[BUG] Segmentation fault with Graphar load

[atberium]

Describe the bug I've got an "SIGSEGV" error when try to load graph in graphar format

To Reproduce

1. Ensure you are running GAE in k8s cluster
2. Ensure all necessary files were mounted properly. At least, you have got all required graphar files and them were mounted properly. In my further example I mount my host path using the following configuration from "graphscope" helm chart:

   volumes:
   enabled: true
   items:
   data:
     type: hostPath
     field:
       type: Directory
       path: {{ graphscope_directories_store_host }}
     mounts:
     - mountPath: {{ graphscope_directories_store_mounted }}

3. Make a python script and add the following code:

   
   import graphscope
   from graphscope import Graph
   from graphscope.framework.loader import Loader

Resolve host and port, using https://graphscope.io/docs/latest/deployment/deploy_graphscope_with_helm#installation tutorial, or hardcode 'em

host="" port=""

session = graphscope.session( addr=f"{host}:{port}", k8s_namespace="{{ k8s_graphscope_namespace }}" )

With my example I used https://github.com/apache/incubator-graphar-testing/tree/955596c325ceba7b607e285738e3dd0ce4ff424e/neo4j files. But the issue could be reproduced with any others from this repository

uri = "graphar+file://{{ graphscope_directories_store_mounted }}/graphar/MovieGraph.graph.yml" graph = Graph.load_from(uri, session) session.close()

4. Run script
6. See error:

Aborted at 1731639742 (unix time) try "date -d @1731639742" if you are using GNU date PC: @ 0x0 (unknown) SIGSEGV (@0x8) received by PID 296 (TID 0x7c1707e00640) from PID 8; stack trace: @ 0x7c17157db046 (unknown) @ 0x7c1712f13520 (unknown) Aborted at 1731639742 (unix time) try "date -d @1731639742" if you are using GNU date @ 0x7c17172e425f _ZZN8vineyard17GARFragmentLoaderIlmNS_14ArrowVertexMapEE18constructVertexMapEvENKUliE_clEi PC: @ 0x0 (unknown) SIGSEGV (@0x8) received by PID 324 (TID 0x77a733400640) from PID 8; stack trace: @ 0x7c17173182a7 vineyard::GARFragmentLoader<>::constructVertexMap() @ 0x77a7416bc046 (unknown) @ 0x7c171731b797 vineyard::GARFragmentLoader<>::LoadFragment() @ 0x77a73edf4520 (unknown) @ 0x7c171731c140 vineyard::GARFragmentLoader<>::LoadFragmentAsFragmentGroup() @ 0x7c17074e1451 LoadGraph @ 0x58967b75c0d1 gs::GrapeInstance::loadGraph() @ 0x58967b761d65 gs::GrapeInstance::OnReceive() @ 0x58967b7f018a gs::Dispatcher::processCmd() @ 0x58967b7f1ebe gs::Dispatcher::subscriberLoop() @ 0x77a7431c525f _ZZN8vineyard17GARFragmentLoaderIlmNS_14ArrowVertexMapEE18constructVertexMapEvENKUliE_clEi @ 0x7c17131f8253 (unknown) @ 0x77a7431f92a7 vineyard::GARFragmentLoader<>::constructVertexMap() @ 0x7c1712f65ac3 (unknown) @ 0x7c1712ff7850 (unknown) @ 0x77a7431fc797 vineyard::GARFragmentLoader<>::LoadFragment() @ 0x77a7431fd140 vineyard::GARFragmentLoader<>::LoadFragmentAsFragmentGroup() @ 0x77a7334e1451 LoadGraph @ 0x5cb0ae4510d1 gs::GrapeInstance::loadGraph() @ 0x5cb0ae456d65 gs::GrapeInstance::OnReceive() @ 0x5cb0ae4e518a gs::Dispatcher::processCmd() @ 0x5cb0ae4e7906 gs::Dispatcher::publisherLoop() @ 0x77a73f0d9253 (unknown) @ 0x77a73ee46ac3 (unknown) @ 0x77a73eed8850 (unknown)

Primary job terminated normally, but 1 process returned a non-zero exit code. Per user-direction, the job has been aborted.

---

mpirun noticed that process rank 1 with PID 296 on node gs-engine-gs-gae-1 exited on signal 11 (Segmentation fault).

RPC failed: rpc _run_step_impl failed: status <_MultiThreadedRendezvous of RPC that terminated with: status = StatusCode.UNAVAILABLE details = "Socket closed. The traceback is: Traceback (most recent call last): File "/home/graphscope/.local/lib/python3.10/site-packages/gscoordinator/servicer/graphscope_one/service.py", line 245, in _RunStep head, bodies = self._operation_executor.run_on_analytical_engine( File "/home/graphscope/.local/lib/python3.10/site-packages/gscoordinator/monitor.py", line 191, in runOnAnalyticalEngineWarp res = func(instance, dag_def, dag_bodies, loader_op_bodies) File "/home/graphscope/.local/lib/python3.10/site-packages/gscoordinator/op_executor.py", line 178, in run_on_analytical_engine response_head, response_bodies = self.run_step(dag_def, dag_bodies) File "/home/graphscope/.local/lib/python3.10/site-packages/gscoordinator/op_executor.py", line 106, in run_step for response in responses: File "/home/graphscope/.local/lib/python3.10/site-packages/grpc/_channel.py", line 543, in next return self._next() File "/home/graphscope/.local/lib/python3.10/site-packages/grpc/_channel.py", line 969, in _next raise self grpc._channel._MultiThreadedRendezvous: <_MultiThreadedRendezvous of RPC that terminated with: status = StatusCode.UNAVAILABLE details = "Socket closed" debug_error_string = "UNKNOWN:Error received from peer {grpc_message:"Socket closed", grpc_status:14, created_time:"2024-11-15T03:02:22.646397916+00:00"}"

" debug_error_string = "UNKNOWN:Error received from peer {grpc_message:"Socket closed. The traceback is: Traceback (most recent call last):\n File \"/home/graphscope/.local/lib/python3.10/site-packages/gscoordinator/servicer/graphscope_one/service.py\", line 245, in _RunStep\n head, bodies = self._operation_executor.run_on_analytical_engine(\n File \"/home/graphscope/.local/lib/python3.10/site-packages/gscoordinator/monitor.py\", line 191, in runOnAnalyticalEngineWarp\n res = func(instance, dag_def, dag_bodies, loader_op_bodies)\n File \"/home/graphscope/.local/lib/python3.10/site-packages/gscoordinator/op_executor.py\", line 178, in run_on_analytical_engine\n response_head, response_bodies = self.run_step(dag_def, dag_bodies)\n File \"/home/graphscope/.local/lib/python3.10/site-packages/gscoordinator/op_executor.py\", line 106, in run_step\n for response in responses:\n File \"/home/graphscope/.local/lib/python3.10/site-packages/grpc/_channel.py\", line 543, in next\n return self._next()\n File \"/home/graphscope/.local/lib/python3.10/site-packages/grpc/_channel.py\", line 969, in _next\n raise self\ngrpc._channel._MultiThreadedRendezvous: <_MultiThreadedRendezvous of RPC that terminated with:\n\tstatus = StatusCode.UNAVAILABLE\n\tdetails = \"Socket closed\"\n\tdebug_error_string = \"UNKNOWN:Error received from peer {grpc_message:\"Socket closed\", grpc_status:14, created_time:\"2024-11-15T03:02:22.646397916+00:00\"}\"\n>\n", grpc_status:14, created_time:"2024-11-15T03:02:22.649210409+00:00"}"

Expected behavior No errors occurred. Screenshots All running pods

Environment:

• GraphScope version: v0.29.0
• OS: Ubuntu
• Version 24.04
• Kubernetes Version 1.28.14
• Python version: 3.11.10 (with following dependencies: graphscope==0.29.0, graphscope-client==0.29.0, pandas==2.0.3, aiohttp, async_timeout)

Additional context The same issue appeared if I run python script on empty cluster (when graphscope library creates all required pods, svc, etc. by itself) like this:

session = graphscope.session(
    k8s_image_registry="{{ docker_image_repository_hosted }}",
    k8s_image_repository="rnd-grapher/graphscope",
    k8s_vineyard_image="{{ docker_image_repository }}/vineyardcloudnative/vineyardd:latest",
    k8s_namespace="{{ k8s_graphscope_namespace }}",
    k8s_volumes={
        "data": {
            "type": "hostPath",
            "field": {
                "path": os.path.expanduser("~/examples/"),
                "type": "Directory"
            },
            "mounts": {
                "mountPath": "/examples/"
            }
        }
    }
)