Sailesh Mukil has posted comments on this change.

Change subject: PREVIEW: IMPALA-2550 Introduce query-wide execution context.
......................................................................


Patch Set 6:

(1 comment)

http://gerrit.cloudera.org:8080/#/c/3817/5/be/src/runtime/query-exec-state.h
File be/src/runtime/query-exec-state.h:

Line 101:   /// referenced as a shared_ptr to allow asynchronous calls to 
CancelPlanFragment()
> they are, and i have a patch out (3988) that embeds the instance index into
Yes, that will make it much simpler to extract the index.

However, I still think that a scenario such as the following could happen:
If there are 6 fragment instances for a query,

frag_inst#1 -> node1
frag_inst#6 -> node1

frag_inst#2 -> node2
frag_inst#3 -> node2

rest goes to node3

So the FragmentExecState vector in node1 will have a gap between index 1 and 
index 6.

Similarly gaps will exist for node2 and node 3.

I verified this by looking at some minicluster logs and the fragment instance 
IDs are not contiguous per backend.


-- 
To view, visit http://gerrit.cloudera.org:8080/3817
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-MessageType: comment
Gerrit-Change-Id: I892091d6401acb2ea91ccb1623af54c6f9635e6c
Gerrit-PatchSet: 6
Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-Owner: Sailesh Mukil <[email protected]>
Gerrit-Reviewer: Dan Hecht <[email protected]>
Gerrit-Reviewer: David Knupp <[email protected]>
Gerrit-Reviewer: Henry Robinson <[email protected]>
Gerrit-Reviewer: Lars Volker <[email protected]>
Gerrit-Reviewer: Marcel Kornacker <[email protected]>
Gerrit-Reviewer: Sailesh Mukil <[email protected]>
Gerrit-HasComments: Yes

Reply via email to