Tim Armstrong has posted comments on this change. Change subject: IMPALA-3242: Remove most usages of RuntimeState::SetMemLimitExceeded() ......................................................................
Patch Set 1: (6 comments) http://gerrit.cloudera.org:8080/#/c/3140/1/be/src/exec/hdfs-rcfile-scanner.cc File be/src/exec/hdfs-rcfile-scanner.cc: Line 267: RC file RCFile scanner http://gerrit.cloudera.org:8080/#/c/3140/1/be/src/exec/partitioned-aggregation-node.cc File be/src/exec/partitioned-aggregation-node.cc: Line 1149: int64_t alloc_size = PAGG_DEFAULT_HASH_TABLE_SZ * HashTable::BucketSize(); I'm not a big fan of using the initial size here, since it means changes elsewhere to the sizing have to be reflected here. Can you just get HashTable::num_buckets() before trying to init the hash table? Line 1251: UNLIKELY I don't think this path is perf-critical enough to justify the annotations. Line 1252: status Dead variable. http://gerrit.cloudera.org:8080/#/c/3140/1/be/src/exec/partitioned-hash-join-node.cc File be/src/exec/partitioned-hash-join-node.cc: Line 636: UNLIKELY Again, not sure this is hot enough to justify annotations. http://gerrit.cloudera.org:8080/#/c/3140/1/be/src/runtime/plan-fragment-executor.cc File be/src/runtime/plan-fragment-executor.cc: Line 533: . > by MemTracker::MemLimitExceeded(). Do we know what places remain that don't do that? This might be related to IMPALA-3090 -- To view, visit http://gerrit.cloudera.org:8080/3140 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: Ic0ca128c768d1e73713866e8c513a1b75e6b4b59 Gerrit-PatchSet: 1 Gerrit-Project: Impala Gerrit-Branch: cdh5-2.6.0_5.8.0 Gerrit-Owner: Michael Ho <[email protected]> Gerrit-Reviewer: Dan Hecht <[email protected]> Gerrit-Reviewer: Tim Armstrong <[email protected]> Gerrit-HasComments: Yes
