Running Elasticsearch in Azure Container Instances - elasticsearch

Question

I'm trying to run ElasticSearch in Azure Container Instances. I've created the container like this using the Azure CLI:
az container create --image elasticsearch:7.4.2 --name $containerGroupName -g $resourceGroup --ip-address public --dns-name-label <mydns> --memory 8 --cpu 2 --ports 9200
The container ends up in a waiting state. When I check out the logs in the Azure portal, I'll see the following error:
ERROR: [2] bootstrap checks failed
[1]: max virtual memory areas vm.max_map_count [65530] is too low, increase to at least [262144]
Apparently, this has something to do with virtual memory: https://www.elastic.co/guide/en/elasticsearch/reference/current/vm-max-map-count.html
But how do I change it inside the container? I cannot connect to it since the container is not running.

Answer 1

Very late to the party but hopefully this will help someone in the future. Elasticsearch if deployed stand-alone requires the environment variable discovery.type to be set. Since Azure does not allow vars with a . (dot) at the moment you'll have to create a custom Docker image and host this e. g. with Azure Container Registry or Dockerhub, ...
Content can be as little as this:
FROM elasticsearch:7.8.1
ENV discovery.type=single-node

Answer 2

I realize that ACI is not the best option to run ElasticSearch. I only need it temporarily for a proof-of-concept while I'm waiting for a stable environment of ElasticSearch elsewhere.
Eventually I got it running by picking an older image of ElasticSearch:
az container create --image elasticsearch:5.6.14-alpine --name $containerGroupName -g $resourceGroup --ip-address public --dns-name-label <mydns> --memory 4 --cpu 2 --ports 9200

Answer 3

The solution to this is simply adding this line to the bound elasticsearch.yml config file:
discovery.type: single-node
For this to be permanent even when you restart the container, the config directory of the elasticsearch container needs to be bound to an Azure Fileshare directory, where you make permanent changes there.
Create a Azure file share and create a elasticsearch folder with another config folder inside it: https://docs.microsoft.com/en-us/azure/storage/files/storage-how-to-create-file-share
Deploy a elasticsearch container with the fileshare volume mounted and bind /mnt/elasticsearch/config to a folder you've created on the new fileshare (template tags: "mountPath": "/mnt/elasticsearch/config", "shareName": "elasticsearch/config"): https://docs.microsoft.com/en-us/azure/container-instances/container-instances-volume-azure-files
While the container doesn't error out and shuts itself down because of this max virtual memory error, copy the entire /usr/share/elasticsearch/config folder and filers to the fileshare folder. As soon as you get terminal access, run cp -R /usr/share/elasticsearch/config /mnt/elasticsearch/config
On your file share folder, you should now have a config directory with files and other folders created by the elastisearch startup process. One of this files is elasticsearch.yml. Open it, add the line discovery.type: single-node and save it.
Finally, change mounting and binding location to the correct location, so elasticsearch starts up and reads our now modified configurations. Bind /usr/share/elasticsearch/config to the fileshare /elasticsearch/config folder (template tags: "mountPath": "/usr/share/elasticsearch/config", "shareName": "elasticsearch/config"). More info here: https://www.elastic.co/guide/en/elasticsearch/reference/master/docker.html#docker-configuration-methods
Start the container with the new mounting locations and now the max virtual memory error should be gone.
Please, only use this trick for proofs-of-concept or very low volume logs. Using network storage like Azure Fileshare in general is not suitable for Elasticsearch and can even crash it.