uwescience / myria

60%

master: 27%

LAST BUILD ON BRANCH exit_when_OOM

branch: exit_when_OOM

CHANGE BRANCH

x

Reset

• exit_when_OOM
• CSVFragmentTupleSource-bug-fix
• FileSink
• IDB_store_and_schema_inference
• MPW_reconfig
• add-binary-support
• add-driver-dir
• add-partition-function-to-ingest
• add-s3-UriSink
• add_PF_in_JSON_ingest
• add_broadcast
• add_code_formatter
• add_count_filter
• add_few_recursive_json
• add_limit_to_dataset_api
• add_myriax_doc
• anonymous-for-local-deployment
• ansible_fix
• blob-udf-new-merge
• blob_expr_UDF
• blob_literal
• byteRangeExpression
• catalog-scan
• concatExpression
• cosmo8Fix
• create-deployment
• cross_compile
• dbExecute
• deleteTables
• deploy_python
• doc-ingest
• dup_elim_hash
• elastic_cluster
• enable_worker_sysgc
• error_reporting
• exclude_old_jersey
• fix-ipc-channel-failure
• fix-null-child-when-generating-schema
• fix-old-jsons
• fix-query-summary
• fix_conf_parsing
• fix_constant_emit_in_statefulapply
• fix_cp_return_code
• fix_dup_opIds
• fix_fragment_assigned_workers
• fix_hanging_query_when_output_unavailable
• fix_some_jsons
• fix_state_col_offset
• fix_tbb_column_size_with_blob
• fixing-CSVFragmentTupleSource
• flatmap_apply
• flatmap_expressions
• gradle-2.3
• gradle-cache
• hash_table_stats
• hdfs-persistence
• hyrkas-msd-code
• imply_clean_catalog
• int-to-boolean-casting
• integration_tests
• java_udf
• jmx
• jmx_symlink_overwrite
• jortiz16-partition
• lbrendanl_elastic
• load_options
• master
• mod_op
• multi_file_scan_script
• myria-jar-deploy-jortiz
• myriaPythonWorker
• new_blob_type
• opid-private
• optimize_aggregates
• optimize_local_join
• optimize_stringcol
• orderbyTests
• orzikhd_sql_resource
• output-encoding
• overwrite_fix
• parallel-ingest-operator-for-raco
• parallelIngest
• parallelIngest-bug
• parallelTest
• parallel_tester
• partitioning_order
• perfenforce-merge
• postgres_float
• pre_refactoring_merge
• production
• profiling-relations
• prune_hash_join
• pythonIO
• pyudf_string
• readme
• reef
• reef-xdgmm
• reef550project
• reef_merge_tmp
• refactor-data-inputs
• refactor_agg
• refactor_deployment
• renameImport
• rename_to_alternate
• restart_cluster
• return_proper_message_for_failed_ingestion
• revert_refactoring
• s3_uri_fix
• several-extensions
• split_operator
• stats_collector
• streaming_query_results
• test-branch-add-worker
• tipsy_hdfs_support
• tipsy_url_support
• track_relations_in_query_scan
• travis-s3
• uda-rewrite-script-eval
• uri_fix
• urisource-http
• workerid_race

Build # 950

Build Type

push

travis-ci

Committed by jingjingwang

Commit Message

make JVM exit when OOM
WHen the JVM is close to an OOM state, the OOM error can be thrown at any time by any thread, even threads that are not used for query processing. The old implementatiob takes a long way to propagate an OOM error from the lowest level to Worker level, and the system state can crash during this process so that the query is still marked as running but the JVM cannot respond anymore. On the other hand, it is unnecessary to clean up after an OOM since the JVM is in an OOM state anyway. So we terminate the worker process when an OOM is detected.

Run Details

26 of 26 new or added lines in 5 files covered. (100.0%)

14006 of 23427 relevant lines covered (59.79%)

1.75 hits per line