Open tancho-mihov-1crew opened 1 year ago
Hi,
When you said it only works when replicaCount: 1
you mean you are getting errors ?
Hi,
When replicaCount: 3
as desired first container cassandra-0 runs as expected, but the second one:
Warning Unhealthy 14m (x12379 over 11d) kubelet Liveness probe failed: nodetool: Failed to connect to '127.0.0.1:7199' - ConnectException: 'Connection refused (Connection refused)'.
Warning BackOff 9m45s (x30884 over 11d) kubelet Back-off restarting failed container cassandra in pod cassandra-1_non-production(1f7ffdca-3c83-4ba6-a9d0-fc7e6179021a)
Warning Unhealthy 4m43s (x51536 over 11d) kubelet Readiness probe failed: nodetool: Failed to connect to '127.0.0.1:7199' - ConnectException: 'Connection refused (Connection refused)'.
because cassandra.yaml
is "hard coded" by setting existingConfiguration
and second and third containers gets exact the same cassandra.yaml
but they expected personal values for listen_address
and broadcast_rpc_address
.
Hi, As you indicated currently this way of passing the configuration is not supported if you have more than one replica. I am opening an internal task to study this further and try to look for a fix/workaround. I can not provide an ETA for this, meanwhile maybe you would like to send a PR to fix this. We will be glad to review and merge it.
Setting changes to cassandra.yaml is essential, and I am having trouble doing the same. I would say most production environments use cassandra in cluster mode.
Hi, Thanks for your feedback
Thanks for maintaining the charts, team 🥇. Unfortunately, not being able to use custom cassandra.yaml
in cluster mode is a blocker for me as well.
Name and Version
bitnami/cassandra 10.4.2
What architecture are you using?
amd64
What steps will reproduce the bug?
The issue is same as issue CASSANDRA: ENABLE SASI INDEXES I want to set
sasi_indexes_enabled: true
When I'm using the suggestion "to be used a configmap with your custom configuration cassandra.yaml" it works only whenreplicaCount: 1
I want to use 3 replicas:replicaCount: 3
Because Usage ofexistingConfiguration
overrides any other Cassandra configuration set in the chart. Properties likelisten_address
andbroadcast_rpc_address
expected to be uncial per replica but they are the same as values from the configmap. And it makes any other instances that first one to not run.Are you using any custom parameters or values?
No response
What is the expected behavior?
No response
What do you see instead?
Any other instances that first one to not run
Additional information
No response