Philip Zeyliger has posted comments on this change. (
http://gerrit.cloudera.org:8080/11574 )
Change subject: Report the top 5 slowest test steps when running on S3
......................................................................
Patch Set 1:
Seems reasonable. I've got no objection here as long as it works :)
If you're interested, I have the following questionable code hanging out.
Combined with a little bit of sorting, you might be able to get a sense of what
tests are slow. There's been some question about whether the "time" attribute
is accurate (and some JIRAs about it that I've not dug up). Jenkins definitely
shows wonky-looking times for some of our tests, but I've not investigated.
If you think this is a useful direction, feel free to take this over and we can
stick it in lib/py.
$python ~/src/impala/parsexunit.py $(find . -name '*.xml') | sort -n -k 1 -r |
head
623.787399769 results TEST-impala-parallel.xml
query_test.test_decimal_fuzz.TestDecimalFuzz
query_test.test_decimal_fuzz.TestDecimalFuzz.test_decimal_ops[exec_option:
{'batch_size': 0, 'num_nodes': 0, 'disable_codegen_rows_threshold': 5000,
'disable_codegen': False, 'abort_on_error': 1, 'debug_action': None,
'exec_single_node_rows_threshold': 0}] success
476.49873209 results TEST-impala-custom-cluster.xml
custom_cluster.test_permanent_udfs.TestUdfPersistence
custom_cluster.test_permanent_udfs.TestUdfPersistence.test_java_udfs_hive_integration
success
286.006864071 results TEST-impala-parallel.xml
query_test.test_udfs.TestUdfExecution
query_test.test_udfs.TestUdfExecution.test_native_functions[exec_option:
{'disable_codegen_rows_threshold': 0, 'disable_codegen': False,
'exec_single_node_rows_threshold': 100, 'enable_expr_rewrites': True} |
table_format: text/none] success
276.820947886 results TEST-impala-parallel.xml
query_test.test_udfs.TestUdfExecution
query_test.test_udfs.TestUdfExecution.test_native_functions[exec_option:
{'disable_codegen_rows_threshold': 0, 'disable_codegen': False,
'exec_single_node_rows_threshold': 0, 'enable_expr_rewrites': True} |
table_format: text/none] success
256.676662683 results TEST-impala-parallel.xml
metadata.test_ddl.TestDdlStatements
metadata.test_ddl.TestDdlStatements.test_alter_table[exec_option: {'sync_ddl':
0, 'batch_size': 0, 'num_nodes': 0, 'disable_codegen_rows_threshold': 0,
'disable_codegen': False, 'abort_on_error': 1, 'debug_action': None,
'exec_single_node_rows_threshold': 0} | table_format:
text/none-unique_database0] success
247.263021231 results TEST-impala-parallel.xml
query_test.test_nested_types.TestNestedTypes
query_test.test_nested_types.TestNestedTypes.test_subplan[exec_option:
{'batch_size': 0, 'num_nodes': 0, 'disable_codegen_rows_threshold': 0,
'disable_codegen': False, 'abort_on_error': 1, 'debug_action': None,
'exec_single_node_rows_threshold': 0} | table_format: parquet/none]
success
227.564936876 results TEST-impala-parallel.xml
query_test.test_spilling.TestSpillingDebugActionDimensions
query_test.test_spilling.TestSpillingDebugActionDimensions.test_spilling[exec_option:
{'debug_action': '-1:OPEN:[email protected]',
'default_spillable_buffer_size': '256k'} | table_format: parquet/none] success
225.499986172 results TEST-impala-parallel.xml
metadata.test_show_create_table.TestShowCreateTable
metadata.test_show_create_table.TestShowCreateTable.test_show_create_table[table_format:
text/none] success
222.650696039 results TEST-impala-parallel.xml
query_test.test_spilling.TestSpillingDebugActionDimensions
query_test.test_spilling.TestSpillingDebugActionDimensions.test_spilling[exec_option:
{'debug_action': None, 'default_spillable_buffer_size': '256k'} |
table_format: parquet/none] success
211.495414019 results TEST-impala-parallel.xml
query_test.test_udfs.TestUdfTargeted
query_test.test_udfs.TestUdfTargeted.test_concurrent_jar_drop_use[exec_option:
{'batch_size': 0, 'num_nodes': 0, 'disable_codegen_rows_threshold': 0,
'disable_codegen': False, 'abort_on_error': 1, 'debug_action': None,
'exec_single_node_rows_threshold': 0} | table_format: text/none] success
$cat parsexunit.py
import sys
import xml.etree.ElementTree as ET
import os
MAX_MESSAGE_LENGTH = 70
def work():
for f in sys.argv[1:]:
tree = ET.parse(file(f))
# TODO: how do we find xfail?
# expected test failure
for testcase in tree.findall(".//testcase"):
state_seen = False
# Default state is SUCCESS
testcase_state = "success"
message = ""
STATES = ("error", "failure", "skipped")
OTHER_CHILDREN = ("system-out", "system-err")
for child in testcase.getchildren():
assert child.tag in STATES or child.tag in OTHER_CHILDREN, child.tag
for state in STATES:
child = testcase.findall(state)
if child:
if state_seen:
# Beware, there are sometimes multiple
# inners.
pass
testcase_state = state
state_seen = True
message = child[0].attrib.get("message", "").replace("\n", "
").replace("\t", " ")
if len(message) > MAX_MESSAGE_LENGTH:
message = message[0:MAX_MESSAGE_LENGTH - 3] + "..."
print "\t".join([
testcase.attrib["time"],
# testsuite.attrib["name"],
# testcase.attrib["status"],
f.split("/")[3],
os.path.basename(f),
testcase.attrib["classname"],
testcase.attrib["classname"] + "." + testcase.attrib["name"],
testcase_state,
message
])
if __name__ == "__main__":
work()
--
To view, visit http://gerrit.cloudera.org:8080/11574
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings
Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I346fb8c7d57d1dbc7f5c2b8214823f7d8af0522b
Gerrit-Change-Number: 11574
Gerrit-PatchSet: 1
Gerrit-Owner: Laszlo Gaal <[email protected]>
Gerrit-Reviewer: Csaba Ringhofer <[email protected]>
Gerrit-Reviewer: Impala Public Jenkins <[email protected]>
Gerrit-Reviewer: Philip Zeyliger <[email protected]>
Gerrit-Reviewer: Vuk Ercegovac <[email protected]>
Gerrit-Comment-Date: Thu, 04 Oct 2018 23:18:08 +0000
Gerrit-HasComments: No
