Red Hat Bugzilla – Bug 534317
test cases leave data in database
Last modified: 2011-04-20 19:50:39 EDT
I just ran a full test on my box. I first ensured I had an empty DB (-Ddbsetup), I then ran: mvn -Pdev,enterprise install After 14 minutes, all tests pass and my build was successful. However, I then looked in the database and found some tables still had data in it that should not be. The data looks like test data which should have been cleaned up after a successful test run. The tables with data in them (that should not be there) is: rhq_config rhq_config_prop_grp_def rhq_config_property rhq_partition_event rhq_qrtz_cron_triggers rhq_qrtz_job_details rhq_qrtz_triggers rhq_resource rhq_resource_type I think some of this leftover data is causing future test runs to fail
Here is some of the data leftover that might help track down the offending tests: * rhq_config_prop_grp_def id name display_name description default_hidden order_index 500050 Basic Group (null) (null) true 0 * rhq_config_property id configuration_id string_value override name dtype parent_list_id parent_map_id list_index error_message 501190 509734 testValue (null) testProperty property (null) (null) (null) (null) 501191 509753 value1 false property1 property (null) (null) (null) (null) 501060 509587 test-value! (null) param1 property (null) (null) (null) (null) 501192 509754 value1 (null) property1 property (null) (null) (null) (null) 501193 509757 value1 false property1 property (null) (null) (null) (null) 501063 509622 group-test (null) param1 property (null) (null) (null) (null) 501064 509622 blah (null) param2 property (null) (null) (null) (null) 501194 509758 value1 (null) property1 property (null) (null) (null) (null) 500552 508593 true false myboolean property (null) (null) (null) (null) 500553 508594 true false myboolean property (null) (null) (null) (null) 500554 508595 false false myboolean property (null) (null) (null) (null) 500555 508596 false false myboolean property (null) (null) (null) (null) 500556 508597 TRUE false myboolean property (null) (null) (null) (null) 500557 508598 TRUE false myboolean property (null) (null) (null) (null) 500558 508599 false false myboolean property (null) (null) (null) (null) 500560 508613 true false myboolean property (null) (null) (null) (null) 500565 508625 true false myboolean property (null) (null) (null) (null) 500566 508626 false false myboolean property (null) (null) (null) (null) 500568 508640 true false myboolean property (null) (null) (null) (null) 500569 508641 true false myboolean property (null) (null) (null) (null) 500570 508642 false false myboolean property (null) (null) (null) (null) 500571 508642 7000 false mysleep property (null) (null) (null) (null) 500575 508649 false false myboolean property (null) (null) (null) (null) 500576 508649 7000 false mysleep property (null) (null) (null) (null) 500579 508659 1 false foo property (null) (null) (null) (null) 500580 508660 2 false bar property (null) (null) (null) (null) 500583 508667 true false myboolean property (null) (null) (null) (null) 500584 508668 true false myboolean property (null) (null) (null) (null) 500586 508674 invalid-boolean false myboolean property (null) (null) (null) java.lang.IllegalArgumentException: Not a valid boolean at org.rhq.enterprise.server.configuration.ConfigurationManagerBeanTest$TestServices$1.run(ConfigurationManagerBeanTest.java:933) 501056 509078 UPDATED (null) updateCSName property (null) (null) (null) (null) * rhq_qrtz_cron_triggers trigger_name trigger_group cron_expression time_zone_id testConstraintViolation--testConstraintViolationCST syncContentSource 0 0 3 * * ? America/New_York * rhq_qrtz_job_details job_name job_group description job_class_name is_durable is_volatile is_stateful requests_recovery job_data testConstraintViolation--testConstraintViolationCST syncContentSource (null) org.rhq.enterprise.server.scheduler.jobs.ContentSourceSyncJob false false true false BINARY, 34 Bytes * rhq_qrtz_triggers trigger_name trigger_group job_name job_group is_volatile description next_fire_time prev_fire_time trigger_state trigger_type start_time end_time calendar_name misfire_instr job_data testConstraintViolation--testConstraintViolationCST syncContentSource testConstraintViolation--testConstraintViolationCST syncContentSource false (null) 1226995200000 -1 WAITING CRON 1226934871000 0 (null) 0 (null) * rhq_resource id resource_type_id uuid name resource_key agent_id inventory_status connected description version ctime mtime itime res_configuration_id plugin_configuration_id modified_by location parent_resource_id product_version_id 505297 501105 A A A (null) COMMITTED false (null) (null) 1226934979703 1226934979703 1226934979703 509655 509656 (null) (null) (null) (null) 505298 501105 B B B (null) COMMITTED false (null) (null) 1226934979703 1226934979703 1226934979703 509657 509658 (null) (null) (null) (null) 505288 501105 3 3 3 (null) COMMITTED false (null) (null) 1226934979703 1226934979703 1226934979703 509637 509638 (null) (null) 505298 (null) 505289 501105 2 2 2 (null) COMMITTED false (null) (null) 1226934979703 1226934979703 1226934979703 509639 509640 (null) (null) 505297 (null) 505290 501105 1 1 1 (null) COMMITTED false (null) (null) 1226934979703 1226934979703 1226934979703 509641 509642 (null) (null) 505297 (null) 505291 501105 a a a (null) COMMITTED false (null) (null) 1226934979703 1226934979703 1226934979703 509643 509644 (null) (null) 505290 (null) 505292 501105 i i i (null) COMMITTED false (null) (null) 1226934979703 1226934979703 1226934979703 509645 509646 (null) (null) 505291 (null) 505293 501105 ii ii ii (null) COMMITTED false (null) (null) 1226934979703 1226934979703 1226934979703 509647 509648 (null) (null) 505291 (null) 505294 501105 b b b (null) COMMITTED false (null) (null) 1226934979703 1226934979703 1226934979703 509649 509650 (null) (null) 505290 (null) 505295 501105 iii iii iii (null) COMMITTED false (null) (null) 1226934979703 1226934979703 1226934979703 509651 509652 (null) (null) 505294 (null) 505296 501105 iv iv iv (null) COMMITTED false (null) (null) 1226934979703 1226934979703 1226934979703 509653 509654 (null) (null) 505294 (null) * rhq_resource_type id name category creation_data_type create_delete_policy singleton supports_manual_add description plugin ctime mtime subcategory_id plugin_config_def_id res_config_def_id 501105 fakeType PLATFORM CONFIGURATION BOTH false false (null) fakePlugin 1226934979703 1226934979703 (null) (null) (null)
those rhq_resource data looks like its from RecursiveResourceGroupTest.getSimpleTree
ConfigurationDefinitionTest persists data across tests - though there are test dependencies that should clean it up, there are places in here where we should combine tests and run inside a begin..rollback block so we don't risk leaking data, even on failure.
ContentSourceManagerBeanTest.testConstraintViolation didn't seem to clean up all the way a content source thus leaving things in the quartz tables.
OperationManagerBeanTest might be leaving data too - found test-value! in DB and its a value the test persists.
fixed the quartz tables leftover data and some of the config data - see rhq-1126
svn rev 2018 fixed ConfigurationDefinitionTest so it doesn't leak data and consolidated tests so it never actually commited data to the db
Fixed a few things, still need to fix these tests: RecursiveResourceGroupTest, FailoverListManagerBeanTest, OperationManagerBeanTest, ConfigurationManagerBeanTest Here's the remainnig data that is still leaked which appear to be from one of the above: * rhq_config (65 rows) * rhq_config_property id configuration_id string_value override name dtype parent_list_id parent_map_id list_index error_message 500052 500093 true false myboolean property (null) (null) (null) (null) 500053 500094 true false myboolean property (null) (null) (null) (null) 500054 500095 false false myboolean property (null) (null) (null) (null) 500055 500096 false false myboolean property (null) (null) (null) (null) 500056 500097 TRUE false myboolean property (null) (null) (null) (null) 500057 500098 TRUE false myboolean property (null) (null) (null) (null) 500058 500099 false false myboolean property (null) (null) (null) (null) 500060 500113 true false myboolean property (null) (null) (null) (null) 500065 500125 true false myboolean property (null) (null) (null) (null) 500066 500126 false false myboolean property (null) (null) (null) (null) 500068 500140 true false myboolean property (null) (null) (null) (null) 500069 500141 true false myboolean property (null) (null) (null) (null) 500070 500142 false false myboolean property (null) (null) (null) (null) 500071 500142 7000 false mysleep property (null) (null) (null) (null) 500075 500149 false false myboolean property (null) (null) (null) (null) 500076 500149 7000 false mysleep property (null) (null) (null) (null) 500079 500159 1 false foo property (null) (null) (null) (null) 500080 500160 2 false bar property (null) (null) (null) (null) 500083 500167 true false myboolean property (null) (null) (null) (null) 500084 500168 true false myboolean property (null) (null) (null) (null) 500086 500174 invalid-boolean false myboolean property (null) (null) (null) java.lang.IllegalArgumentException: Not a valid boolean at org.rhq.enterprise.server.configuration.ConfigurationManagerBeanTest$TestServices$1.run(ConfigurationManagerBeanTest.java:933) 500190 501334 testValue (null) testProperty property (null) (null) (null) (null) 500191 501353 value1 false property1 property (null) (null) (null) (null) 500192 501354 value1 (null) property1 property (null) (null) (null) (null) 500193 501357 value1 false property1 property (null) (null) (null) (null) 500194 501358 value1 (null) property1 property (null) (null) (null) (null) 500560 501187 test-value! (null) param1 property (null) (null) (null) (null) 500563 501222 group-test (null) param1 property (null) (null) (null) (null) 500564 501222 blah (null) param2 property (null) (null) (null) (null) * rhq_partition_event (216 rows, same as before) * rhq_resource (11 rows, same as before) * rhq_resource_type (1 row, same as before)
Improved FailoverListManagerBeanTest cleanup logic. It was leaving behind some PartitionEvent rows, which in turn left PartitionDetails, FailoverLists and FailoverDetails. This should all be good, now.
r2036 Updated tests to clean up debris in rhq_config (22 of 65)), rhq_resource (11 of 11) and rhq_resource_type (1 of 1)
This bug was previously known as http://jira.rhq-project.org/browse/RHQ-1125 This bug relates to RHQ-1128 This bug is incorporated by RHQ-1126
mass move off the qa triage list. These are tasks for dev.
jay, what's the current state of this - is it still an issue?
I'm sure there must still be stuff left in the db after a set of test runs. It was a year and a half ago when this last saw activity, and at that time we had cleaned it up pretty well. Now, I don't know. I'd close this unless you want a serious effort put into a new analysis. Our test runs don't seem to suffer too much from this sort of thing from what I've seen lately. And now that we have maven set up to easily use separate test and dev dbs it has reduced issues of this sort by isolating the test data.
no longer relevant per jay's comments.