[
https://issues.apache.org/jira/browse/CASSANDRA-14190?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16341772#comment-16341772
]
Samuel Fink commented on CASSANDRA-14190:
-----------------------------------------
I've uploaded an updated patch which adds the try/catch around getSeeds in a
similar way to how [~jjirsa] did it in the CASSANDRA-13018 patch and updated
the unit test to also exercise the exception handling code path.
I changed the return for reloadseeds to be a list (or null on failure) to make
user facing error reporting more feasible now that the reload operation catches
the likely errors. The user facing messages in both getseeds and reloadseeds
have been improved to be more intuitive in the case where there are no
non-localhost seed nodes.
> Non-disruptive seed node list reload
> ------------------------------------
>
> Key: CASSANDRA-14190
> URL: https://issues.apache.org/jira/browse/CASSANDRA-14190
> Project: Cassandra
> Issue Type: Improvement
> Components: Configuration, Lifecycle
> Reporter: Samuel Fink
> Assignee: Samuel Fink
> Priority: Minor
> Fix For: 4.x
>
> Attachments: 14190-trunk.patch, 14190-trunk.patch, 14190-trunk.patch
>
>
> Add a mechanism for reloading the Gossiper in-memory seed node IP list
> without requiring a service restart.
> The Gossiper keeps an in-memory copy of the seed node IP list and uses it
> during a gossip round to determine if the random node that was gossiped to is
> a seed node and for picking a seed node to gossip to in maybeGossipToSeed.
> Currently the Gossiper seed node list is only updated when an endpoint is
> removed, at the start of a shadow round, and on startup. Those scenarios
> don’t handle the case of seed nodes changing IP addresses (eg. DHCP lease
> changes) or additional seed nodes being added to the cluster.
> As described in CASSANDRA-3829 the current way to ensure that all nodes in
> the cluster have the same seed node list when there has been a change is to
> do a rolling restart of every node in the cluster. In large clusters rolling
> restarts can be very complicated to manage and can have performance impacts
> because the caches get flushed.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
