Adar Dembo has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/10612 )

Change subject: [rebalancer] 'auto' mode for RF=1 tablet movements
......................................................................


Patch Set 7:

(8 comments)

Can you articulate how the existing tests provide sufficient coverage for 
--move_single_replicas=auto?

http://gerrit.cloudera.org:8080/#/c/10612/7/src/kudu/tools/tool_action_cluster.cc
File src/kudu/tools/tool_action_cluster.cc:

http://gerrit.cloudera.org:8080/#/c/10612/7/src/kudu/tools/tool_action_cluster.cc@101
PS7, Line 101: It's best to keep the default value for this "
             :               "flag unless you know what you are doing."
This can be implied using the right flag tag (perhaps advanced or experimental).


http://gerrit.cloudera.org:8080/#/c/10612/7/src/kudu/tools/tool_action_cluster.cc@173
PS7, Line 173:                            nullptr, is_343_scheme);
Nit: add an inline arg to explain what this nullptr means.


http://gerrit.cloudera.org:8080/#/c/10612/7/src/kudu/tools/tool_action_cluster.cc@180
PS7, Line 180:     DCHECK(boost::iequals(FLAGS_move_single_replicas, "enabled") 
||
             :            boost::iequals(FLAGS_move_single_replicas, 
"disabled"));
Nit: I think this would be simpler as a single DCHECK on "disabled", on L185.


http://gerrit.cloudera.org:8080/#/c/10612/7/src/kudu/tools/tool_action_cluster.cc@220
PS7, Line 220: That's because write
             :   // operations in progress (if any) cannot be committed until 
the destination
             :   // replica (added as a voter) is caught up. Catching up the 
destination
             :   // voter replica might take much longer than write operation's 
timeout.
I'll be honest; I don't understand this explanation. Could you clarify further?


http://gerrit.cloudera.org:8080/#/c/10612/7/src/kudu/tools/tool_action_cluster.cc@224
PS7, Line 224:   {
Nit: what's the point of this extra scope?


http://gerrit.cloudera.org:8080/#/c/10612/7/src/kudu/tools/tool_action_cluster.cc@226
PS7, Line 226: is_343_scheme
Don't we have to test this too? If not, why bother setting it to non-null?


http://gerrit.cloudera.org:8080/#/c/10612/7/src/kudu/util/version_util.h
File src/kudu/util/version_util.h:

http://gerrit.cloudera.org:8080/#/c/10612/7/src/kudu/util/version_util.h@60
PS7, Line 60: bool VersionSupportsRF1Movement(const std::string& version_str);
I don't think this belongs here. This is util code, so one would expect the 
functions and classes to be Kudu-agnostic.


http://gerrit.cloudera.org:8080/#/c/10612/7/src/kudu/util/version_util.cc
File src/kudu/util/version_util.cc:

http://gerrit.cloudera.org:8080/#/c/10612/7/src/kudu/util/version_util.cc@85
PS7, Line 85: std::
Nit: drop



--
To view, visit http://gerrit.cloudera.org:8080/10612
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: kudu
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: Id5a06b137cb34d9351f7e2fb819ed52790e2ee7e
Gerrit-Change-Number: 10612
Gerrit-PatchSet: 7
Gerrit-Owner: Alexey Serbin <[email protected]>
Gerrit-Reviewer: Adar Dembo <[email protected]>
Gerrit-Reviewer: Kudu Jenkins
Gerrit-Reviewer: Mike Percy <[email protected]>
Gerrit-Reviewer: Tidy Bot
Gerrit-Reviewer: Will Berkeley <[email protected]>
Gerrit-Comment-Date: Thu, 21 Jun 2018 23:48:11 +0000
Gerrit-HasComments: Yes

Reply via email to