HugeGraph-Computer Configuration Reference

Computer Config Options

The defaults in the tables come from ComputerOptions.java in computer-api; explicit conf/computer.properties overrides are shown as “code default (distribution: actual value)”. Common options such as rpc.* come from HugeGraph Commons; distribution values are listed separately below.

Configuration Sources

Source template: computer/computer-dist/src/assembly/static/conf/computer.properties, with log4j2.xml in the same directory. Maven package copies these into distribution conf/, runtime dependencies into lib/, and built-in algorithm JARs into algorithm/. Templates are maintained in source; there is no separate configuration generator.
Standalone and YARN: startup scripts read distribution conf/computer.properties by default. Override it with bin/start-computer.sh -c <configuration-path>; master and workers need the same job parameters.
Kubernetes Operator: users supply spec.computerConf in the CRD; the Operator writes a ConfigMap mounted as computer.properties. It supplies job ID, worker count, Pod addresses, and etcd when unspecified. Unspecified or zero transfer/RPC ports become 8099/8190; transport.server_host and rpc.server_host become Pod IPs. Job startup scripts substitute only ${POD_IP}, ${HOSTNAME}, ${POD_NAME}, and ${POD_NAMESPACE}.
Kubernetes job images must contain the Computer runtime. Bundle algorithm JARs and select them with jarFile, or use an HTTP(S) remoteJarUri for startup download. See the Computer Quick Start and CRD table below.

Apache downloads provide versioned Computer source archives without separate precompiled binaries. Release 1.7.0 source and distribution names include -incubating-; run mvn clean package -DskipTests in the tagged computer/ project to build them. Current master omits that marker and takes its version from the POM. Source archives cannot be started as built distributions.

Warning

Empty HugeGraph credentials and example MinIO keys below are for local demonstrations. In production, enable Server authentication and authorization, retain Server audit-*.log, use a Computer account with minimum required Server permissions, and configure a Server source IP allowlist. Do not use example MinIO keys in production.

1. Basic Configuration

Core job settings for HugeGraph-Computer.

config option	default value	description
hugegraph.url	http://127.0.0.1:8080	The HugeGraph server URL to load data and write results back.
hugegraph.name	hugegraph	The graph name to load data and write results back.
hugegraph.username	"" (empty)	The username for HugeGraph authentication (leave empty if authentication is disabled).
hugegraph.password	"" (empty)	The password for HugeGraph authentication (leave empty if authentication is disabled).
job.id	local_0001 (packaged: local_001)	The job identifier on YARN cluster or K8s cluster.
job.namespace	"" (empty)	Optional etcd job-key prefix for namespace isolation; distinct from Kubernetes metadata.namespace and not populated automatically by the Operator.
job.workers_count	1	The number of workers for one graph algorithm job. In K8s, this option is set by the Operator.
job.partitions_count	1	The number of partitions for computing one graph algorithm job.
job.partitions_thread_nums	4	The number of threads for partition parallel compute.

2. Algorithm Configuration

Algorithm-specific configuration for computation logic.

config option	default value	description
algorithm.params_class	`ComputerOptions.Null` placeholder class	Required. The class used to pass algorithm parameters before the algorithm runs.
algorithm.result_class	`ComputerOptions.Null` placeholder class	The vertex value class used to store computation results.
algorithm.message_class	`ComputerOptions.Null` placeholder class	The message class passed while computing a vertex.

3. Input Configuration

Configuration for loading input data from HugeGraph or other sources.

3.1 Input Source

config option	default value	description
input.source_type	hugegraph-server	The source type to load input data, allowed values: [‘hugegraph-server’, ‘hugegraph-loader’]. The ‘hugegraph-loader’ means use hugegraph-loader to load data from HDFS or file. If using ‘hugegraph-loader’, please configure ‘input.loader_struct_path’ and ‘input.loader_schema_path’.
input.loader_struct_path	"" (empty)	The structure path for Loader input. It takes effect only when `input.source_type=hugegraph-loader`.
input.loader_schema_path	"" (empty)	The schema path for Loader input. It takes effect only when `input.source_type=hugegraph-loader`.

3.2 Input Splits

config option	default value	description
input.split_size	1048576 (1 MB)	The input split size in bytes.
input.split_max_splits	10000000	The maximum number of input splits.
input.split_page_size	500	The page size for streamed load input split data.
input.split_fetch_timeout	300	The timeout in seconds to fetch input splits.

3.3 Input Processing

config option	default value	description
input.filter_class	org.apache.hugegraph.computer.core.input.filter.DefaultInputFilter	The class to create input-filter object. Input-filter is used to filter vertex edges according to user needs.
input.edge_direction	OUT	The direction of edges to load, allowed values: [OUT, IN, BOTH]. When the value is BOTH, edges in both OUT and IN directions will be loaded.
input.edge_freq	MULTIPLE	The frequency of edges that can exist between a pair of vertices, allowed values: [SINGLE, SINGLE_PER_LABEL, MULTIPLE]. SINGLE means only one edge can exist between a pair of vertices (identified by sourceId + targetId); SINGLE_PER_LABEL means each edge label can have one edge between a pair of vertices (identified by sourceId + edgeLabel + targetId); MULTIPLE means many edges can exist between a pair of vertices (identified by sourceId + edgeLabel + sortValues + targetId).
input.max_edges_in_one_vertex	200	The maximum number of adjacent edges allowed to be attached to a vertex. The adjacent edges will be stored and transferred together as a batch unit.

3.4 Input Performance

config option	default value	description
input.send_thread_nums	4	The number of threads for parallel sending of vertices or edges.

4. Snapshot & Storage Configuration

HugeGraph-Computer supports snapshot functionality to save vertex/edge partitions to local storage or MinIO object storage, enabling checkpoint recovery or accelerating repeated computations.

4.1 Basic Snapshot Configuration

config option	default value	description
snapshot.write	false	Whether to write snapshots of input vertex/edge partitions.
snapshot.load	false	Whether to load from snapshots of vertex/edge partitions.
snapshot.name	"" (empty)	User-defined snapshot name to distinguish different snapshots.

4.2 MinIO Integration (Optional)

MinIO can be used as a distributed object storage backend for snapshots in K8s deployments.

config option	default value	description
snapshot.minio_endpoint	"" (empty)	MinIO service endpoint (e.g., `http://minio:9000`). Required when using MinIO.
snapshot.minio_access_key	minioadmin	MinIO access key for authentication.
snapshot.minio_secret_key	minioadmin	MinIO secret key for authentication.
snapshot.minio_bucket_name	"" (empty)	MinIO bucket name for storing snapshot data.

Usage Scenarios:

Checkpoint Recovery: Resume from snapshots after job failures, avoiding data reloading
Repeated Computations: Load data from snapshots when running the same algorithm multiple times
A/B Testing: Save multiple snapshot versions of the same dataset to test different algorithm parameters

Example: Local Snapshot (in computer.properties):

snapshot.write=true
snapshot.name=pagerank-snapshot-20260201

Example: MinIO Snapshot (in K8s CRD computerConf):

computerConf:
  snapshot.write: "true"
  snapshot.name: "pagerank-snapshot-v1"
  snapshot.minio_endpoint: "http://minio:9000"
  snapshot.minio_access_key: "my-access-key"
  snapshot.minio_secret_key: "my-secret-key"
  snapshot.minio_bucket_name: "hugegraph-snapshots"

5. Worker & Master Configuration

Configuration for worker and master computation logic.

5.1 Master Configuration

config option	default value	description
master.computation_class	org.apache.hugegraph.computer.core.master.DefaultMasterComputation	Master-computation is computation that can determine whether to continue to the next superstep. It runs at the end of each superstep on the master.

5.2 Worker Computation

config option	default value	description
worker.computation_class	org.apache.hugegraph.computer.core.config.Null	The class to create worker-computation object. Worker-computation is used to compute each vertex in each superstep.
worker.combiner_class	org.apache.hugegraph.computer.core.config.Null	Combiner can combine messages into one value for a vertex. For example, PageRank algorithm can combine messages of a vertex to a sum value.
worker.partitioner	org.apache.hugegraph.computer.core.graph.partition.HashPartitioner	The partitioner that decides which partition a vertex should be in, and which worker a partition should be in.

5.3 Worker Combiners

config option	default value	description
worker.vertex_properties_combiner_class	org.apache.hugegraph.computer.core.combiner.OverwritePropertiesCombiner	The combiner can combine several properties of the same vertex into one properties at input step.
worker.edge_properties_combiner_class	org.apache.hugegraph.computer.core.combiner.OverwritePropertiesCombiner	The combiner can combine several properties of the same edge into one properties at input step.

5.4 Worker Buffers

config option	default value	description
worker.received_buffers_bytes_limit	104857600 (100 MB)	The limit bytes of buffers of received data. The total size of all buffers can’t exceed this limit. If received buffers reach this limit, they will be merged into a file (spill to disk).
worker.write_buffer_capacity	52428800 (50 MB)	The initial size of write buffer that used to store vertex or message.
worker.write_buffer_threshold	52428800 (50 MB)	The threshold of write buffer. Exceeding it will trigger sorting. The write buffer is used to store vertex or message.

5.5 Worker Data & Timeouts

config option	default value	description
worker.data_dirs	[jobs]	The directories separated by ‘,’ that received vertices and messages can persist into.
worker.wait_sort_timeout	600000 (10 minutes)	The max timeout (in ms) for message-handler to wait for sort-thread to sort one batch of buffers.
worker.wait_finish_messages_timeout	86400000 (24 hours)	The max timeout (in ms) for message-handler to wait for finish-message of all workers.

6. I/O & Output Configuration

Configuration for output computation results.

6.1 Output Class & Result

config option	default value	description
output.output_class	org.apache.hugegraph.computer.core.output.LogOutput	The class to output the computation result of each vertex. Called after iteration computation.
output.result_name	value	The value is assigned dynamically by #name() of instance created by WORKER_COMPUTATION_CLASS.
output.result_write_type	OLAP_COMMON	The result write-type to output to HugeGraph, allowed values: [OLAP_COMMON, OLAP_SECONDARY, OLAP_RANGE].

6.2 Output Behavior

config option	default value	description
output.with_adjacent_edges	false	Whether to output the adjacent edges of the vertex.
output.with_vertex_properties	false	Whether to output the properties of the vertex.
output.with_edge_properties	false	Whether to output the properties of the edge.

6.3 Batch Output

config option	default value	description
output.batch_size	500	The batch size of output.
output.batch_threads	1	The number of threads used for batch output.
output.single_threads	1	The number of threads used for single output.

6.4 HDFS Output

config option	default value	description
output.hdfs_url	hdfs://127.0.0.1:9000	The HDFS URL for output.
output.hdfs_user	hadoop	The HDFS user for output.
output.hdfs_path_prefix	/hugegraph-computer/results	The directory of HDFS output results.
output.hdfs_delimiter	, (comma)	The delimiter of HDFS output.
output.hdfs_merge_partitions	true	Whether to merge output files of multiple partitions.
output.hdfs_replication	3	The replication number of HDFS.
output.hdfs_core_site_path	"" (empty)	The HDFS core site path.
output.hdfs_site_path	"" (empty)	The HDFS site path.
output.hdfs_kerberos_enable	false	Whether Kerberos authentication is enabled for HDFS.
output.hdfs_kerberos_principal	"" (empty)	The HDFS principal for Kerberos authentication.
output.hdfs_kerberos_keytab	"" (empty)	The HDFS keytab file for Kerberos authentication.
output.hdfs_krb5_conf	/etc/krb5.conf	Kerberos configuration file path.

6.5 Retry & Timeout

config option	default value	description
output.retry_times	3	The retry times when output fails.
output.retry_interval	10	The retry interval (in seconds) when output fails.
output.thread_pool_shutdown_timeout	60	The timeout (in seconds) of output thread pool shutdown.

7. Network & Transport Configuration

Configuration for network communication between workers and master.

7.1 Server Configuration

config option	default value	description
transport.server_host	127.0.0.1	Worker data listener and advertised address. It must be reachable by other workers across hosts; the Operator sets it to the Pod IP.
transport.server_port	0 (distribution: 0; K8s Operator: 8099)	Worker data port. Locally, `0` lets the OS assign a port; the Operator uses fixed port `8099` by default for Pod communication.
transport.server_threads	4	The number of transport threads for server.

7.2 Client Configuration

config option	default value	description
transport.client_threads	4	The number of transport threads for client.
transport.client_connect_timeout	3000	The timeout (in ms) of client connect to server.

7.3 Protocol Configuration

config option	default value	description
transport.provider_class	org.apache.hugegraph.computer.core.network.netty.NettyTransportProvider	The transport provider, currently only supports Netty.
transport.io_mode	AUTO	The network IO mode, allowed values: [NIO, EPOLL, AUTO]. AUTO means selecting the appropriate mode automatically.
transport.tcp_keep_alive	true	Whether to enable TCP keep-alive.
transport.transport_epoll_lt	false	Whether to enable EPOLL level-trigger (only effective when io_mode=EPOLL).

7.4 Buffer Configuration

config option	default value	description
transport.send_buffer_size	0	The size of socket send-buffer in bytes. 0 means using system default value.
transport.receive_buffer_size	0	The size of socket receive-buffer in bytes. 0 means using system default value.
transport.write_buffer_high_mark	67108864 (64 MB)	The high water mark for write buffer in bytes. It will trigger sending unavailable if the number of queued bytes > write_buffer_high_mark.
transport.write_buffer_low_mark	33554432 (32 MB)	The low water mark for write buffer in bytes. It will trigger sending available if the number of queued bytes < write_buffer_low_mark.

7.5 Flow Control

config option	default value	description
transport.max_pending_requests	8	The max number of client unreceived ACKs. It will trigger sending unavailable if the number of unreceived ACKs >= max_pending_requests.
transport.min_pending_requests	6	The minimum number of client unreceived ACKs. It will trigger sending available if the number of unreceived ACKs < min_pending_requests.
transport.min_ack_interval	200	The minimum interval (in ms) of server reply ACK.

7.6 Timeouts

config option	default value	description
transport.close_timeout	10000	The timeout (in ms) of close server or close client.
transport.sync_request_timeout	10000	The timeout (in ms) to wait for response after sending sync-request.
transport.finish_session_timeout	0	The timeout (in ms) to finish session. 0 means using (transport.sync_request_timeout × transport.max_pending_requests).
transport.write_socket_timeout	3000	The timeout (in ms) to write data to socket buffer.
transport.server_idle_timeout	360000 (6 minutes)	The max timeout (in ms) of server idle.

7.7 Heartbeat

config option	default value	description
transport.heartbeat_interval	20000 (20 seconds)	The minimum interval (in ms) between heartbeats on client side.
transport.max_timeout_heartbeat_count	120	The maximum times of timeout heartbeat on client side. If the number of timeouts waiting for heartbeat response continuously > max_timeout_heartbeat_count, the channel will be closed from client side.

7.8 Advanced Network Settings

config option	default value	description
transport.max_syn_backlog	511	The capacity of SYN queue on server side. 0 means using system default value.
transport.recv_file_mode	true	Whether to enable receive buffer-file mode. It will receive buffer and write to file from socket using zero-copy if enabled. Note: Requires OS support for zero-copy (e.g., Linux sendfile/splice).
transport.network_retries	3	The number of retry attempts for network communication if network is unstable.

7.9 Master RPC Configuration

These options come from HugeGraph Commons RPC configuration; the distribution template explicitly sets host and port. Kubernetes Operator advertises the Pod IP and defaults an unspecified or zero port to 8190.

Option	Distribution value	Description
rpc.server_host	127.0.0.1 (K8s: Pod IP)	Master RPC address, reachable by workers.
rpc.server_port	8190 (K8s default: 8190)	Master RPC listener port; allow worker connections in security groups and network policies.

8. Storage & Persistence Configuration

Configuration for HGKV (HugeGraph Key-Value) storage engine and value files.

8.1 HGKV Configuration

config option	default value	description
hgkv.max_file_size	2147483648 (2 GB)	The max number of bytes in each HGKV file.
hgkv.max_data_block_size	65536 (64 KB)	The max byte size of HGKV file data block.
hgkv.max_merge_files	10	The max number of files to merge at one time.
hgkv.temp_file_dir	/tmp/hgkv	This folder is used to store temporary files during the file merging process.

8.2 Value File Configuration

config option	default value	description
valuefile.max_segment_size	1073741824 (1 GB)	The max number of bytes in each segment of value-file.

9. BSP & Coordination Configuration

Configuration for Bulk Synchronous Parallel (BSP) protocol and etcd coordination.

config option	default value	description
bsp.etcd_endpoints	http://localhost:2379 (distribution: http://127.0.0.1:2379)	Comma-separated etcd client endpoints. The Operator uses `INTERNAL_ETCD_URL` only when `computerConf` does not specify this key.
bsp.max_super_step	10 (packaged: 2)	The max super step of the algorithm.
bsp.register_timeout	300000 (packaged: 100000)	The max timeout (in ms) to wait for master and workers to register.
bsp.wait_workers_timeout	86400000 (24 hours)	The max timeout (in ms) to wait for workers BSP event.
bsp.wait_master_timeout	86400000 (24 hours)	The max timeout (in ms) to wait for master BSP event.
bsp.log_interval	30000 (30 seconds)	The log interval (in ms) to print the log while waiting for BSP event.

10. Performance Tuning Configuration

Configuration for performance optimization.

config option	default value	description
allocator.max_vertices_per_thread	10000	Maximum number of vertices per thread processed in each memory allocator.
sort.thread_nums	4	The number of threads performing internal sorting.

11. System Administration Configuration

Kubernetes Operator fills or overrides these options. Avoid overriding them unless customizing the Operator or network. job.namespace is optional user configuration; see the basic configuration table.

Option	Managed by	Description
bsp.etcd_endpoints	K8s Operator	Uses `INTERNAL_ETCD_URL` only when absent from `computerConf`.
transport.server_host	K8s Operator	Overrides with Pod IP; workers must reach each other.
transport.server_port	K8s Operator	Defaults unspecified or zero values to `8099`, not a random port.
job.id	K8s Operator	Set from CRD job ID.
job.workers_count	K8s Operator	Set from CRD `workerInstances`.
rpc.server_host	K8s Operator	Overrides with master Pod IP.
rpc.server_port	K8s Operator	Defaults unspecified or zero values to `8190`.
rpc.remote_url	Computer startup	Removed when configuration is read; do not set as job configuration.

Why these values must match:

BSP/RPC: Must match deployed etcd/RPC services for coordination.
Job settings: Must match the CRD to provide the correct worker count.
Transport: Workers need mutually reachable Pod IPs and ports; the K8s default is 8099.

K8s Operator Config Options

NOTE: Option needs to be converted through environment variable settings, e.g. k8s.internal_etcd_url => INTERNAL_ETCD_URL

config option	default value	description
k8s.auto_destroy_pod	true	Delete the job CR after completion or failure; CR deletion triggers cleanup of associated computing resources.
k8s.close_reconciler_timeout	120	The max timeout (in ms) to close reconciler.
k8s.internal_etcd_url	Code: `http://127.0.0.1:2379`; manifest: `http://hugegraph-computer-operator-etcd.hugegraph-computer-operator-system:2379`	etcd URL used by Operator jobs; the supplied manifest uses the etcd Service address.
k8s.internal_minio_url	Code: `http://127.0.0.1:9000`; manifest: `http://hugegraph-computer-operator-minio.hugegraph-computer-operator-system:9000`	MinIO address supplied to jobs by the Operator; needed only for MinIO snapshots.
k8s.max_reconcile_retry	3	The max retry times of reconcile.
k8s.probe_backlog	50	The maximum backlog for serving health probes.
k8s.probe_port	9892	The port that the controller binds to for serving health probes.
k8s.ready_check_internal	1000	The time interval (ms) of check ready.
k8s.ready_timeout	30000	The max timeout (in ms) of check ready.
k8s.reconciler_count	Code: `Runtime.getRuntime().availableProcessors()`; manifest: `6`	Maximum reconciler thread count.
k8s.resync_period	600000	The minimum frequency at which watched resources are reconciled.
k8s.timezone	Asia/Shanghai	The timezone of computer job and operator.
k8s.watch_namespace	hugegraph-computer-operator-system	Namespace watched for custom resources; also used by the supplied manifest. Use `*` to watch all namespaces.

HugeGraph-Computer CRD

1.7.0 CRD: https://github.com/apache/hugegraph-computer/blob/1.7.0/computer/computer-k8s-operator/manifest/hugegraph-computer-crd.v1.yaml

spec	default value	description	required
algorithmName		The name of algorithm.	true
jobId		The job id.	true
image		Job container image, which must include the Computer runtime. Bundle algorithm JARs or download them with `remoteJarUri`.	true
computerConf		The map of computer config options.	true
workerInstances		The number of worker instances, it will override the ‘job.workers_count’ option.	true
pullPolicy	Kubernetes selects `Always` for `latest`, otherwise `IfNotPresent`, when unset	Explicit values: `Always`, `Never`, or `IfNotPresent`. The CRD has no default; see image pull policy.	false
pullSecrets		The pull-secrets of Image, detail please refer to: https://kubernetes.io/docs/concepts/containers/images/#specifying-imagepullsecrets-on-a-pod	false
masterCpu		The cpu limit of master, the unit can be ’m’ or without unit detail please refer to: https://kubernetes.io/docs/concepts/configuration/manage-resources-containers/#meaning-of-cpu	false
workerCpu		The cpu limit of worker, the unit can be ’m’ or without unit detail please refer to: https://kubernetes.io/docs/concepts/configuration/manage-resources-containers/#meaning-of-cpu	false
masterMemory		The memory limit of master, the unit can be one of Ei、Pi、Ti、Gi、Mi、Ki detail please refer to: https://kubernetes.io/docs/concepts/configuration/manage-resources-containers/#meaning-of-memory	false
workerMemory		The memory limit of worker, the unit can be one of Ei、Pi、Ti、Gi、Mi、Ki detail please refer to: https://kubernetes.io/docs/concepts/configuration/manage-resources-containers/#meaning-of-memory	false
log4jXml		The content of log4j.xml for computer job.	false
jarFile		Path to the algorithm JAR inside the image.	false
remoteJarUri		HTTP(S) algorithm JAR URL, downloaded and loaded by the startup script.	false
jvmOptions		The java startup parameters of computer job.	false
envVars		please refer to: https://kubernetes.io/docs/tasks/inject-data-application/define-interdependent-environment-variables/	false
envFrom		please refer to: https://kubernetes.io/docs/tasks/inject-data-application/define-environment-variable-container/	false
masterCommand	bin/start-computer.sh	The run command of master, equivalent to ‘Entrypoint’ field of Docker.	false
masterArgs	["-r master", “-d k8s”]	The run args of master, equivalent to ‘Cmd’ field of Docker.	false
workerCommand	bin/start-computer.sh	The run command of worker, equivalent to ‘Entrypoint’ field of Docker.	false
workerArgs	["-r worker", “-d k8s”]	The run args of worker, equivalent to ‘Cmd’ field of Docker.	false
volumes		Please refer to: https://kubernetes.io/docs/concepts/storage/volumes/	false
volumeMounts		Please refer to: https://kubernetes.io/docs/concepts/storage/volumes/	false
secretPaths		The map of k8s-secret name and mount path.	false
configMapPaths		The map of k8s-configmap name and mount path.	false
podTemplateSpec		Please refer to: https://kubernetes.io/docs/reference/kubernetes-api/workload-resources/pod-template-v1/#PodTemplateSpec	false
securityContext		Please refer to: https://kubernetes.io/docs/tasks/configure-pod-container/security-context/	false

KubeDriver Config Options

config option	default value	description
k8s.build_image_bash_path		The path of command used to build image.
k8s.enable_internal_algorithm	true	Whether enable internal algorithm.
k8s.framework_image_url	hugegraph/hugegraph-computer:latest	The image url of computer framework.
k8s.image_repository_password		The password for login image repository.
k8s.image_repository_registry		The address for login image repository.
k8s.image_repository_url	hugegraph/hugegraph-computer	The url of image repository.
k8s.image_repository_username		The username for login image repository.
k8s.internal_algorithm	[pageRank]	The name list of all internal algorithm. Note: Algorithm names use camelCase here (e.g., `pageRank`), but algorithm implementations return underscore_case (e.g., `page_rank`).
k8s.internal_algorithm_image_url	hugegraph/hugegraph-computer:latest	The image url of internal algorithm.
k8s.jar_file_dir	/cache/jars/	The directory where the algorithm jar will be uploaded.
k8s.kube_config	~/.kube/config	The path of k8s config file.
k8s.log4j_xml_path		The log4j.xml path for computer job.
k8s.namespace	hugegraph-computer-operator-system	Namespace of the HugeGraph-Computer system.
k8s.pull_secret_names	[]	The names of pull-secret for pulling image.