Azure Marketplace cluster - upgrade to 6.4.0

[djl2]

Is there an ETA on when the Azure Marketplace cluster will be update to deploy 6.4.0? It's still is configured for v6.3.1.

elastic/azure-marketplace

[forloop (Russ Cam)]

Hey @djl2, No ETA on when the template will be updated; we will endeavour to update it as soon as we can.

[djl2]

Thanks Russ.

In the meantime, could you provide some guidance with upgrading the v6.3.1 cluster to v6.4? I had x-pack enabled on all nodes before I attempted a rolling upgrade. I followed this URL for guidance--
https://www.elastic.co/guide/en/elasticsearch/reference/current/rolling-upgrades.html#rolling-upgrades

I just updated the master-0 node to v6.4. Now at ES service start up, this 'waiting' message is thrown in the logs.

[2018-08-27T21:46:03,353][INFO ][o.e.n.Node ] [ctesdmaster-0] started
[2018-08-27T21:46:03,420][INFO ][o.e.x.m.e.l.LocalExporter] waiting for elected master node [{ctesdmaster-1}{5aIDf_OfR6iOXeX-1Vmjzg}{BbeRU7LmSPihTmdU1Bx4wQ}{10.0.0.6}{10.0.0.6:9300}{update_domain=0, fault_domain=0, xpack.installed=true}] to setup local exporter [default_local] (does it have x-pack installed?)

In Kibana Monitoring page is now failing as well. This error appears--

Monitoring Request Error
[search_phase_execution_exception] all shards failed: Check the Elasticsearch
Monitoring cluster network connection or the load level of the nodes.
HTTP 503

Note I also tried updating all the nodes in another cluster (non-rolling) and had similar errors.

Note I'm currently using the x-pack trial version.

UPDATE: I rebooted the master-0 VM. The original 'waiting' message no longer appears in the log, but the error still exists in Kibana. Also, in ES-Head, I get the message "cluster health: red (25 of 27).

Thanks!

[forloop (Russ Cam)]

The process for upgrading a template deployment is no different from an upgrade of Debian package to 6.4.0 running on Ubuntu 16.04 anywhere, but note that the Elasticsearch deployment uses monit to monitor the Elasticsearch process, so Elasticsearch should be stopped with

sudo monit stop elasticsearch

and started with

sudo monit start elasticsearch

Going from 6.3.1 to 6.4.0, you'd be able to perform a rolling upgrade.

With Kibana, you'll need to stop the Kibana service with

sudo service kibana stop

update Kibana with 6.4.0 Debian package, then start again

sudo service kibana start