Windows Server Core and Nano Server images for Elasticsearch.
Ths images uses version 5.6 of Elasticsearch. Later versions do not work with Docker volumes in Windows containers.
They also specify very small memory allocations (see the
ES_JAVA_OPTSenvironment variable in the Dockerfile, so they're suitable for dev. You should override that value for more realistic environments.
Run in the background to start a containerized Elasticsearch node:
docker container run -d -p 9200:9200 -p 9300:9300 --name elasticsearch rajeshmuraleedharan/elasticsearch:nanoserverOther containers in the Docker network can work with the server using hostname elasticsearch and port 9200.
You can query the Elasticsearch API from the container host using its IP address:
$ip = docker container inspect -f '{{ .NetworkSettings.Networks.nat.IPAddress }}' elasticsearch
iwr "http://$($ip):9200"The image uses a Docker volume at c:\data to store the Elasticsearch data. To store data on the host instead, mount the host directory when you run a container:
docker container run -d -p 9200:9200 -p 9300:9300 -v c:\es-data:c:\data --name elasticsearch rajeshmuraleedharan/elasticsearch:nanoserverRunning a container may fail on Windows 10 hosts if they don't have enough RAM. The Hyper-V container doesn't allocate enough memory for the JVM and you'll see an error in the container logs:
Error occurred during initialization of VM
Could not reserve enough space for object heap
This doesn't happen with the Nano Server image, but it does with the Windows Server Core image on Windows 10. You can force the memory allocation using the --memory option. On my 8GB Surface I find 2GB is enough to get Elasticsearch running:
docker container run -d -p 9200:9200 -p 9300:9300 -m 2G --name elasticsearch rajeshmuraleedharan/elasticsearch:windowsservercore
Windows Server 2016 is fine.
Note: this repo is a Fork copy from sixeyed/dockerfiles-windows