Failed to reconfigure Kibana in ECE 1.1.3

[jmtrc (Jin)]

I was trying to add an xpack option to kibana via ECE admin console:

xpack.reporting.queue.timeout: 60000

However, kibana does not come back and reports:

[2018-05-17 18:40:42,682][INFO ][no.found.cluster.plan.kibana.6cce5e34496841a0bc4123508dfdce79] Waiting for ephemeral nodes to disappear for instance [instance-0000000002]: [List((inspect,1550483376799,1550483376799,1526581524697,1526581524697,0,0,0,720575942518046807,5234,0,1550483376799

[2018-05-17 18:40:45,305][INFO ][no.found.runner.docker.ContainerInspector] Container not found. {"ec_container_kind":"kibana","ec_container_group":"6cce5e34496841a0bc4123508dfdce79","ec_container_name":"instance-0000000002"}

Then I created a brand new cluster and applied the same configuration to kibana. It was a success.

Does these two messages mean anything?

Thanks.

Jin.

[jmtrc (Jin)]

ECE admin console shows: "Waiting until nodes are deallocated".

[Yuri (Yuri Tceretian)]

Hi Jin,
when you change a cluster's configuration, ECE applies a plan that causes cluster nodes to be re-created. That's why you see the message Waiting until nodes are deallocated. You need to wait until plan finishes to see changes. If it does not work or plan gets stuck on that message, please contact our support team that will help you troubleshoot the issue.

[jmtrc (Jin)]

It was stuck at this message for hours.

I tried to increase memory allocation from 4GB to 8GB. Then kibana took the xpack timeout option this time.

Thanks.

[Yuri (Yuri Tceretian)]

What OS and version of Docker do you use?

[jmtrc (Jin)]

CentOS 7.3 with stock kernel and docker 1.11.