Murtadha Hubail has posted comments on this change. Change subject: ASTERIXDB-1340: fix node group nodes order ......................................................................
Patch Set 1: (1 comment) https://asterix-gerrit.ics.uci.edu/#/c/708/1/asterix-common/src/main/java/org/apache/asterix/common/config/AsterixPropertiesAccessor.java File asterix-common/src/main/java/org/apache/asterix/common/config/AsterixPropertiesAccessor.java: Line 140: return nodeNames; > >> "Some query ..., and the others"? To tell you the truth, I was hoping you could tell me :P What I know is that the operators that have location constrains take it from the locations of file splits in SplitsAndConstraintsUtil.splitsForDataset(.), while others take it from the already sorted list in APIFramework.compileQuery(.). But what specifies the order in which the task starts, I'm not sure. I spent sometime looking at the plans in the query that works and the one that doesn't, but couldn't find something suspicious. I saw the JIRA that you opened and I believe implementing the suggestion is going to eliminate those position specific file splits. -- To view, visit https://asterix-gerrit.ics.uci.edu/708 To unsubscribe, visit https://asterix-gerrit.ics.uci.edu/settings Gerrit-MessageType: comment Gerrit-Change-Id: I6e93aedb6cdc57d98f83991ab910476b82bacb15 Gerrit-PatchSet: 1 Gerrit-Project: asterixdb Gerrit-Branch: master Gerrit-Owner: Murtadha Hubail <[email protected]> Gerrit-Reviewer: Jenkins <[email protected]> Gerrit-Reviewer: Murtadha Hubail <[email protected]> Gerrit-Reviewer: Yingyi Bu <[email protected]> Gerrit-HasComments: Yes
