Hao Hao has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/8804 )

Change subject: KUDU-1704: add READ_OWN_WRITES scan mode
......................................................................


Patch Set 2:

(20 comments)

http://gerrit.cloudera.org:8080/#/c/8804/2/src/kudu/common/common.proto
File src/kudu/common/common.proto:

http://gerrit.cloudera.org:8080/#/c/8804/2/src/kudu/common/common.proto@222
PS2, Line 222: BOUNDED_READ
> READ_OWN_WRITES ?
Done


http://gerrit.cloudera.org:8080/#/c/8804/2/src/kudu/common/common.proto@222
PS2, Line 222: pick an arbitrarily
> nit: "arbitrarily pick" or "pick an arbitrary"
Right, the picked snapshot is subject to the criteria you mentioned. Maybe 
'arbitrary' is not a good word here. I will remove it. The intention was to 
specify the snapshot is not something specifically controlled by user, even the 
rule to chose one is specific.


http://gerrit.cloudera.org:8080/#/c/8804/2/src/kudu/common/common.proto@225
PS2, Line 225: staleness bound that allows execution of the reads
> This "staleness bound" isn't mentioned anywhere else, so I'm not sure exact
Done


http://gerrit.cloudera.org:8080/#/c/8804/2/src/kudu/common/common.proto@229
PS2, Line 229: Boundedly stale reads
> nit: maybe "Reads in this mode"?
Done


http://gerrit.cloudera.org:8080/#/c/8804/2/src/kudu/tserver/tablet_server-test.cc
File src/kudu/tserver/tablet_server-test.cc:

http://gerrit.cloudera.org:8080/#/c/8804/2/src/kudu/tserver/tablet_server-test.cc@a1616
PS2, Line 1616:
              :
              :
> nit: various spurious deletions of newlines
I thought only one line is needed between tests?


http://gerrit.cloudera.org:8080/#/c/8804/2/src/kudu/tserver/tablet_server-test.cc@1835
PS2, Line 1835: TestBoundedScan
> nit: this should be renamed (same with other tests)
Done


http://gerrit.cloudera.org:8080/#/c/8804/2/src/kudu/tserver/tablet_server-test.cc@1843
PS2, Line 1843:   RpcController rpc;
> nit: move this under the 'Send the call' scope.
Done


http://gerrit.cloudera.org:8080/#/c/8804/2/src/kudu/tserver/tablet_server-test.cc@1855
PS2, Line 1855: s
> nit: make it capital
Done


http://gerrit.cloudera.org:8080/#/c/8804/2/src/kudu/tserver/tablet_server-test.cc@1866
PS2, Line 1866: the no snapshot
> nit: typo, same in other tests.
Done


http://gerrit.cloudera.org:8080/#/c/8804/2/src/kudu/tserver/tablet_server-test.cc@1869
PS2, Line 1869: resp.has_propagated_timestamp()
> nit: maybe pull into a local variable?
I assume you mean 'resp.propagated_timestamp()'?


http://gerrit.cloudera.org:8080/#/c/8804/2/src/kudu/tserver/tablet_server-test.cc@1878
PS2, Line 1878: num_rows
> micro-nit: kNumRows?
Done


http://gerrit.cloudera.org:8080/#/c/8804/2/src/kudu/tserver/tablet_server-test.cc@1899
PS2, Line 1899:   req.set_batch_size_bytes(0); // so it won't return data right 
away
> Is this needed here? I thought we only needed it to postpone draining the s
Done


http://gerrit.cloudera.org:8080/#/c/8804/2/src/kudu/tserver/tablet_server-test.cc@1917
PS2, Line 1917:   ASSERT_LT(pre_scan_ts.ToUint64(), 
resp.propagated_timestamp());
> Should the result scan contain any rows?  Does it make sense to assert on t
I do not think it makes sense to assert on that, since mvcc clean timestamp is 
hard to predefined.


http://gerrit.cloudera.org:8080/#/c/8804/2/src/kudu/tserver/tablet_server-test.cc@1922
PS2, Line 1922: post
> ?
Done


http://gerrit.cloudera.org:8080/#/c/8804/2/src/kudu/tserver/tablet_server-test.cc@1968
PS2, Line 1968: ASSERT_NO_FATAL_FAILURE
> nit: I think you can use the shorter "NO_FATALS" here
Done


http://gerrit.cloudera.org:8080/#/c/8804/2/src/kudu/tserver/tablet_service.h
File src/kudu/tserver/tablet_service.h:

http://gerrit.cloudera.org:8080/#/c/8804/2/src/kudu/tserver/tablet_service.h@156
PS2, Line 156:   Status PickAndVerifyTimeStamp(const NewScanRequestPB& scan_pb,
> nit: add comments on what these two methods do.
Done


http://gerrit.cloudera.org:8080/#/c/8804/2/src/kudu/tserver/tablet_service.cc
File src/kudu/tserver/tablet_service.cc:

http://gerrit.cloudera.org:8080/#/c/8804/2/src/kudu/tserver/tablet_service.cc@1619
PS2, Line 1619: BOUNDED_READ
> nit: READ_OWN_WRITES?
Done


http://gerrit.cloudera.org:8080/#/c/8804/2/src/kudu/tserver/tablet_service.cc@2096
PS2, Line 2096: BOUNDED_READ
> nit: READ_OWN_WRITES
Done


http://gerrit.cloudera.org:8080/#/c/8804/2/src/kudu/tserver/tablet_service.cc@2153
PS2, Line 2153: BOUNDED_READ
> nit: READ_OWN_WRITES ?
Done


http://gerrit.cloudera.org:8080/#/c/8804/2/src/kudu/tserver/tablet_service.cc@2162
PS2, Line 2162:       tmp_snap_timestamp = tmp_snap_timestamp > 
mvcc_manager->GetCleanTimestamp() ?
              :                            tmp_snap_timestamp : 
mvcc_manager->GetCleanTimestamp();
> nit: maybe, call GetCleanTimestamp() just once, storing the result in the t
Done



--
To view, visit http://gerrit.cloudera.org:8080/8804
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: kudu
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I84ddb981a1a0f199d4e66f5d5097318f8c785a48
Gerrit-Change-Number: 8804
Gerrit-PatchSet: 2
Gerrit-Owner: Hao Hao <[email protected]>
Gerrit-Reviewer: Alexey Serbin <[email protected]>
Gerrit-Reviewer: Andrew Wong <[email protected]>
Gerrit-Reviewer: David Ribeiro Alves <[email protected]>
Gerrit-Reviewer: Hao Hao <[email protected]>
Gerrit-Reviewer: Kudu Jenkins
Gerrit-Comment-Date: Thu, 14 Dec 2017 07:33:46 +0000
Gerrit-HasComments: Yes

Reply via email to