cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Joel Knighton (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-9692) Print sensible units for all log messages
Date Tue, 15 Mar 2016 21:09:33 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-9692?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15196264#comment-15196264
] 

Joel Knighton commented on CASSANDRA-9692:
------------------------------------------

Thanks [~giampaolo] - if you want to ping me on JIRA and the autocomplete isn't working, you
can use {{jkni}} for my username.

This patch looks good - I pushed another review commit on my branch at [9692-trunk|https://github.com/jkni/cassandra/tree/9692-trunk].
This fixes some minor whitespace/formatting issues to bring the patch in line with project
conventions and proposes one fairly major change described in the following paragraph.

I'm not sure about the return value for 0 divisor in FBUtilities.bytesPerSeconds. While returning
-1 might make sense for statistics, I'm a little wary of putting a method like that in FBUtilities
where it is likely to be used for others, possibly for calculations where we would rather
throw an exception. Instead, I propose two rate pretty printing functions - one that takes
memory and time and returns NaN if time is zero, and one takes a rate calculated elsewhere.
This removes the need to choose one way to handle bytesPerSeconds 0 divisors and allows each
caller to calculate the rate as needed. At the same time, it sanely handles the common case
of pretty printing a rate that needs no further calculation. If you have any concerns with
this approach, let me know. I've also renamed the prettyPrintRateInSeconds method name to
prettyPrintMemoryPerSecond to better reflect its purpose.

Since most of these changes seem likely to be agreed upon, I've pushed the patch for a first
round of CI.
||branch||testall||dtest||
|[9692-trunk|https://github.com/jkni/cassandra/tree/9692-trunk]|[testall|http://cassci.datastax.com/view/Dev/view/jkni/job/jkni-9692-trunk-testall]|[dtest|http://cassci.datastax.com/view/Dev/view/jkni/job/jkni-9692-trunk-dtest]|



> Print sensible units for all log messages
> -----------------------------------------
>
>                 Key: CASSANDRA-9692
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-9692
>             Project: Cassandra
>          Issue Type: Improvement
>            Reporter: Benedict
>            Assignee: Giampaolo
>            Priority: Minor
>              Labels: lhf
>             Fix For: 3.x
>
>         Attachments: Cassandra9692-Rev1-trunk-giampaolo-trapasso-at-radicalbit-io.diff,
Cassandra9692-trunk-giampaolo-trapasso-at-radicalbit-io.diff
>
>
> Like CASSANDRA-9691, this has bugged me too long. it also adversely impacts log analysis.
I've introduced some improvements to the bits I touched for CASSANDRA-9681, but we should
do this across the codebase. It's a small investment for a lot of long term clarity in the
logs.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message