[
https://issues.apache.org/jira/browse/HADOOP-18950?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17943597#comment-17943597
]
ASF GitHub Bot commented on HADOOP-18950:
-----------------------------------------
steveloughran commented on PR #4854:
URL: https://github.com/apache/hadoop/pull/4854#issuecomment-2797124945
> If we need to support users who want to do their own Avro serialization of
Hadoop classes, then I think we should abandon this PR. I think it would be far
easier to just upgrade the actual Avro jars that Hadoop uses and give up on
shading it.
We now have a shaded avro jar, which we use internally -right?
we still have that problem in branch-3.4 of the 1.9.2 dependency in
serialization though.
I'm thinking: we pull all public classes related to nonshaded avro into a
hadoop-serialization module which is not bundled into hadoop distros; there if
you want it, but not by default
> upgrade avro to 1.11.3 due to CVE
> ---------------------------------
>
> Key: HADOOP-18950
> URL: https://issues.apache.org/jira/browse/HADOOP-18950
> Project: Hadoop Common
> Issue Type: Bug
> Components: common
> Reporter: Xuze Yang
> Priority: Major
> Labels: pull-request-available
>
> [https://nvd.nist.gov/vuln/detail/CVE-2023-39410]
> When deserializing untrusted or corrupted data, it is possible for a reader
> to consume memory beyond the allowed constraints and thus lead to out of
> memory on the system. This issue affects Java applications using Apache Avro
> Java SDK up to and including 1.11.2. Users should update to apache-avro
> version 1.11.3 which addresses this issue.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
