hadoop-hdfs-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Andrew Wang <andrew.w...@cloudera.com>
Subject Re: HADOOP-13636 and io.bytes.per.checksum
Date Fri, 23 Sep 2016 17:55:07 GMT
Have you git blamed to dig up the original JIRA conversation? I think that
deprecation predates many of us, so you might not get much historical
perspective from the mailing list.

I'm happy to lend a +1 though, since like you said, it doesn't seem like
that config key is going anywhere.

On Fri, Sep 23, 2016 at 1:52 AM, Steve Loughran <stevel@hortonworks.com>

> I got silence from HDFS dev here, so I'm raising it on common dev.
> Why is HDFS tagging as deprecated " io.bytes.per.checksum ", given its an
> option being set in core-default, and used by other filesystems?
> >INFO  Configuration.deprecation (Configuration.java:warnOnceIfDeprecated(1182))
> - io.bytes.per.checksum is deprecated. Instead, use dfs.bytes-per-checksum
> I don't see why it should be deprecated. If it isn't what HDFS likes, then
> the code there could be smarter: look for the dfs value, and if not found
> then grab the io.bytes one —warning the user.
> I will volunteer to write this code if I get a promise that someone agrees
> with the premise and is willing to help nurture it in.
> Begin forwarded message:
> From: Steve Loughran <stevel@hortonworks.com<mailto:stevel@hortonworks.com
> >>
> Subject: HADOOP-13636 and io.bytes.per.checksum
> Date: 21 September 2016 at 17:12:00 BST
> To: "hdfs-dev@hadoop.apache.org<mailto:hdfs-dev@hadoop.apache.org>" <
> hdfs-dev@hadoop.apache.org<mailto:hdfs-dev@hadoop.apache.org>>
> I'm getting told off for using the deprecated option: io.bytes.per.checksum
> https://issues.apache.org/jira/browse/HADOOP-13636
> Except: I'm not. FileSystem.getServerDefaults() is, which is used by Trash
> to work out where to delete things.
> It strikes me that the system is inconsitent: HdfsConfiguration is
> deprecating a property that everything else is happy to use; I see it in
> four places in production, and various tests, plus core-default.xml
> Is it really deprecated? If so, are there any volunteers to remove it from
> the codebase, while pulling up the default value into core-default?
> otherwise: how about the complaint is turned off?

  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message