[
https://issues.apache.org/jira/browse/CASSANDRA-14162?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17378073#comment-17378073
]
Stefan Miklosovic edited comment on CASSANDRA-14162 at 7/9/21, 1:43 PM:
------------------------------------------------------------------------
The backport contains yet another "feature" which is sorting of the output.
This sorting was firstly introduced in 3.2 in
https://issues.apache.org/jira/browse/CASSANDRA-10464 and since then it is
sorted one way or the other but together with [~e.dimitrova] (thanks for
spotting this) we actually think that it should not be the part of the backport
as that would be the introduction of behaviour which would not address the
original issue this ticket is trying to solve.
Unless we agree that we may introduce sorting (either doing it directly in this
ticket or creating new one just for 3.0.25), we drift towards removing sorting
to be strictly just about solving the long names.
However, I doubt that the introduction of sorting would be "bad". It only
improve the user experience and I can not see any reason why a user would be
suddenly "confused and angry" that we started to sort it for him since 3.0.25.
The commit which deletes sorting is here:
https://github.com/apache/cassandra/pull/1103/commits/415a029657fe9a3873cdd256d43fe2e9ca8be384
was (Author: stefan.miklosovic):
The backport contains yet another "feature" which is sorting of the output.
This sorting was firstly introduced in 3.2 in
https://issues.apache.org/jira/browse/CASSANDRA-10464 and since then it is
sorted one way or the other but together with [~e.dimitrova] (thanks for
spotting this) we actually think that it should not be the part of the backport
as that would be the introduction of behaviour which would not address the
original issue this ticket is trying to solve.
Unless we agree that we may introduce sorting (either doing it directly in this
ticket or creating new one just for 3.0.25), we drift towards removing sorting
to be strictly just about solving the long names.
However, I doubt that the introduction of sorting would be "bad". It only
improve the user experience and I can not see any reason why a user would be
suddenly "confused and angry" that we started to sort it for him since 3.0.25.
> Backport 7950 (Output of nodetool compactionstats and compactionhistory does
> not work well with long keyspace and column family names)
> --------------------------------------------------------------------------------------------------------------------------------------
>
> Key: CASSANDRA-14162
> URL: https://issues.apache.org/jira/browse/CASSANDRA-14162
> Project: Cassandra
> Issue Type: Bug
> Components: Legacy/Tools
> Reporter: Kurt Greaves
> Assignee: Stefan Miklosovic
> Priority: Low
> Fix For: 3.0.25
>
> Attachments: 14162-3.0.patch, Screenshot from 2018-01-11
> 01-02-02.png, Screenshot from 2018-01-11 01-02-46.png, Screenshot from
> 2018-01-11 01-02-51.png
>
>
> Colleagues have had issues with output of listsnapshots/compactionstats
> because of things with really long names. Mostly cosmetic but I see no reason
> we shouldn't backport CASSANDRA-7950 to 3.0. It's practically a bugfix. I've
> attached a patch and a bunch of images to show the relevant commands working
> as intended after applying the patch.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
