable());
 :   WaitForTSAndReplicas();
 :   ASSERT_FALSE(tablet_replicas_.empty());
 :   tablet_id_ = (*tablet_replicas_.begin()).first;
 :
 :   // Fail under half of the servers so we always maintain a 
healthy majority.
 :   for (int i = 0; i < cluster_->num_tablet_servers() / 2 - 1; 
i++) {
 : ExternalTabletServer* ts = cluster_->tablet_server(i);
 : // Fail blocks in all of the directories, avoiding the 
metadata directory.
 : vector dirs_to_fail(ts->data_dirs().begin() + 1, 
ts->data_dirs().end());
 : ASSERT_OK(cluster_->SetFlag(ts, "env_inject_eio_globs",
 : JoinStrings(JoinPathSegmentsV(dirs_to_fail, "**"), 
",")));
 :
 : // Fail only a fraction of the time to get better 
failure-path coverage.
 : ASSERT_OK(cluster_->SetF
> not entirely your fault, but can we add an assertion that the failures did
Done



--
To view, visit http://gerrit.cloudera.org:8080/8388
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: kudu
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I966b41b71e39827f8f78fcca59a2208f13f71e53
Gerrit-Change-Number: 8388
Gerrit-PatchSet: 4
Gerrit-Owner: Andrew Wong 
Gerrit-Reviewer: Andrew Wong 
Gerrit-Reviewer: David Ribeiro Alves 
Gerrit-Reviewer: Kudu Jenkins
Gerrit-Reviewer: Tidy Bot
Gerrit-Comment-Date: Tue, 28 Nov 2017 23:20:30 +
Gerrit-HasComments: Yes

[kudu-CR] exactly once writes-itest for faulty disks

[Andrew Wong (Code Review)]

Hello Tidy Bot, David Ribeiro Alves, Kudu Jenkins,

I'd like you to reexamine a change. Please visit

http://gerrit.cloudera.org:8080/8388

to look at the new patch set (#4).

Change subject: exactly_once_writes-itest for faulty disks
..

exactly_once_writes-itest for faulty disks

This patch adds an ExactlyOnceWrites test that exercises intermittently
crashy disks. This serves to not only test disk failures during writes,
but also test disk failures during tablet copies, all while maintaining
exactly-once semantics.

exactly_once_writes-itest is also updated to use std::thread instead of
kudu::Thread, and to favor returning Status instead of void.

Change-Id: I966b41b71e39827f8f78fcca59a2208f13f71e53
---
M src/kudu/integration-tests/exactly_once_writes-itest.cc
M src/kudu/util/test_macros.h
2 files changed, 228 insertions(+), 115 deletions(-)


  git pull ssh://gerrit.cloudera.org:29418/kudu refs/changes/88/8388/4
--
To view, visit http://gerrit.cloudera.org:8080/8388
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: kudu
Gerrit-Branch: master
Gerrit-MessageType: newpatchset
Gerrit-Change-Id: I966b41b71e39827f8f78fcca59a2208f13f71e53
Gerrit-Change-Number: 8388
Gerrit-PatchSet: 4
Gerrit-Owner: Andrew Wong 
Gerrit-Reviewer: David Ribeiro Alves 
Gerrit-Reviewer: Kudu Jenkins
Gerrit-Reviewer: Tidy Bot

[kudu-CR] exactly once writes-itest for faulty disks

[David Ribeiro Alves (Code Review)]

David Ribeiro Alves has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/8388 )

Change subject: exactly_once_writes-itest for faulty disks
..


Patch Set 3:

(3 comments)

http://gerrit.cloudera.org:8080/#/c/8388/3/src/kudu/integration-tests/exactly_once_writes-itest.cc
File src/kudu/integration-tests/exactly_once_writes-itest.cc:

http://gerrit.cloudera.org:8080/#/c/8388/3/src/kudu/integration-tests/exactly_once_writes-itest.cc@115
PS3, Line 115:   Status ValidateLogs();
 :
 :   void ValidateWriteResponses(const WriteResponsePBsPerThread& 
responses,
 :   bool* mismatched);
docs


http://gerrit.cloudera.org:8080/#/c/8388/3/src/kudu/integration-tests/exactly_once_writes-itest.cc@342
PS3, Line 342: TestWritesWithExactlyOnceSemanticsWithCrashNodes
nit: TestWritesWithExactlyOnceSemanticsWithCrashyNodes


http://gerrit.cloudera.org:8080/#/c/8388/3/src/kudu/integration-tests/exactly_once_writes-itest.cc@382
PS3, Line 382:   // Fail under half of the servers so we always maintain a 
healthy majority.
 :   for (int i = 0; i < cluster_->num_tablet_servers() / 2 - 1; 
i++) {
 : ExternalTabletServer* ts = cluster_->tablet_server(i);
 : // Fail blocks in all of the directories, avoiding the 
metadata directory
 : // so we don't crash.
 : vector dirs_to_fail(ts->data_dirs().begin() + 1, 
ts->data_dirs().end());
 : ASSERT_OK(cluster_->SetFlag(ts, "env_inject_eio_globs",
 : JoinStrings(JoinPathSegmentsV(dirs_to_fail, "**"), 
",")));
 :
 : // Fail only a fraction of the time to get better 
failure-path coverage.
 : ASSERT_OK(cluster_->SetFlag(ts, "env_inject_eio", "0.025"));
 :   }
 :
 :   // All servers should stay alive and complete their write 
workload.
 :   ASSERT_OK(DoTestWritesWithExactlyOnceSemantics(num_batches, 
false /* Allow crashes */));
 :   ASSERT_OK(ValidateLogs());
not entirely your fault, but can we add an assertion that the failures did 
indeed happen somehow?



--
To view, visit http://gerrit.cloudera.org:8080/8388
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: kudu
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I966b41b71e39827f8f78fcca59a2208f13f71e53
Gerrit-Change-Number: 8388
Gerrit-PatchSet: 3
Gerrit-Owner: Andrew Wong 
Gerrit-Reviewer: David Ribeiro Alves 
Gerrit-Reviewer: Kudu Jenkins
Gerrit-Reviewer: Tidy Bot
Gerrit-Comment-Date: Tue, 28 Nov 2017 21:15:28 +
Gerrit-HasComments: Yes

[kudu-CR] exactly once writes-itest for faulty disks

[Andrew Wong (Code Review)]

Hello Tidy Bot, David Ribeiro Alves, Kudu Jenkins,

I'd like you to reexamine a change. Please visit

http://gerrit.cloudera.org:8080/8388

to look at the new patch set (#3).

Change subject: exactly_once_writes-itest for faulty disks
..

exactly_once_writes-itest for faulty disks

This patch adds an ExactlyOnceWrites test that exercises intermittently
crashy disks. This serves to not only test disk failures during writes,
but also test disk failures during tablet copies, all while maintaining
exactly-once semantics.

exactly_once_writes-itest is also updated to use std::thread instead of
kudu::Thread, and to favor returning Status instead of void.

Change-Id: I966b41b71e39827f8f78fcca59a2208f13f71e53
---
M src/kudu/integration-tests/exactly_once_writes-itest.cc
M src/kudu/util/test_macros.h
2 files changed, 205 insertions(+), 112 deletions(-)


  git pull ssh://gerrit.cloudera.org:29418/kudu refs/changes/88/8388/3
--
To view, visit http://gerrit.cloudera.org:8080/8388
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: kudu
Gerrit-Branch: master
Gerrit-MessageType: newpatchset
Gerrit-Change-Id: I966b41b71e39827f8f78fcca59a2208f13f71e53
Gerrit-Change-Number: 8388
Gerrit-PatchSet: 3
Gerrit-Owner: Andrew Wong 
Gerrit-Reviewer: David Ribeiro Alves 
Gerrit-Reviewer: Kudu Jenkins
Gerrit-Reviewer: Tidy Bot