Ted Yu created HBASE-21279:
------------------------------
Summary: Split TestAdminShell into several tests
Key: HBASE-21279
URL: https://issues.apache.org/jira/browse/HBASE-21279
Project: HBase
Issue Type: Test
Reporter: Ted Yu
In the flaky test board, TestAdminShell often timed out
(https://builds.apache.org/job/HBASE-Find-Flaky-Tests/job/branch-2/lastSuccessfulBuild/artifact/dashboard.html).
I ran the test on Linux with SSD and reproduced the timeout (see attached test
output).
{code}
2018-10-08 02:36:09,146 DEBUG [main] hbase.HBaseTestingUtility(351): Setting
hbase.rootdir to
/mnt/disk2/a/2-hbase/hbase-shell/target/test-data/a103d8e4-695c-a5a9-6690-1ef2580050f9
...
2018-10-08 02:49:09,093 DEBUG
[RpcServer.default.FPBQ.Fifo.handler=27,queue=0,port=33337]
master.MasterRpcServices(1171): Checking to see if procedure is done pid=871
Took 0.7262 seconds2018-10-08 02:49:09,324 DEBUG [PEWorker-1]
util.FSTableDescriptors(684): Wrote into
hdfs://localhost:43859/user/hbase/test-data/cefc73d9-cc37-d2a6-b92b-
d935316c9241/.tmp/data/default/hbase_shell_tests_table/.tabledesc/.tableinfo.0000000001
2018-10-08 02:49:09,328 INFO
[RegionOpenAndInitThread-hbase_shell_tests_table-1] regionserver.HRegion(7004):
creating HRegion hbase_shell_tests_table HTD ==
'hbase_shell_tests_table', {NAME => 'x', VERSIONS => '5', EVICT_BLOCKS_ON_CLOSE
=> 'false', NEW_VERSION_BEHAVIOR => 'false', KEEP_DELETED_CELLS => 'FALSE',
CACHE_DATA_ON_WRITE => 'false', DATA_BLOCK_ENCODING => 'NONE',
TTL => 'FOREVER', MIN_VERSIONS => '0', REPLICATION_SCOPE => '0', BLOOMFILTER =>
'ROW', CACHE_INDEX_ON_WRITE => 'false', IN_MEMORY => 'false',
CACHE_BLOOMS_ON_WRITE => 'false', PREFETCH_BLOCKS_ON_OPEN => 'false',
COMPRESSION => 'NONE', BLOCKCACHE => 'true', BLOCKSIZE => '65536'}, {NAME
=> 'y', VERSIONS => '1', EVICT_BLOCKS_ON_CLOSE => 'false', NEW_VERSION_BEHAVIOR
=> 'false', KEEP_DELETED_CELLS => 'FALSE', CACHE_DATA_ON_WRITE => 'false',
DATA_BLOCK_ENCODING => 'NONE', TTL => 'FOREVER', MIN_VERSIONS => '0',
REPLICATION_SCOPE => '0', BLOOMFILTER => 'ROW', CACHE_INDEX_ON_WRITE =>
'false', IN_MEMORY => 'false', CACHE_BLOOMS_ON_WRITE => 'false',
PREFETCH_BLOCKS_ON_OPEN => 'false', COMPRESSION => 'NONE', BLOCKCACHE =>
'true', BLOCKSIZE => '65536'} RootDir = hdfs://localhost:43859/
user/hbase/test-data/cefc73d9-cc37-d2a6-b92b-d935316c9241/.tmp Table name ==
hbase_shell_tests_table
^[[38;5;226mE^[[0m
===============================================================================
Error: ^[[48;5;16;38;5;226;1mtest_Get_simple_status(Hbase::StatusTest)^[[0m:
Java::JavaIo::InterruptedIOException: Interrupt while waiting on Operation:
CREATE, Table Name: default:hbase_shell_tests_table, procId: 871
2018-10-08 02:49:09,361 INFO [Block report processor]
blockmanagement.BlockManager(2645): BLOCK* addStoredBlock: blockMap updated:
127.0.0.1:41338 is added to
blk_1073742193_1369{UCState=COMMITTED, truncateBlock=null, primaryNodeIndex=-1,
replicas=[ReplicaUC[[DISK]DS-ecc89143-e0a5-4a1c-b552-120be2561334:NORMAL:127.0.0.1:
41338|RBW]]} size 58
====> TEST TIMED OUT. PRINTING THREAD DUMP. <====
{code}
We can see that the procedure #871 wasn't stuck - the timeout cut in and
stopped the test.
We should separate the current test into two (or more) test files (with
corresponding .rb) so that the execution time consistently would not exceed
limit.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
