[
https://issues.apache.org/jira/browse/CASSANDRA-20104?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17904184#comment-17904184
]
Stefan Miklosovic commented on CASSANDRA-20104:
-----------------------------------------------
[~manish_m_pillai] thank you for the patch. I took a look and a lot of things
may be simplified as it is overly complicated.
(1)
https://github.com/instaclustr/cassandra/commit/cde7f344cba3f71459e18cbaff5c44b84b1b5bb4
I would still expect the tests to be improved. We are actually not testing
anything if there is just one node in nodetool's output. Maybe writing a
distributed in-jvm dtest would be better so we might start e.g. two nodes, add
data in such a way that we would biased it towards one node and then we would
check the load etc ...
Anyway, we should test that the sorting works at least for two nodes, in
general.
> Sort output in nodetool status
> ------------------------------
>
> Key: CASSANDRA-20104
> URL: https://issues.apache.org/jira/browse/CASSANDRA-20104
> Project: Apache Cassandra
> Issue Type: New Feature
> Components: Tool/nodetool
> Reporter: Stefan Miklosovic
> Assignee: Manish M Pillai
> Priority: Normal
> Fix For: 5.x
>
> Time Spent: 10m
> Remaining Estimate: 0h
>
> Lets have this output:
> {code}
> nodetool status
> Datacenter: dc1
> ===============
> Status=Up/Down
> |/ State=Normal/Leaving/Joining/Moving
> -- Address Load Tokens Owns (effective) Host ID
> Rack
> UN 172.19.0.9 385.29 KiB 256 66.6%
> 6d194555-f6eb-41d0-c000-000000000003 rack3
> UN 172.19.0.5 609.07 KiB 256 68.4%
> 6d194555-f6eb-41d0-c000-000000000001 rack1
> UN 172.19.0.8 377.5 KiB 256 65.0%
> 6d194555-f6eb-41d0-c000-000000000002 rack2
> {code}
> This is a very simple case but it will serve well for the explanation.
> What I would like to see is to have this output sorted. If I am looking for a
> particular node, e.g. by IP address, these entries are not sorted by IP in
> general, so when I have 20-nodes cluster and I want to check a particular
> node, it is like a needle in a haystack so I would welcome if they are
> ordered by IP.
> This can be generalized to other columns - sort by Load, for example, to see
> what nodes are the most loaded ones, sort by host id would be cool too. TCM
> introduced kind of predictable node ids so sorting them should give you the
> order in which nodes have been added etc.
> There might be a new switch added like "--sort=ip/id/load/rack". For loads,
> it would sort it from the most loaded one to least loaded one. For IPs and
> IDs, it would sort them from the "lowest" to the "highest" node in terms of
> IP address and ID. We might also group nodes by rack and by node id inside a
> rack.
> This should be done _per DC_.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
