virajjasani commented on pull request #2932:
URL: https://github.com/apache/hadoop/pull/2932#issuecomment-822431445
Before creating this PR, local build went well. Hence I raised PR and since
we have updated License-binary file, we will have full QA build taking 17+ hr
worth of time. Hopef
virajjasani commented on pull request #2932:
URL: https://github.com/apache/hadoop/pull/2932#issuecomment-823055329
For the consistently failing tests, I see mostly timeout issues. As for the
release notes of wildfly-openssl 2.x, it's not apparently well accessible. I
tried looking into th
virajjasani commented on pull request #2932:
URL: https://github.com/apache/hadoop/pull/2932#issuecomment-824987122
> The reason we use wildfly in the azure and -aws JARs is for performance by
picking up those native bindings. if you use wildfly ssl as your socket
factory, when openssl bin
virajjasani commented on pull request #2932:
URL: https://github.com/apache/hadoop/pull/2932#issuecomment-824996410
Btw, the main reason behind this upgrade is CVE:
https://nvd.nist.gov/vuln/detail/CVE-2020-25644
```
A memory leak flaw was found in WildFly OpenSSL in versions prior to
virajjasani commented on pull request #2932:
URL: https://github.com/apache/hadoop/pull/2932#issuecomment-830580979
@steveloughran If you suggest to not move forward due to the possibility of
regression in Azure and AWS Jars, I can close this PR and maybe we can give
this a shot later, pro