Dan Hecht has posted comments on this change. Change subject: IMPALA-2399: Check for mem limit in allocations in parquet scanner and decompressor ......................................................................
Patch Set 13: (4 comments) http://gerrit.cloudera.org:8080/#/c/2203/13/be/src/exec/hdfs-scanner.cc File be/src/exec/hdfs-scanner.cc: Line 198: RETURN_IF_ERROR(parse_status_); as we discussed, using parse_status_ here is a step backwards -- now this central code is making assumptions about the various scanner implementations. Let's keep the parse_status_ grossness isolated to the scanner implementations (and one day maybe remove it) by restructuring the parquet scanner control flow. Line 201: GetQueryState Should this say GetQueryStatus()? http://gerrit.cloudera.org:8080/#/c/2203/13/be/src/exec/hdfs-scanner.h File be/src/exec/hdfs-scanner.h: Line 228: scanner_inited_ scanner_prepared_. Or how about just calling it prepared_ since "scanner" is implied by the class? Also I don't think we want to get into the habit of defining debug only fields - it makes making sure the builds always work a bit tougher. As we discussed, maybe clearer anyway to just make this always defined, and then in the Close() implementations make the guard: "if (prepared_)". http://gerrit.cloudera.org:8080/#/c/2203/13/be/src/runtime/runtime-state.h File be/src/runtime/runtime-state.h: Line 242: inline why? this path is only used when generating an error so it's not a perf critical path. -- To view, visit http://gerrit.cloudera.org:8080/2203 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: Ic70400407b7662999332448f4d1bce2cc344ca89 Gerrit-PatchSet: 13 Gerrit-Project: Impala Gerrit-Branch: cdh5-trunk Gerrit-Owner: Michael Ho <[email protected]> Gerrit-Reviewer: Dan Hecht <[email protected]> Gerrit-Reviewer: Michael Ho <[email protected]> Gerrit-Reviewer: Skye Wanderman-Milne <[email protected]> Gerrit-Reviewer: Tim Armstrong <[email protected]> Gerrit-HasComments: Yes
