David Handermann created NIFI-10989:
---------------------------------------
Summary: Intermittent Failures in TestHashContent
Key: NIFI-10989
URL: https://issues.apache.org/jira/browse/NIFI-10989
Project: Apache NiFi
Issue Type: Bug
Reporter: David Handermann
Assignee: David Handermann
The {{TestHashContent.testSHA1}} method fails intermittently on GitHub
automated builds, and also fails consistently when running as an individual
test class with the following error:
{noformat}
Error: Tests run: 3, Failures: 1, Errors: 0, Skipped: 0, Time elapsed: 0.063 s
<<< FAILURE! - in org.apache.nifi.processors.standard.TestHashContent
Error: org.apache.nifi.processors.standard.TestHashContent.testSHA1 Time
elapsed: 0.016 s <<< FAILURE!
org.opentest4j.AssertionFailedError:
Processor has 1 validation failures:
'Hash Algorithm' validated against 'SHA-1' is invalid because Given value not
found in allowed set 'SHA3-512, SHA-384, SHA, SHA3-384, SHA-224, SHA-512/256,
SHA-256, MD2, SHA-512/224, SHA3-256, SHA-512, MD5, SHA3-224'
{noformat}
The problem is the result of {{Security.getAlgorithms("MessageDigest")}}
returning different results in different scenarios. The standard JDK returns
{{SHA}} but the Bouncy Castle Security Provider returns {{SHA-1}}. In some
Maven builds, other classes invoke {{Security.addProvider(new
BouncyCastleProvider())}} which changes the registered security providers for
the entire Java Runtime. This allows the current test method to run
successfully. In scenarios where the Bouncy Castle Provider is not registered,
the test fails.
The {{HashContent}} Processor is deprecated in favor of
{{CryptographicHashContent}} and will be removed in subsequent major releases.
Although one solution would involve registering the Bouncy Castle Provider,
removing the test method avoids intermittent failures and simplifies future
maintenance with impending removal the Processor.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
