[
https://issues.apache.org/jira/browse/HADOOP-11892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15238684#comment-15238684
]
Xiao Chen commented on HADOOP-11892:
------------------------------------
Thanks for the comments Wei-Chiu and Andrew.
bq. And you can also use lsof /dev/urandom to list open fds.
Yes, we can get the process opening /dev/urandom using {{lsof}}. But for
details on how those fds are kept open, I think we'd need to check jstack as
well to help diagnose.
bq. IMO, fixing this isn't really useful.
Fair point. As Andrew pointed out, /dev/urandom doesn't block. But we did see
the process has thousands of /dev/urandom fd open, which indicates something
else. I feel this jira could improve the situation in some degree.
OTOH, so far the occurrence of this issue are all without HADOOP-11891, so it's
hard to tell whether there are other issues left (e.g. some rare case leaking
etc.). If we don't feel much value of this jira, maybe we could revisit this if
we see any reproduction with the existence of HADOOP-11891, and analyze the
jstack then.
> CryptoCodec#getInstance always returns a new instance of CryptoCodec. This
> could be expensive
> ---------------------------------------------------------------------------------------------
>
> Key: HADOOP-11892
> URL: https://issues.apache.org/jira/browse/HADOOP-11892
> Project: Hadoop Common
> Issue Type: Bug
> Affects Versions: 2.7.0
> Reporter: Arun Suresh
> Assignee: Arun Suresh
>
> {{CryptoCodec#getInsatnce}} should be able to return possibly cached
> instances of the CryptoCodec implementation as instantiating a new instance
> every time could be expensive
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
