The only difference I found is that the new nodes have been installed with
Debian-12 and OpenJDK-17 while the rest of the cluster is running on Debian-11
with OpenJDK-11.
> On 9. May 2025, at 07:34, Tibor Répási wrote:
>
> Rechecked that, the upgrade was complete, all nodes are runn
the database.
>
> https://github.com/apache/cassandra/blob/trunk/src/java/org/apache/cassandra/net/OutboundConnectionInitiator.java#L451-L467
>
> – Scott
>
>> On May 8, 2025, at 5:18 AM, Tibor Répási wrote:
>>
>>
>> Hi all,
>>
>> Recently
Hi all,
Recently I face some issues when a node is joining an already existing cluster,
the bootstrap ends up incomplete and in failed state due to some streams have
failed.
Using Cassandra 5.0.3 running on OpenJDK 17.0.15:
Logs show:
WARN [NonPeriodicTasks:1] 2025-05-08 13:23:26,145 StreamR
available anymore. The question now is whether this is a bug or do we have
another way to monitor the process?
Best regards,
Tibor Répási
That doesn’t sound like a RHEL issue. Swapping can severely impact performance,
that’s why it is recommended to turn off swap on Cassandra nodes. This was
recommended in the Cassandra documentation by DataStax[1]. Apache Cassandra is
testing that and log a warning on startup [2] if swap is found
Usually, it’s a good practice to resemble the real datacenter in the Cassandra
topology, thus nodes mounted to distinct racks are know with different rack
names to Cassandra. This is due to the usual datacenter infrastructure, having
a single point of failure in each rack - e.g. a network switch
