[jira] [Commented] (CASSANDRA-15677) Topology events are not sent to clients if the nodes use the same network interface

["Alan Boudreault (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CASSANDRA-15677?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17104902#comment-17104902 ] 

Alan Boudreault commented on CASSANDRA-15677:
---------------------------------------------

I've unassigned myself in case someone else is looking for a task to pick since I haven't been able to get back to this yet. This ticket only affect a cluster that is using a single interface for all nodes (different ports). From my discussion with Ariel. CASSANDRA-7544 was tested by running manually all dtests with the single interface ccm option. There is actually no specific dtests  for that and it's not something that is run regularly. So to complete this ticket:
 * If there is no way yet to easily run the dtests with the single interface mode, it might be a good idea to add something for this.
 * Check if there is a test for those topology events (expected)
 * Run it and see why it is not failing with the single interface mode
 * Run the entire dtests suite with the single interface mode and ensure everything is OK

> Topology events are not sent to clients if the nodes use the same network interface
> -----------------------------------------------------------------------------------
>
>                 Key: CASSANDRA-15677
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-15677
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Messaging/Client
>            Reporter: Alan Boudreault
>            Priority: Normal
>              Labels: pull-request-available
>             Fix For: 4.0-rc
>
>          Time Spent: 20m
>  Remaining Estimate: 0h
>
> *This bug only happens when the cassandra nodes are configured to use a single network interface (ip) but different ports.  See CASSANDRA-7544.*
> Issue: The topology events aren't sent to clients. The problem is that the port is not taken into account when determining if we send it or not:
> https://github.com/apache/cassandra/blob/trunk/src/java/org/apache/cassandra/transport/Server.java#L624
> To reproduce:
> {code}
> # I think the cassandra-test branch is required to get the -S option (USE_SINGLE_INTERFACE)
> ccm create -n4 local40 -v 4.0-alpha2 -S
> {code}
>  
> Then run this small python driver script:
> {code}
> import time
> from cassandra.cluster import Cluster
> cluster = Cluster()
> session = cluster.connect()
> while True:
>     print(cluster.metadata.all_hosts())
>     print([h.is_up for h in cluster.metadata.all_hosts()])
>     time.sleep(5)
> {code}
> Then decommission a node:
> {code}
> ccm node2 nodetool disablebinary
> ccm node2 nodetool decommission
> {code}
>  
> You should see that the node is never removed from the client cluster metadata and the reconnector started.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@cassandra.apache.org
For additional commands, e-mail: commits-help@cassandra.apache.org