You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hugegraph.apache.org by "shirleyStorage (via GitHub)" <gi...@apache.org> on 2024/09/22 06:48:44 UTC
[I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
shirleyStorage opened a new issue, #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667
### Problem Type (问题类型)
struct / logic (架构 / 逻辑设计问题)
### Before submit
- [X] 我已经确认现有的 [Issues](https://github.com/apache/hugegraph/issues) 与 [FAQ](https://hugegraph.apache.org/docs/guides/faq/) 中没有相同 / 重复问题 (I have confirmed and searched that there are no similar problems in the historical issue and documents)
### Environment (环境信息)
- Server Version: 1.0.0 (Apache Release Version)
- Backend: RocksDB x nodes, HDD or SSD
- OS: xx CPUs, xx G RAM, Ubuntu 2x.x / CentOS 7.x
- Data Size: xx vertices, xx edges <!-- (like 1000W 点, 9000W 边) -->
### Your Question (问题描述)
How to disable legacy master-worker mechanism in multi-server
### Vertex/Edge example (问题点 / 边数据举例)
_No response_
### Schema [VertexLabel, EdgeLabel, IndexLabel] (元数据结构)
_No response_
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: dev-unsubscribe@hugegraph.apache.org.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "dosubot[bot] (via GitHub)" <gi...@apache.org>.
dosubot[bot] commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2380548021
<!-- Answer -->
To make a `hugegraph/hugegraph` Docker image from your own HugeGraph source code for backend=rocksdb and start a standalone cluster, follow these steps:
1. **Build the Docker Image:**
First, you need to build the Docker image from your HugeGraph source code. Assuming you have Docker installed and your HugeGraph source code is ready, create a `Dockerfile` in the root of your HugeGraph source directory with the following content:
```dockerfile
FROM openjdk:8-jre
# Set the working directory
WORKDIR /hugegraph
# Copy the HugeGraph source code into the container
COPY . /hugegraph
# Install necessary dependencies and build HugeGraph
RUN ./build.sh
# Expose the port
EXPOSE 8080
# Set the entrypoint
ENTRYPOINT ["./bin/start-hugegraph.sh"]
```
Then, build the Docker image using the following command:
```bash
docker build -t hugegraph/hugegraph .
```
2. **Start the Standalone Cluster:**
You can start the HugeGraph server using Docker with RocksDB as the backend. Use the following `docker run` command to start the server:
```bash
docker run -itd --name=graph -p 8080:8080 hugegraph/hugegraph
```
Optional:
- Use `docker exec -it graph bash` to enter the container to perform operations.
- Use `docker run -itd --name=graph -p 8080:8080 -e PRELOAD="true" hugegraph/hugegraph` to start with a built-in example graph. You can use the RESTful API to verify the result.
- Use `-e PASSWORD=123456` to enable auth mode and set the password for admin.
Alternatively, you can use Docker Compose to manage the deployment. Create a `docker-compose.yaml` file with the following content:
```yaml
version: '3'
services:
graph:
image: hugegraph/hugegraph
ports:
- 8080:8080
```
Then, start the server using Docker Compose:
```bash
docker-compose up -d
```
This will start a standalone HugeGraph server with RocksDB as the backend <sup>[[1]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/hugegraph-dist/docker/README.md)</sup><sup>[[2]](https://github.com/apache/incubator-hugegraph-doc/blob/master/content/en/docs/quickstart/hugegraph-server.md)</sup>.
<!-- Dosu Comment Footer -->
*To continue talking to [Dosu](https://dosu.dev), mention @dosu.*
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "shirleyStorage (via GitHub)" <gi...@apache.org>.
shirleyStorage commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2380588946
@dosu
docker run a 'backend=rocksdb ,standalone' cluster
1. The first time I started properly, the service worked.
2. docker restart hugegraph-Container, error message: The server with name 'server-1' already in cluster.
3. Run docker restart hugegraph-Container again, the service starts normally again. What is the reason?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "dosubot[bot] (via GitHub)" <gi...@apache.org>.
dosubot[bot] commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2365581755
<!-- Answer -->
To disable the legacy master-worker mechanism in a multi-server setup using Apache HugeGraph, you need to set the `server.role` configuration option to "computer". This can be done in your configuration file as follows:
```properties
server.role=computer
```
This setting ensures that the `RoleElectionStateMachine` is not initialized, effectively disabling the legacy master-worker mechanism. Here is the relevant configuration option:
```java
public static final ConfigOption<String> SERVER_ROLE =
new ConfigOption<>(
"server.role",
"The role of nodes in the cluster, available types are " +
"[master, worker, computer]",
allowValues("master", "worker", "computer"),
"master"
);
```
By setting `server.role` to "computer", the `supportRoleElection` will be false, and the role state machine will not be initialized <sup>[[1]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/hugegraph-api/src/main/java/org/apache/hugegraph/config/ServerOptions.java)</sup><sup>[[2]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/hugegraph-api/src/main/java/org/apache/hugegraph/core/GraphManager.java)</sup>.
<!-- Dosu Comment Footer -->
*To continue talking to [Dosu](https://dosu.dev), mention @dosu.*
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "shirleyStorage (via GitHub)" <gi...@apache.org>.
shirleyStorage commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2379662278
@dosu How to run standalone mode in docker after hugegraph is built locally?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "shirleyStorage (via GitHub)" <gi...@apache.org>.
shirleyStorage commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2380543148
@dosu How do I make a hugegraph/hugegraph docker image from my own hugegraph source code？only for backend=rocksdb and start a standalone cluster
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "shirleyStorage (via GitHub)" <gi...@apache.org>.
shirleyStorage commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2407241251
@dosu
I want to deprecated ```server.role=master```, so I can't set ```server.role=master```.I have to set server.role=computer,
, then I set backend=rocksdb , start a standalone cluster.
To request curl -DELETE "http://localhost:8080/graphs/policy/schema/vertexlabels/person
it will return "can't schedule task on non-master server".
How can I fix it by improving the code?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "dosubot[bot] (via GitHub)" <gi...@apache.org>.
dosubot[bot] commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2380397397
<!-- Answer -->
To create a `hugegraph/hugegraph` Docker image from your own HugeGraph source code, follow these steps:
1. **Ensure you have Docker and Maven installed on your system.**
2. **Navigate to the root directory of your HugeGraph source code.**
3. **Create a Dockerfile** in the root directory with the following content:
```dockerfile
#
# Licensed to the Apache Software Foundation (ASF) under one or more
# contributor license agreements. See the NOTICE file distributed with
# this work for additional information regarding copyright ownership.
# The ASF licenses this file to You under the Apache License, Version 2.0
# (the "License"); you may not use this file except in compliance with
# the License. You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#
# Dockerfile for HugeGraph Server
# 1st stage: build source code
FROM maven:3.9.0-eclipse-temurin-11 AS build
COPY . /pkg
WORKDIR /pkg
ARG MAVEN_ARGS
RUN mvn package $MAVEN_ARGS -e -B -ntp -Dmaven.test.skip=true -Dmaven.javadoc.skip=true && pwd && ls -l && rm \
./hugegraph-server/*.tar.gz && rm ./hugegraph-pd/*.tar.gz && rm ./hugegraph-store/*.tar.gz
# 2nd stage: runtime env
# Note: ZGC (The Z Garbage Collector) is only supported on ARM-Mac with java > 13
FROM openjdk:11-slim
COPY --from=build /pkg/hugegraph-server/apache-hugegraph-server-incubating-*/ /hugegraph-server/
LABEL maintainer="HugeGraph Docker Maintainers <de...@hugegraph.apache.org>"
# TODO: use g1gc or zgc as default
ENV JAVA_OPTS="-XX:+UnlockExperimentalVMOptions -XX:+UseContainerSupport -XX:MaxRAMPercentage=50 -XshowSettings:vm" \
HUGEGRAPH_HOME="hugegraph-server"
#COPY . /hugegraph/hugegraph-server
WORKDIR /hugegraph-server/
# 1. Install environment and init HugeGraph Sever
RUN set -x \
&& apt-get -q update \
&& apt-get -q install -y --no-install-recommends --no-install-suggests \
dumb-init \
procps \
curl \
lsof \
vim \
cron \
&& apt-get clean \
&& rm -rf /var/lib/apt/lists/* \
&& service cron start \
&& pwd && cd /hugegraph-server/ \
&& sed -i "s/^restserver.url.*$/restserver.url=http:\/\/0.0.0.0:8080/g" ./conf/rest-server.properties
# 2. Init docker script
COPY hugegraph-server/hugegraph-dist/docker/scripts/remote-connect.groovy ./scripts
COPY hugegraph-server/hugegraph-dist/docker/scripts/detect-storage.groovy ./scripts
COPY hugegraph-server/hugegraph-dist/docker/docker-entrypoint.sh .
RUN chmod 755 ./docker-entrypoint.sh
EXPOSE 8080
VOLUME /hugegraph-server
ENTRYPOINT ["/usr/bin/dumb-init", "--"]
CMD ["./docker-entrypoint.sh"]
```
4. **Build the Docker image** using the following command:
```sh
docker build -t hugegraph/hugegraph .
```
This command will create a Docker image named `hugegraph/hugegraph` from your HugeGraph source code <sup>[[1]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-pd/Dockerfile)</sup><sup>[[2]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/Dockerfile)</sup>.
<!-- Dosu Comment Footer -->
*To continue talking to [Dosu](https://dosu.dev), mention @dosu.*
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "dosubot[bot] (via GitHub)" <gi...@apache.org>.
dosubot[bot] commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2374469722
<!-- Answer -->
To start a minimal cluster with each of the three `server.role` values (`master`, `worker`, `computer`), you need to configure the `server.role` option in the `rest-server.properties` configuration file for each server instance. Here is an example of how you can set up the configuration for each role:
1. **Master Node Configuration:**
```properties
server.id=server-master
server.role=master
restserver.url=http://127.0.0.1:8080
```
2. **Worker Node Configuration:**
```properties
server.id=server-worker
server.role=worker
restserver.url=http://127.0.0.1:8081
```
3. **Computer Node Configuration:**
```properties
server.id=server-computer
server.role=computer
restserver.url=http://127.0.0.1:8082
```
Each configuration file should be saved separately and used to start the respective server instance. The `server.id` and `restserver.url` values should be unique for each server to avoid conflicts.
Refer to the `rest-server.properties` configuration file for additional options that may be relevant to your setup, such as `ssl.keystore_file`, `restserver.max_worker_threads`, and `gremlinserver.url` <sup>[[1]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/hugegraph-api/src/main/java/org/apache/hugegraph/config/ServerOptions.java)</sup><sup>[[2]](https://github.com/apache/incubator-hugegraph-doc/blob/master/content/cn/docs/config/config-option.md)</sup><sup>[[3]](https://github.com/apache/incubator-hugegraph-doc/blob/master/content/en/docs/config/config-option.md)</sup>.
<!-- Dosu Comment Footer -->
*To continue talking to [Dosu](https://dosu.dev), mention @dosu.*
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "dosubot[bot] (via GitHub)" <gi...@apache.org>.
dosubot[bot] commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2380624304
<!-- Answer -->
Yes, if the physical machine suddenly goes down, the `removeSelfServerInfo` method will not be called, and the server info will stay there indefinitely. This is because the `removeSelfServerInfo` method is called during the `close` method, which is executed when the server is shutting down gracefully. If the machine goes down unexpectedly, this method will not be invoked.
```java
public synchronized boolean close() {
this.closed = true;
if (!this.dbExecutor.isShutdown()) {
this.removeSelfServerInfo();
this.call(() -> {
try {
this.tx().close();
} catch (ConnectionException ignored) {
// ConnectionException means no connection established
}
this.graph.closeTx();
return null;
});
}
return true;
}
```
The `removeSelfServerInfo` method is responsible for removing the server info, and it is only called within the `close` method. Therefore, if the machine goes down suddenly, the server info will not be removed <sup>[[1]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/hugegraph-core/src/main/java/org/apache/hugegraph/task/ServerInfoManager.java)</sup>.
<!-- Dosu Comment Footer -->
*To continue talking to [Dosu](https://dosu.dev), mention @dosu.*
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "dosubot[bot] (via GitHub)" <gi...@apache.org>.
dosubot[bot] commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2407595104
<!-- Answer -->
To determine if your HugeGraph cluster is running in standalone mode, you can check the configuration and code logic related to task scheduling and server roles.
1. **Configuration Check**:
- In the `rest-server.properties` file, ensure that the `server.role` is set to a role that supports standalone operation, such as `computer` or `master` for a single-node setup.
2. **Code Logic**:
- In the `TaskManager` class, standalone mode is typically indicated by the use of `StandardTaskScheduler` instead of `DistributedTaskScheduler`. This is determined by the `schedulerType` in `HugeGraphParams`.
```java
case "local":
default: {
TaskScheduler scheduler = new StandardTaskScheduler(
graph,
this.taskExecutor,
this.taskDbExecutor,
this.serverInfoDbExecutor
);
this.schedulers.put(graph, scheduler);
break;
}
```
3. **Single Node Flag**:
- The `ServerInfoManager` class may use a flag like `onlySingleNode` to indicate a single-node setup, affecting task scheduling.
By checking these configurations and code paths, you can determine if your cluster is running in standalone mode <sup>[[1]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/hugegraph-core/src/main/java/org/apache/hugegraph/task/TaskManager.java)</sup>.
<!-- Dosu Comment Footer -->
*To continue talking to [Dosu](https://dosu.dev), mention @dosu.*
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "dosubot[bot] (via GitHub)" <gi...@apache.org>.
dosubot[bot] commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2377135730
<!-- Answer -->
The recommended JDK version for compiling HugeGraph is JDK 1.8.0_111. However, for running HugeGraph-Server, it is recommended to use Java 11 (JDK 11) to ensure better security, especially in production environments or those exposed to the public network <sup>[[1]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-commons/hugegraph-common/src/test/resources/META-INF/MANIFEST.MF)</sup><sup>[[2]](https://github.com/apache/incubator-hugegraph-doc/blob/master/content/en/docs/quickstart/hugegraph-server.md)</sup><sup>[[3]](https://github.com/apache/incubator-hugegraph-doc/blob/master/content/en/docs/changelog/hugegraph-1.3.0-release-notes.md)</sup>.
<!-- Dosu Comment Footer -->
*To continue talking to [Dosu](https://dosu.dev), mention @dosu.*
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "shirleyStorage (via GitHub)" <gi...@apache.org>.
shirleyStorage commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2374420563
@dosu What do the three server.role configurations mean？
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "dosubot[bot] (via GitHub)" <gi...@apache.org>.
dosubot[bot] commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2380591926
<!-- Answer -->
To simulate the situation where restarting a HugeGraph container results in the error "The server with name 'server-1' already in cluster", you can follow these steps:
1. **Initial Setup**: Start a HugeGraph container with the default configuration where `server.id` is set to `server-1`.
2. **Restart the Container**: Restart the HugeGraph container without changing the `server.id`.
Here is an example using Docker to simulate this:
```bash
# Step 1: Start the HugeGraph container
docker run -d --name hugegraph -p 8080:8080 -p 8182:8182 hugegraph/hugegraph
# Step 2: Restart the HugeGraph container
docker restart hugegraph
```
The configuration file `rest-server.properties` contains the following line which sets the server ID:
```properties
server.id=server-1
```
When the container restarts, it will attempt to rejoin the cluster with the same `server.id`, leading to the error "The server with name 'server-1' already in cluster".
According to the HugeGraph documentation, the `server.id` is used for license verification and must be unique within the cluster. This is why restarting the container with the same `server.id` causes the conflict.
Additionally, you can use `docker-compose` to manage HugeGraph and related instances. Here is an example `docker-compose.yml`:
```yaml
version: '3'
services:
server:
image: hugegraph/hugegraph
container_name: server
ports:
- 8080:8080
```
To start the services, use:
```bash
docker-compose up -d
```
This setup ensures that the server configuration is managed consistently, and any restart will follow the same configuration, potentially leading to the same error if `server.id` is not unique <sup>[[1]](https://github.com/apache/incubator-hugegraph-doc/blob/master/content/en/docs/config/config-option.md)</sup><sup>[[2]](https://github.com/apache/incubator-hugegraph-doc/blob/master/content/cn/docs/config/config-option.md)</sup><sup>[[3]](https://github.com/apache/incubator-hugegraph-doc/blob/master/content/en/docs/quickstart/hugegraph-server.md)</sup>.
<!-- Dosu Comment Footer -->
*To continue talking to [Dosu](https://dosu.dev), mention @dosu.*
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "dosubot[bot] (via GitHub)" <gi...@apache.org>.
dosubot[bot] commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2380712735
<!-- Answer -->
The master-worker pattern and the `backend=hstore` pattern in HugeGraph serve different purposes and are implemented differently. Here are the key differences:
### Master-Worker Pattern
1. **Purpose**:
- The master-worker pattern is used for task distribution and coordination within a HugeGraph cluster. It ensures that tasks are efficiently managed and executed across multiple nodes.
2. **Components**:
- **Master Node**: Responsible for coordinating the cluster, managing metadata, and scheduling tasks.
- **Worker Nodes**: Execute tasks as directed by the master node.
- **Task Executors**: Various executors for different types of tasks (e.g., schema tasks, OLAP tasks).
- **Schedulers**: Manage the scheduling and execution of tasks.
3. **Implementation**:
- Implemented in classes like `TaskManager`, `DistributedTaskScheduler`, and `ServerInfoManager`.
- Uses methods like `scheduleOrExecuteJob`, `onAsRoleMaster`, and `onAsRoleWorker` to manage task distribution and execution <sup>[[1]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/hugegraph-core/src/main/java/org/apache/hugegraph/task/TaskManager.java)</sup><sup>[[2]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/hugegraph-core/src/main/java/org/apache/hugegraph/task/ServerInfoManager.java)</sup>.
4. **Role Management**:
- Roles are managed using the `NodeRole` enum, which defines roles like `MASTER`, `WORKER`, and `COMPUTER`.
- Role transitions and state management are handled by the `RoleListener` interface and `RoleElectionStateMachine` <sup>[[3]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/hugegraph-core/src/main/java/org/apache/hugegraph/masterelection/StandardRoleElectionStateMachine.java)</sup>.
### Backend=Hstore Pattern
1. **Purpose**:
- The `backend=hstore` pattern is used for data storage and retrieval in HugeGraph. It defines how data is stored, queried, and managed in the Hstore backend.
2. **Components**:
- **HstoreStore**: Manages the interaction with the Hstore backend, including session management and table operations.
- **HstoreTable**: Provides methods for querying, inserting, deleting, and updating entries in the Hstore backend.
- **HstoreProvider**: Responsible for providing the Hstore backend store and creating new schema, graph, and system stores.
3. **Implementation**:
- Implemented in classes like `HstoreStore`, `HstoreTable`, and `HstoreProvider`.
- Uses methods like `mutate`, `query`, and `registerTableManager` to handle data operations <sup>[[4]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/hugegraph-hstore/src/main/java/org/apache/hugegraph/backend/store/hstore/HstoreStore.java)</sup><sup>[[5]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/hugegraph-hstore/src/main/java/org/apache/hugegraph/backend/store/hstore/HstoreTable.java)</sup><sup>[[6]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/hugegraph-hstore/src/main/java/org/apache/hugegraph/backend/store/hstore/HstoreProvider.java)</sup>.
4. **Data Management**:
- Manages data distribution and storage using the Hstore backend.
- Handles mutations, queries, and table management specific to the Hstore backend.
### Summary
- **Master-Worker Pattern**: Focuses on task distribution and coordination within a cluster, ensuring efficient task execution across multiple nodes.
- **Backend=Hstore Pattern**: Focuses on data storage and retrieval, defining how data is managed in the Hstore backend.
Each pattern addresses different aspects of HugeGraph's functionality, with the master-worker pattern handling task management and the `backend=hstore` pattern handling data storage.
<!-- Dosu Comment Footer -->
*To continue talking to [Dosu](https://dosu.dev), mention @dosu.*
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "dosubot[bot] (via GitHub)" <gi...@apache.org>.
dosubot[bot] commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2380589239
<!-- Answer -->
The error "The server with name 'server-1' already in cluster" on the first restart but not on the second restart is likely due to the `server.id` configuration in the `rest-server.properties` file. The `server.id` is set to `server-1`, and it seems that on the first restart, the server might not have fully released the previous instance, causing a conflict. By the second restart, the previous instance is fully terminated, allowing the new instance to start without issues.
To avoid this, ensure that the server is fully stopped before restarting it. You can also try changing the `server.id` to a unique value if running multiple instances.
Here is the relevant configuration from `rest-server.properties`:
```properties
server.id=server-1
server.role=master
```
Additionally, ensure that the `restserver.url` and `gremlinserver.url` configurations are consistent with the host and port settings in `gremlin-server.yaml` to avoid any potential conflicts during the server startup process.
For example, in `rest-server.properties`:
```properties
restserver.url=http://127.0.0.1:8080
gremlinserver.url=http://127.0.0.1:8182
```
And in `gremlin-server.yaml`:
```yaml
host: 127.0.0.1
port: 8182
```
Moreover, the issue might also be related to how RocksDB handles its sessions and locks. According to the `RocksDBStore` class in HugeGraph, the database sessions are managed and opened using a thread pool. If the previous instance of RocksDB has not fully released its locks or sessions, it could cause the "already in cluster" error. The second restart works because the previous instance has had enough time to release all resources.
To mitigate this, ensure that the RocksDB instance is properly closed before restarting the server. You can also check the logs for any issues related to session handling or lock availability in RocksDB <sup>[[1]](https://github.com/apache/incubator-hugegraph-doc/blob/master/content/en/docs/config/config-guide.md)</sup><sup>[[2]](https://github.com/apache/incubator-hugegraph-doc/blob/master/content/cn/docs/config/config-guide.md)</sup><sup>[[3]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/hugegraph-rocksdb/src/main/java/org/apache/hugegraph/backend/store/rocksdb/RocksDBStore.java)</sup>.
<!-- Dosu Comment Footer -->
*To continue talking to [Dosu](https://dosu.dev), mention @dosu.*
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "dosubot[bot] (via GitHub)" <gi...@apache.org>.
dosubot[bot] commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2380394779
<!-- Answer -->
To run a standalone HugeGraph server locally, you can use the `hugegraph` Docker image. Here is the command to start a HugeGraph server with RocksDB as the backend:
```bash
docker run -itd --name=graph -p 8080:8080 hugegraph/hugegraph
```
This command will start the HugeGraph server in the background, and you can access it at `http://127.0.0.1:8080` <sup>[[1]](https://github.com/apache/incubator-hugegraph-doc/blob/master/content/en/docs/quickstart/hugegraph-server.md)</sup>.
If you prefer to use a startup script, follow these steps:
1. **Update the `hugegraph.properties` file** with the desired backend configuration. For example, for RocksDB:
```properties
backend=rocksdb
serializer=binary
rocksdb.data_path=.
rocksdb.wal_path=.
```
2. **Initialize the database** (required on first startup):
```bash
cd *hugegraph-${version}
bin/init-store.sh
```
3. **Start the server**:
```bash
bin/start-hugegraph.sh
```
This will start the HugeGraph server and connect it to the configured backend storage <sup>[[1]](https://github.com/apache/incubator-hugegraph-doc/blob/master/content/en/docs/quickstart/hugegraph-server.md)</sup>.
<!-- Dosu Comment Footer -->
*To continue talking to [Dosu](https://dosu.dev), mention @dosu.*
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "shirleyStorage (via GitHub)" <gi...@apache.org>.
shirleyStorage commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2380591622
@dosu Can you give an example to simulate this situation
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "dosubot[bot] (via GitHub)" <gi...@apache.org>.
dosubot[bot] commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2365739451
<!-- Answer -->
The legacy master-worker role code design in Apache HugeGraph is implemented using an enumeration called `NodeRole`. This enum defines three roles: `MASTER`, `WORKER`, and `COMPUTER`. Each role is associated with a unique byte code and a name. The `NodeRole` enum also includes methods to check the role type and retrieve the role's code and name.
Here is the relevant code:
```java
public enum NodeRole implements SerialEnum {
MASTER(1, "master"),
WORKER(2, "worker"),
COMPUTER(3, "computer");
private final byte code;
private final String name;
NodeRole(int code, String name) {
assert code < 256;
this.code = (byte) code;
this.name = name;
}
static {
SerialEnum.register(NodeRole.class);
}
@Override
public byte code() {
return this.code;
}
public String string() {
return this.name;
}
public boolean master() {
return this == MASTER;
}
public boolean worker() {
return this == WORKER;
}
public boolean computer() {
return this == COMPUTER;
}
}
```
- **Roles Defined**:
- `MASTER`: Represents the master node with code `1` and name `"master"`.
- `WORKER`: Represents the worker node with code `2` and name `"worker"`.
- `COMPUTER`: Represents the computer node with code `3` and name `"computer"`.
- **Methods**:
- `code()`: Returns the byte code of the role.
- `string()`: Returns the name of the role.
- `master()`: Checks if the role is `MASTER`.
- `worker()`: Checks if the role is `WORKER`.
- `computer()`: Checks if the role is `COMPUTER`.
Additionally, the `RoleListener` interface in the `org.apache.hugegraph.masterelection` package defines methods for handling role changes and errors within the system. This interface includes methods such as `onAsRoleMaster`, `onAsRoleWorker`, `onAsRoleCandidate`, `unknown`, `onAsRoleAbdication`, and `error`, each taking a `StateMachineContext` parameter to manage the state transitions and handle errors.
Here is the relevant code for the `RoleListener` interface:
```java
package org.apache.hugegraph.masterelection;
public interface RoleListener {
void onAsRoleMaster(StateMachineContext context);
void onAsRoleWorker(StateMachineContext context);
void onAsRoleCandidate(StateMachineContext context);
void unknown(StateMachineContext context);
void onAsRoleAbdication(StateMachineContext context);
void error(StateMachineContext context, Throwable e);
}
```
The `RoleElectionStateMachineTest` class provides a test implementation for the role election state machine, demonstrating how roles transition and how the system logs these transitions. The `LogEntry` class within this test captures the epoch, node, and role for each state change, and the `TestConfig` class provides configuration details for the state machine.
Here is an excerpt from the `RoleElectionStateMachineTest` class:
```java
public class RoleElectionStateMachineTest {
private static class LogEntry {
private final Integer epoch;
private final String node;
private final Role role;
enum Role {
master,
worker,
candidate,
abdication,
unknown
}
public LogEntry(Integer epoch, String node, Role role) {
this.epoch = epoch;
this.node = node;
this.role = role;
}
@Override
public boolean equals(Object obj) {
if (this == obj) {
return true;
}
if (!(obj instanceof LogEntry)) {
return false;
}
LogEntry logEntry = (LogEntry) obj;
return Objects.equals(this.epoch, logEntry.epoch) &&
Objects.equals(this.node, logEntry.node) &&
this.role == logEntry.role;
}
@Override
public int hashCode() {
return Objects.hash(this.epoch, this.node, this.role);
}
@Override
public String toString() {
return "LogEntry{" +
"epoch=" + this.epoch +
", node='" + this.node + '\'' +
", role=" + this.role +
'}';
}
}
private static class TestConfig implements Config {
private final String node;
public TestConfig(String node) {
this.node = node;
}
@Override
public String node() {
return this.node;
}
@Override
public String url() {
return "http://127.0.0.1:8080";
}
@Override
public int exceedsFailCount() {
return 2;
}
@Override
public long randomTimeoutMillisecond() {
return 400;
}
@Override
public long heartBeatIntervalSecond() {
return 1;
}
@Override
public int masterDeadTimes() {
return 5;
}
@Override
public long baseTimeoutMillisecond() {
return 100;
}
}
@Test
public void testStateMachine() throws InterruptedException {
final int MAX_COUNT = 200;
CountDownLatch stop = new CountDownLatch(4);
List<LogEntry> logRecords = Collections.synchronizedList(new ArrayList<>(MAX_COUNT));
List<String> masterNodes = Collections.synchronizedList(new ArrayList<>(MAX_COUNT));
RoleListener callback = new RoleListener() {
@Override
public void onAsRoleMaster(StateMachineContext context) {
Integer epochId = context.epoch();
String node = context.node();
logRecords.add(new LogEntry(epochId, node, LogEntry.Role.master));
if (logRecords.size() > MAX_COUNT) {
context.stateMachine().shutdown();
}
Utils.println("master node: " + node);
masterNodes.add(node);
}
@Override
public void onAsRoleWorker(StateMachineContext context) {
Integer epochId = context.epoch();
String node = context.node();
logRecords.add(new LogEntry(epochId, node, LogEntry.Role.worker));
if (logRecords.size() > MAX_COUNT) {
context.stateMachine().shutdown();
}
}
@Override
public void onAsRoleCandidate(StateMachineContext context) {
Integer epochId = context.epoch();
String node = context.node();
logRecords.add(new LogEntry(epochId, node, LogEntry.Role.candidate));
if (logRecords.size() > MAX_COUNT) {
context.stateMachine().shutdown();
}
}
};
}
}
```
This design allows for easy role identification, role-specific logic implementation, and handling of role transitions and errors within the HugeGraph system <sup>[[1]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/hugegraph-core/src/main/java/org/apache/hugegraph/type/define/NodeRole.java)</sup><sup>[[2]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/hugegraph-core/src/main/java/org/apache/hugegraph/masterelection/RoleListener.java)</sup><sup>[[3]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/hugegraph-test/src/main/java/org/apache/hugegraph/core/RoleElectionStateMachineTest.java)</sup>.
<!-- Dosu Comment Footer -->
*To continue talking to [Dosu](https://dosu.dev), mention @dosu.*
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "shirleyStorage (via GitHub)" <gi...@apache.org>.
shirleyStorage commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2407232243
@dosu If I set server.role=computer,and use backend=rocksdb , start a standalone cluster.
Then request ```curl -DELETE "http://localhost:8080/graphs/policy/schema/vertexlabels/person ```
it will return "can't schedule task on non-master server".how to resolve it?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "dosubot[bot] (via GitHub)" <gi...@apache.org>.
dosubot[bot] commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2374498133
<!-- Answer -->
To start a single HugeGraph cluster with `backend=rocksdb`, follow these steps:
1. **Update `hugegraph.properties`**:
```properties
backend=rocksdb
serializer=binary
rocksdb.data_path=.
rocksdb.wal_path=.
```
2. **Initialize the database** (required on first startup or if a new configuration was manually added under 'conf/graphs/'):
```bash
cd *hugegraph-${version}
bin/init-store.sh
```
3. **Update `rest-server.properties`**:
```properties
restserver.url=http://127.0.0.1:8080
gremlinserver.url=http://127.0.0.1:8182
graphs=[hugegraph:conf/hugegraph.properties]
```
4. **Update `gremlin-server.yaml`**:
```yaml
host: 127.0.0.1
port: 8182
graphs: {
hugegraph: conf/hugegraph.properties
}
```
5. **Start the server**:
```bash
bin/start-hugegraph.sh
Starting HugeGraphServer...
Connecting to HugeGraphServer (http://127.0.0.1:8080/graphs)....OK
```
6. **Verify the server and graphs**:
```bash
curl http://127.0.0.1:8080/graphs/
```
This will list the available graphs, including the one configured with RocksDB.
This will start the HugeGraph server with RocksDB as the backend <sup>[[1]](https://github.com/apache/incubator-hugegraph-doc/blob/master/content/en/docs/quickstart/hugegraph-server.md)</sup><sup>[[2]](https://github.com/apache/incubator-hugegraph-doc/blob/master/content/en/docs/config/config-guide.md)</sup><sup>[[3]](https://github.com/apache/incubator-hugegraph-doc/blob/master/content/cn/docs/config/config-guide.md)</sup>.
<!-- Dosu Comment Footer -->
*To continue talking to [Dosu](https://dosu.dev), mention @dosu.*
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "shirleyStorage (via GitHub)" <gi...@apache.org>.
shirleyStorage commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2380668916
@dosu If I set server.role=computer, does it not check if serverinfo exists?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "dosubot[bot] (via GitHub)" <gi...@apache.org>.
dosubot[bot] commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2380581088
<!-- Answer -->
To resolve the `java.lang.IllegalArgumentException: The server with name 'server-1' already in cluster` error when restarting a HugeGraph container, you need to ensure that the server information is properly removed before restarting. The `ServerInfoManager` class has a method `removeSelfServerInfo` that can be used to remove the server information.
Here is a code snippet that demonstrates how to remove the server information before restarting the container:
```java
public synchronized boolean close() {
this.closed = true;
if (!this.dbExecutor.isShutdown()) {
this.removeSelfServerInfo();
this.call(() -> {
try {
this.tx().close();
} catch (ConnectionException ignored) {
// ConnectionException means no connection established
}
this.graph.closeTx();
return null;
});
}
return true;
}
```
Additionally, ensure that the `server.id` in your `rest-server.properties` configuration file is unique for each server in the cluster. The default value is `server-1`, and if multiple servers use the same ID, it will cause conflicts. You can change the `server.id` to a unique value for each server:
```properties
server.id=unique-server-id
```
Ensure that this method is called before the container is restarted to remove the server information and avoid the `IllegalArgumentException`. Also, verify that the `server.id` is unique across your cluster configuration.
Refer to the HugeGraph configuration documentation for more details on the `rest-server.properties` file and other configuration options:
```properties
# rest-server.properties
server.id=server-1
server.role=master
restserver.url=http://127.0.0.1:8080
# Other configurations...
```
By following these steps, you can avoid conflicts and ensure a smooth restart of your HugeGraph container <sup>[[1]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/hugegraph-core/src/main/java/org/apache/hugegraph/task/ServerInfoManager.java)</sup><sup>[[2]](https://github.com/apache/incubator-hugegraph-doc/blob/master/content/en/docs/config/config-option.md)</sup><sup>[[3]](https://github.com/apache/incubator-hugegraph-doc/blob/master/content/cn/docs/config/config-option.md)</sup>.
<!-- Dosu Comment Footer -->
*To continue talking to [Dosu](https://dosu.dev), mention @dosu.*
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "shirleyStorage (via GitHub)" <gi...@apache.org>.
shirleyStorage commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2377131099
@dosu What is the recommended jdk version for compiling hugegraph?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "dosubot[bot] (via GitHub)" <gi...@apache.org>.
dosubot[bot] commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2379665210
<!-- Answer -->
To run HugeGraph in standalone mode using Docker after building it locally, follow these steps:
1. **Build the Docker image**:
Ensure you are in the directory containing the `Dockerfile` and run the following command to build the Docker image:
```sh
docker build -t hugegraph-pd .
```
2. **Run the Docker container**:
Once the image is built, you can run the container in standalone mode using the following command:
```sh
docker run -itd --name=graph -p 8080:8080 hugegraph-pd
```
This command will start the HugeGraph server in a Docker container, exposing port 8080 for access. The `docker-entrypoint.sh` script will be executed as specified in the `Dockerfile`.
Alternatively, you can use Docker Compose to manage the deployment:
1. Create a `docker-compose.yaml` file with the following content:
```yaml
version: '3'
services:
graph:
image: hugegraph-pd
ports:
- 8080:8080
```
2. Run the following command to start the HugeGraph server:
```sh
docker-compose up -d
```
This will start the HugeGraph server and manage other HugeGraph-related instances if needed <sup>[[1]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-pd/Dockerfile)</sup><sup>[[2]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/hugegraph-dist/docker/README.md)</sup>.
<!-- Dosu Comment Footer -->
*To continue talking to [Dosu](https://dosu.dev), mention @dosu.*
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "shirleyStorage (via GitHub)" <gi...@apache.org>.
shirleyStorage commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2407740084
@dosu If I set backend=rocksdb, is there definitely only one server node?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "shirleyStorage (via GitHub)" <gi...@apache.org>.
shirleyStorage commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2380392754
@dosu Why use hugegraph-pd instead of hugegraph-server?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "shirleyStorage (via GitHub)" <gi...@apache.org>.
shirleyStorage commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2380580707
@dosu I ran into the following problem when restarting container after compiling with the latest master branch code:
```
java.lang.IllegalArgumentException: The server with name 'server-1' already in cluster
at com.google.common.base.Preconditions.checkArgument(Preconditions.java:167) ~[guava-31.0.1-android.jar:?]
at org.apache.hugegraph.util.E.checkArgument(E.java:52) ~[hugegraph-common-1.5.0.jar:1.5.0]
at org.apache.hugegraph.task.ServerInfoManager.initServerInfo(ServerInfoManager.java:110) ~[hugegraph-core-1.5.0.jar:1.5.0]
at org.apache.hugegraph.StandardHugeGraph.serverStarted(StandardHugeGraph.java:288) ~[hugegraph-core-1.5.0.jar:1.5.0]
at org.apache.hugegraph.core.GraphManager.serverStarted(GraphManager.java:484) ~[hugegraph-api-1.5.0.jar:0.71.0.0]
at org.apache.hugegraph.core.GraphManager.init(GraphManager.java:125) ~[hugegraph-api-1.5.0.jar:0.71.0.0]
at org.apache.hugegraph.server.ApplicationConfig$GraphManagerFactory$1.onEvent(ApplicationConfig.java:143) ~[hugegraph-api-1.5.0.jar:0.71.0.0]
at org.glassfish.jersey.server.internal.monitoring.CompositeApplicationEventListener.onEvent(CompositeApplicationEventListener.java:49) ~[jersey-server-3.0.3.jar:?]
at org.glassfish.jersey.server.internal.monitoring.MonitoringContainerListener.onStartup(MonitoringContainerListener.java:56) ~[jersey-server-3.0.3.jar:?]
at org.glassfish.jersey.server.ApplicationHandler.onStartup(ApplicationHandler.java:711) ~[jersey-server-3.0.3.jar:?]
at org.glassfish.jersey.grizzly2.httpserver.GrizzlyHttpContainer.start(GrizzlyHttpContainer.java:330) ~[jersey-container-grizzly2-http-3.0.3.jar:?]
at org.glassfish.grizzly.http.server.HttpHandlerChain.start(HttpHandlerChain.java:376) ~[grizzly-http-server-3.0.1.jar:3.0.1]
at org.glassfish.grizzly.http.server.HttpServer.setupHttpHandler(HttpServer.java:268) ~[grizzly-http-server-3.0.1.jar:3.0.1]
at org.glassfish.grizzly.http.server.HttpServer.start(HttpServer.java:245) ~[grizzly-http-server-3.0.1.jar:3.0.1]
at org.apache.hugegraph.server.RestServer.start(RestServer.java:70) ~[hugegraph-api-1.5.0.jar:0.71.0.0]
at org.apache.hugegraph.server.RestServer.start(RestServer.java:177) ~[hugegraph-api-1.5.0.jar:0.71.0.0]
at org.apache.hugegraph.dist.HugeRestServer.start(HugeRestServer.java:31) ~[hugegraph-dist-1.5.0.jar:1.5.0]
at org.apache.hugegraph.dist.HugeGraphServer.<init>(HugeGraphServer.java:60) ~[hugegraph-dist-1.5.0.jar:1.5.0]
at org.apache.hugegraph.dist.HugeGraphServer.main(HugeGraphServer.java:120) ~[hugegraph-dist-1.5.0.jar:1.5.0]
```
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "dosubot[bot] (via GitHub)" <gi...@apache.org>.
dosubot[bot] commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2380639815
<!-- Answer -->
To clear `serverInfo` manually if the server goes down unexpectedly, you can use the `removeServerInfo` method provided in the `ServerInfoManager` class. This method removes the server information for a given server ID.
Here is how you can do it:
```java
// Assuming you have an instance of ServerInfoManager
ServerInfoManager serverInfoManager = ...;
// The ID of the server you want to remove
Id serverId = ...;
// Remove the server info
serverInfoManager.removeServerInfo(serverId);
```
The `removeServerInfo` method will handle the removal of the server information from the backend store.
Supporting evidence from the provided code:
```java
private HugeServerInfo removeServerInfo(Id serverId) {
if (serverId == null) {
return null;
}
LOG.info("Remove server info: {}", serverId);
return this.call(() -> {
Iterator<Vertex> vertices = this.tx().queryServerInfos(serverId);
Vertex vertex = QueryResults.one(vertices);
if (vertex == null) {
return null;
}
this.tx().removeVertex((HugeVertex) vertex);
return HugeServerInfo.fromVertex(vertex);
});
}
```
This method ensures that the server information is properly removed from the backend store <sup>[[1]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/hugegraph-core/src/main/java/org/apache/hugegraph/task/ServerInfoManager.java)</sup>.
<!-- Dosu Comment Footer -->
*To continue talking to [Dosu](https://dosu.dev), mention @dosu.*
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "shirleyStorage (via GitHub)" <gi...@apache.org>.
shirleyStorage commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2380671110
@dosu Does it still trigger `The server with name 'server-1' already in cluster`
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "github-actions[bot] (via GitHub)" <gi...@apache.org>.
github-actions[bot] commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2439735906
Due to the lack of activity, the current issue is marked as stale and will be closed after 20 days, any update will remove the stale label
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "shirleyStorage (via GitHub)" <gi...@apache.org>.
shirleyStorage commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2374465841
@dosu How would you start a minimal cluster with each of the three server.role values?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "dosubot[bot] (via GitHub)" <gi...@apache.org>.
dosubot[bot] commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2380669832
<!-- Answer -->
If you set `server.role=computer`, the system will still check if `serverinfo` exists. The `heartbeat` method in the `ServerInfoManager` class explicitly checks for the existence of `serverinfo` regardless of the server role:
```java
public synchronized void heartbeat() {
assert this.graphIsReady();
HugeServerInfo serverInfo = this.selfServerInfo();
if (serverInfo != null) {
// Update heartbeat time for this server
serverInfo.updateTime(DateUtil.now());
this.save(serverInfo);
return;
}
/* ServerInfo is missing */
if (this.selfNodeId() == null) {
// Ignore if ServerInfo is not initialized
LOG.info("ServerInfo is missing: {}, may not be initialized yet");
return;
}
if (this.selfIsMaster()) {
// On master node, just wait for ServerInfo re-init
LOG.warn("ServerInfo is missing: {}, may be cleared before",
this.selfNodeId());
return;
}
/*
* Missing server info on non-master node, may be caused by graph
* truncated on master node then synced by raft.
* TODO: we just patch it here currently, to be improved.
*/
serverInfo = this.saveServerInfo(this.selfNodeId(), this.selfNodeRole());
assert serverInfo != null;
}
```
This method checks if `serverInfo` is `null` and handles the situation accordingly, indicating that the existence of `serverinfo` is always checked <sup>[[1]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/hugegraph-core/src/main/java/org/apache/hugegraph/task/ServerInfoManager.java)</sup>.
<!-- Dosu Comment Footer -->
*To continue talking to [Dosu](https://dosu.dev), mention @dosu.*
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "shirleyStorage (via GitHub)" <gi...@apache.org>.
shirleyStorage commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2380394315
@dosu But I only need hugegraph server to run standalone hugegraph cluster locally?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "shirleyStorage (via GitHub)" <gi...@apache.org>.
shirleyStorage commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2380624024
@dosu If the physical machine suddenly goes down, this removeSelfServerInfo will not be called and the serverinfo will stay there forever, right?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "github-actions[bot] (via GitHub)" <gi...@apache.org>.
github-actions[bot] closed issue #2667: [Question] How to disable legacy master-worker mechanism in multi-server
URL: https://github.com/apache/incubator-hugegraph/issues/2667
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: dev-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "dosubot[bot] (via GitHub)" <gi...@apache.org>.
dosubot[bot] commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2407742492
<!-- Answer -->
Using `backend=rocksdb` does not inherently limit you to a single server node. However, RocksDB is typically used in standalone setups due to its lack of built-in support for distributed storage. If you want to use multiple nodes, you would need to implement additional mechanisms for clustering, such as using Raft for consensus, which HugeGraph supports <sup>[[1]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-store/hg-store-core/src/main/java/org/apache/hugegraph/store/options/RaftRocksdbOptions.java)</sup>.
<!-- Dosu Comment Footer -->
*To continue talking to [Dosu](https://dosu.dev), mention @dosu.*
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "dosubot[bot] (via GitHub)" <gi...@apache.org>.
dosubot[bot] commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2407233931
<!-- Answer -->
To resolve the error "can't schedule task on non-master server" in a standalone HugeGraph cluster, you need to set the server as the master. This error occurs because tasks can only be scheduled on the master server.
Here's how to fix it:
1. **Update the Configuration**: Set `server.role=master` in your `rest-server.properties` configuration file.
```properties
server.role=master
```
2. **Restart the Server**: After updating the configuration, restart your HugeGraph server to apply the changes.
This will allow task scheduling operations to proceed without encountering the error <sup>[[1]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/hugegraph-core/src/main/java/org/apache/hugegraph/task/StandardTaskScheduler.java)</sup>.
<!-- Dosu Comment Footer -->
*To continue talking to [Dosu](https://dosu.dev), mention @dosu.*
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "shirleyStorage (via GitHub)" <gi...@apache.org>.
shirleyStorage commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2407592009
@dosu How do I know if my cluster is standalone in the code？
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "dosubot[bot] (via GitHub)" <gi...@apache.org>.
dosubot[bot] commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2374424445
<!-- Answer -->
The three `server.role` configurations define the role of nodes in the HugeGraph cluster. The available types are:
1. **master**: This node acts as the master in the cluster, typically responsible for coordinating the cluster and managing metadata.
2. **worker**: This node handles the actual data processing tasks in the cluster.
3. **computer**: This node is used for computational tasks that are distinct from the roles of master and worker nodes.
Here is the relevant configuration option:
```java
public static final ConfigOption<String> SERVER_ROLE =
new ConfigOption<>(
"server.role",
"The role of nodes in the cluster, available types are " +
"[master, worker, computer]",
allowValues("master", "worker", "computer"),
"master"
);
```
By setting the `server.role` configuration, you can define the specific role of each node in your HugeGraph cluster <sup>[[1]](https://github.com/apache/incubator-hugegraph-doc/blob/master/content/cn/docs/config/config-option.md)</sup><sup>[[2]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/hugegraph-api/src/main/java/org/apache/hugegraph/config/ServerOptions.java)</sup>.
<!-- Dosu Comment Footer -->
*To continue talking to [Dosu](https://dosu.dev), mention @dosu.*
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "shirleyStorage (via GitHub)" <gi...@apache.org>.
shirleyStorage commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2380587254
@dosu If the server is killed suddenly. How to execute removeSelfServerInfo? This ServerInfo will always be there?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "dosubot[bot] (via GitHub)" <gi...@apache.org>.
dosubot[bot] commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2380587752
<!-- Answer -->
To ensure `removeSelfServerInfo` is executed when the server is killed suddenly, you should handle the server shutdown gracefully by ensuring that the `close` method is called. The `close` method in `ServerInfoManager` already calls `removeSelfServerInfo` if the `dbExecutor` is not shut down.
Here is the relevant part of the `close` method:
```java
public synchronized boolean close() {
this.closed = true;
if (!this.dbExecutor.isShutdown()) {
this.removeSelfServerInfo();
this.call(() -> {
try {
this.tx().close();
} catch (ConnectionException ignored) {
// ConnectionException means no connection established
}
this.graph.closeTx();
return null;
});
}
return true;
}
```
To handle sudden server termination, you can use a shutdown hook in your application to ensure that the `close` method is called:
```java
Runtime.getRuntime().addShutdownHook(new Thread(() -> {
try {
serverInfoManager.close();
} catch (Exception e) {
e.printStackTrace();
}
}));
```
This shutdown hook will be executed when the JVM is shutting down, ensuring that `removeSelfServerInfo` is called to remove the ServerInfo <sup>[[1]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/hugegraph-core/src/main/java/org/apache/hugegraph/task/ServerInfoManager.java)</sup>.
<!-- Dosu Comment Footer -->
*To continue talking to [Dosu](https://dosu.dev), mention @dosu.*
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "dosubot[bot] (via GitHub)" <gi...@apache.org>.
dosubot[bot] commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2380671816
<!-- Answer -->
Yes, setting `server.role=computer` does not prevent the error "The server with name 'server-1' already in cluster" from being triggered. The system still checks for the existence of `serverinfo` regardless of the server role. If the server info for `server-1` already exists in the cluster, you will encounter this error.
To avoid this issue, ensure that each server in the cluster has a unique `server.id` in the `rest-server.properties` configuration file:
```properties
server.id=unique-server-id
server.role=computer
```
This will prevent conflicts and ensure that each server can join the cluster without triggering the error <sup>[[1]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/hugegraph-core/src/main/java/org/apache/hugegraph/task/ServerInfoManager.java)</sup>.
<!-- Dosu Comment Footer -->
*To continue talking to [Dosu](https://dosu.dev), mention @dosu.*
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "dosubot[bot] (via GitHub)" <gi...@apache.org>.
dosubot[bot] commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2380715442
<!-- Answer -->
Yes, if you disable the master-worker pattern in HugeGraph, there are alternative task scheduling and execution mechanisms available:
1. **StandardTaskScheduler with a Single Node Setup**:
- In a single-node setup, tasks are submitted and executed immediately without the need for master-worker coordination.
- Example:
```java
if (this.serverManager().onlySingleNode() && !task.computer()) {
task.status(TaskStatus.QUEUED);
task.server(this.serverManager().selfNodeId());
this.save(task);
return this.submitTask(task);
}
```
2. **Local Task Scheduler**:
- Manages task execution within the same node without distributing tasks across multiple nodes. Suitable for environments where a single node handles all tasks.
- Example:
```java
TaskScheduler scheduler = new StandardTaskScheduler(
graph,
this.taskExecutor,
this.taskDbExecutor,
this.serverInfoDbExecutor
);
this.schedulers.put(graph, scheduler);
```
3. **DistributedTaskScheduler**:
- Handles tasks in a distributed manner even without the master-worker pattern. Uses different executor services for various types of tasks (e.g., schema tasks, OLAP tasks, Gremlin tasks) and a scheduled thread pool executor for periodic task scheduling.
- Example:
```java
public DistributedTaskScheduler(HugeGraphParams graph,
ScheduledThreadPoolExecutor schedulerExecutor,
ExecutorService taskDbExecutor,
ExecutorService schemaTaskExecutor,
ExecutorService olapTaskExecutor,
ExecutorService gremlinTaskExecutor,
ExecutorService ephemeralTaskExecutor,
ExecutorService serverInfoDbExecutor) {
super(graph, serverInfoDbExecutor);
this.taskDbExecutor = taskDbExecutor;
this.schemaTaskExecutor = schemaTaskExecutor;
this.olapTaskExecutor = olapTaskExecutor;
this.gremlinTaskExecutor = gremlinTaskExecutor;
this.ephemeralTaskExecutor = ephemeralTaskExecutor;
this.schedulerExecutor = schedulerExecutor;
this.closed.set(false);
this.schedulePeriod = this.graph.configuration()
.get(CoreOptions.TASK_SCHEDULE_PERIOD);
this.cronFuture = this.schedulerExecutor.scheduleWithFixedDelay(
() -> {
LockUtil.lock("", LockUtil.GRAPH_LOCK);
try {
this.cronSchedule();
} catch (Throwable t) {
LOG.info("cronScheduler exception graph: {}", this.graphName(), t);
} finally {
LockUtil.unlock("", LockUtil.GRAPH_LOCK);
}
},
10L, schedulePeriod,
TimeUnit.SECONDS);
}
```
These mechanisms ensure that tasks are handled efficiently even when the master-worker pattern is not in use <sup>[[1]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/hugegraph-core/src/main/java/org/apache/hugegraph/task/StandardTaskScheduler.java)</sup><sup>[[2]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/hugegraph-core/src/main/java/org/apache/hugegraph/task/TaskManager.java)</sup><sup>[[3]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/hugegraph-core/src/main/java/org/apache/hugegraph/task/DistributedTaskScheduler.java)</sup>.
<!-- Dosu Comment Footer -->
*To continue talking to [Dosu](https://dosu.dev), mention @dosu.*
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "imbajin (via GitHub)" <gi...@apache.org>.
imbajin commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2365729589
> To disable the legacy master-worker mechanism in a multi-server setup using Apache HugeGraph, you need to set the `server.role` configuration option to "computer". This can be done in your configuration file as follows:
>
> ```ini
> server.role=computer
> ```
>
> This setting ensures that the `RoleElectionStateMachine` is not initialized, effectively disabling the legacy master-worker mechanism. Here is the relevant configuration option:
>
> ```java
> public static final ConfigOption<String> SERVER_ROLE =
> new ConfigOption<>(
> "server.role",
> "The role of nodes in the cluster, available types are " +
> "[master, worker, computer]",
> allowValues("master", "worker", "computer"),
> "master"
> );
> ```
>
> By setting `server.role` to "computer", the `supportRoleElection` will be false, and the role state machine will not be initialized [[1]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/hugegraph-api/src/main/java/org/apache/hugegraph/config/ServerOptions.java?rgh-link-date=2024-09-22T06%3A49%3A54Z)[[2]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/hugegraph-api/src/main/java/org/apache/hugegraph/core/GraphManager.java?rgh-link-date=2024-09-22T06%3A49%3A54Z).
>
> _To continue talking to [Dosu](https://dosu.dev), mention @dosu._
@dosu Thanks for the direct solution, could u also provide more details about the legacy master-worker role code design & usage (As detailed and systematic as possible)?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "dosubot[bot] (via GitHub)" <gi...@apache.org>.
dosubot[bot] commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2407243272
<!-- Answer -->
To allow task scheduling on a server with `server.role=computer` in a standalone HugeGraph cluster using `backend=rocksdb`, you need to modify the code to permit scheduling on a computer node. Here's how you can do it:
1. **Modify the Task Scheduling Check**: Update the logic in the `StandardTaskScheduler` class to allow scheduling on a computer node.
Locate the `schedule` method and modify the master node check to include the computer role:
```java
// Check this is on master or computer for normal task schedule
if (!this.serverManager().selfIsMaster() && !this.serverManager().selfIsComputer()) {
throw new HugeException("Can't schedule task on non-master and non-computer server");
}
```
2. **Ensure Role Checking Method Exists**: Make sure the `ServerInfoManager` class has a method `selfIsComputer()` to verify if the current server role is `computer`.
Example implementation:
```java
public boolean selfIsComputer() {
return this.selfNodeRole() != null && this.selfNodeRole().computer();
}
```
By making these changes, the system will allow task scheduling on a server with the `computer` role, resolving the error you encountered <sup>[[1]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/hugegraph-core/src/main/java/org/apache/hugegraph/task/StandardTaskScheduler.java)</sup><sup>[[2]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-server/hugegraph-core/src/main/java/org/apache/hugegraph/task/ServerInfoManager.java)</sup>.
<!-- Dosu Comment Footer -->
*To continue talking to [Dosu](https://dosu.dev), mention @dosu.*
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "shirleyStorage (via GitHub)" <gi...@apache.org>.
shirleyStorage commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2374489866
@dosu How to start a single cluster with backend=rocksdb？
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "dosubot[bot] (via GitHub)" <gi...@apache.org>.
dosubot[bot] commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2380393297
<!-- Answer -->
HugeGraph PD is used instead of HugeGraph Server because it is a meta server responsible for service discovery, partition information storage, and node scheduling <sup>[[1]](https://github.com/apache/incubator-hugegraph/blob/master/hugegraph-pd/README.md)</sup>.
<!-- Dosu Comment Footer -->
*To continue talking to [Dosu](https://dosu.dev), mention @dosu.*
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "shirleyStorage (via GitHub)" <gi...@apache.org>.
shirleyStorage commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2380396858
@dosu How do I make a hugegraph/hugegraph docker image from my own hugegraph source code？
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "shirleyStorage (via GitHub)" <gi...@apache.org>.
shirleyStorage commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2380639566
@dosu If this is the case, how do I clear serverInfo manually?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "shirleyStorage (via GitHub)" <gi...@apache.org>.
shirleyStorage commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2380715009
@dosu If disable the master-worker pattern, are there other implementations that will work instead?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org
Re: [I] [Question] How to disable legacy master-worker mechanism in multi-server [incubator-hugegraph]
Posted by "shirleyStorage (via GitHub)" <gi...@apache.org>.
shirleyStorage commented on issue #2667:
URL: https://github.com/apache/incubator-hugegraph/issues/2667#issuecomment-2380712342
@dosu What is the difference between hugegraph using master-worker pattern and backend=hstore pattern？
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@hugegraph.apache.org
For additional commands, e-mail: issues-help@hugegraph.apache.org