Adar Dembo has posted comments on this change.

Change subject: design-docs: multi-master for 1.0 release
......................................................................


Patch Set 1:

(5 comments)

http://gerrit.cloudera.org:8080/#/c/2527/1/docs/design-docs/multi-master-1.0.md
File docs/design-docs/multi-master-1.0.md:

Line 120: f
> It's OK if you don't change 1353, but it would be good to have a jira to tr
As we discussed in person, I'm more inclined to have a 1:1 ratio between root 
causes and JIRAs, so I'll modify KUDU-1353 to reflect the DeleteTable() problem.


Line 130: ###
> then it's weird cuz this doc is called multi-master-1.0. Either make no ref
OK, I'll mention next to each "feature" how important I think it is for 1.0.


Line 200: 2. All destructive actions taken by a tserver must be "fenced". That 
is, the
> creating a tablet? splitting a tablet in the future....
I see. I guess which phrase makes more sense depends on whether you want it to 
include operations that are already properly fenced (i.e. tablet creation).

But generally speaking, I think "state changes" is easier to understand, so 
I'll change the wording.


Line 206: Ensure that the leader master replicates via Raft before triggering an
        :       action. It doesn't matter what is replicated (a no-op would 
suffice);
        :       a successful replication asserts that this master is still the 
leader.
> I'm ok if we don't flesh this out right now. think we need to ponder this f
OK, I've reworded this section to suggest this as an alternative, but to 
emphatically propose #1 as the path forward.


Line 212: partially replicated
        : operations
> post your findings here?
As far as raw sizing is concerned, I hacked up create-table-stress-test to 
create a table with a thousand tablets. The Write issued for the addition of 
the thousand tablets was about 114 KB, which is well under the 8 MB RPC limit. 
So that's not an issue, nor will it be an issue for the conceivable future.

I've also updated this section in great detail to reflect the analysis I 
performed, and attached the analysis itself below.


-- 
To view, visit http://gerrit.cloudera.org:8080/2527
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-MessageType: comment
Gerrit-Change-Id: Iad76012977a45370b72a04d608371cecf90442ef
Gerrit-PatchSet: 1
Gerrit-Project: kudu
Gerrit-Branch: master
Gerrit-Owner: Adar Dembo <[email protected]>
Gerrit-Reviewer: Adar Dembo <[email protected]>
Gerrit-Reviewer: Dan Burkert <[email protected]>
Gerrit-Reviewer: David Ribeiro Alves <[email protected]>
Gerrit-Reviewer: Kudu Jenkins
Gerrit-Reviewer: Todd Lipcon <[email protected]>
Gerrit-HasComments: Yes

Reply via email to