AspNetCore.IgniteServer (extracted from Tarzan and improved)
AspNetCore.IgniteServer is an implementation of Apache Ignite 2 embedded mode that references necessary assemblies to provide a server-side component for the ASP.NET Core platform. Kubernetes ready, authentication and SSL enabled with auto-discovery clustering capabilities. Usually, every node runs one or more instances of the server.
The server can be executed with various options:
dotnet AspNetCore.IgniteServer.dll [options]
Options:
-?|-Help Show help information
-ConfigFile XML configuration file. If not specified then default configuration is used.
-OffHeap Size of off-heap memory given in megabytes.
-OnHeap Size of on-heap memory given in megabytes.
-SpiPort Specifies port for Discovery Spi.
-Cluster Specifies IP address and port of a cluster node. Multiple nodes can be specified.
-ConsistentId Specifies as a consistent id of the node. This value is used in topology.
The default configuration uses the default parameters of the Ignite environment.
There are two predefined configuration files available in config folder:
default-config.xml- this is minimal configuration for running memory only Ignite server.persistent-config.xml- this is minimal configuration for running persistent Ignite server.
- Runs the server locally with the specified amount of memory. It uses 1G of on-heap memory (for computation and queries) and 2G of off-heap memory (data and indexes). It uses default settings, e.g., the Spi port will be the first available port starting from 47500.
dotnet AspNetCore.IgniteServer.dll -Onheap 1024 -Offheap 2048 -Cluster "127.0.0.1:47500"
If the server is run in persistent mode, the situation is slightly more complicated and depends on whether the cluster is run for the first time or it is resumed. It is because, for running a cluster with persistence enabled, Ignite needs to form a topology before the cluster is activated. If the cluster is created for the first time, the cluster has to be activated when it reaches the required topology. The cluster records the topology information, which is used when the cluster is resumed.
memory calculation: ON_HEAP_MEMORY (JVM Heap max size)
+ 256MB (JVM Metaspace max size)
+ 240MB (JVM ReservedCodeCache size)
+ OFF_HEAP_MEMORY (default region max size; data + indexes (basic use cases will add a 30% increase on top of data))
+ MIN(256MB, OFF_HEAP_MEMORY) (OFF_HEAP_MEMORY < 1 GB)
|| OFF_HEAP_MEMORY/4 (OFF_HEAP_MEMORY between 1 GB and 8 GB)
|| 2GB (OFF_HEAP_MEMORY > 8 GB)
(default region checkpointing buffer size)
+ 100MB (sysMemPlc region max size)
+ 100MB (metastoreMemPlc region max size)
+ 100MB (TxLog region max size)
+ 100MB (volatileDsMemPlc region max size)
+ ~512MB (other overheads; e.g., thread stacks, GC, symbols, etc.)
docker build -t dradoaica/aspnetcore.ignite-server:2.17.0 -f Dockerfile .
docker push dradoaica/aspnetcore.ignite-server:2.17.0
docker run -p 0.0.0.0:10800:10800/tcp --name aspnetcore.ignite-server -d dradoaica/aspnetcore.ignite-server:2.17.0
docker rm -f aspnetcore.ignite-server
docker rmi dradoaica/aspnetcore.ignite-server:2.17.0
https://github.com/dradoaica/helm-charts/tree/main/charts/aspnetcore-ignite-server