关于TEEU 上手指南仿真模式中的TEEU 安全聚合示例的疑问

[beinnnn]

Issue Type

Build/Install

Source

binary

Secretflow Version

secretflow 1.5.0

OS Platform and Distribution

Ubuntu 20

Python version

3.10

Bazel version

No response

GCC/Compiler version

No response

What happend and What you expected to happen.

运行该示例时，在一台机器上起了三个容器，每个容器分配一个ip地址以代表alice、bob、carol三人，每个容器里面启动一个ray集群，想请问这种部署方式是否可行？
实际运行时，alice和bob无法连接到carol，carol能ping通alice和bob，但是运行不出结果。

Reproduction code to reproduce the issue.

Alice和Bob的报错信息：
(raylet) [2024-04-17 02:39:59,135 E 326 366] (raylet) file_system_monitor.cc:111: /tmp/ray/session_2024-04-15_02-26-34_133505_135 is over 95% full, available space: 1074393088; capacity: 2360042831872. Object creation will fail if spilling is required.
(SenderProxyActor pid=5387) 2024-04-17 02:40:00.849 ERROR barriers.py:171 [alice] -- [Anonymous_job] Failed to send data to seq_id ping of carol from ping, error: <AioRpcError of RPC that terminated with:
(SenderProxyActor pid=5387) 	status = StatusCode.DEADLINE_EXCEEDED
(SenderProxyActor pid=5387) 	details = "Deadline Exceeded"
(SenderProxyActor pid=5387) 	debug_error_string = "UNKNOWN:Deadline Exceeded {created_time:"2024-04-17T02:40:00.848690293+00:00", grpc_status:4}"
(SenderProxyActor pid=5387) >
2024-04-17 02:40:00.900 INFO barriers.py:520 [alice] -- [Anonymous_job] Try ping ['carol'] at 22 attemp, up to 3600 attemps.

carol的运行结果：
2024-04-17 02:42:14 INFO barriers.py:353 [carol] --  RecverProxy was successfully created.
2024-04-17 02:42:17 INFO barriers.py:388 [carol] --  SendProxy was successfully created.
2024-04-17 02:42:17 INFO barriers.py:463 [carol] --  Try ping ['alice', 'bob'] at 0 attemp, up to 3600 attemps.
2024-04-17 02:42:17 INFO barriers.py:444 [carol] --  Succeeded to ping alice on 10.10.0.2:20001, the result: .
2024-04-17 02:42:17 INFO barriers.py:444 [carol] --  Succeeded to ping bob on 10.10.0.3:20001, the result: .

[Chrisdehe]

@beinnnn hey，请确认你部署的 authmanger 能被其他节点访问到。

[beinnnn]

@beinnnn hey，请确认你部署的 authmanger 能被其他节点访问到。

检查过了，alice、bob、carol三个节点都是可以访问到authmanger的

[zheyang0825]

这样部署方式是可以的，但是你可能需要确认一下，就是 Alice 和 Bob 配置的 Carol 地址是不是正确的。

[github-actions]

Stale issue message. Please comment to remove stale tag. Otherwise this issue will be closed soon.

[liouxiao]

怀疑遇到类似的问题，但Alice/Bob的报错信息和楼主的不太一样：

ValueError: Failed to look up actor with name 'SenderProxyActor'. This could because 1. You are trying to look up a named actor you didn't create. 2. The named actor died. 3. You did not use a namespace matching the namespace of the actor.

Carol异常退出：

/opt/occlum/build/bin/occlum: lin 354: 528 Killed RUST_BACKTRACE=1 "$instance_dir/build/bin/occlum-run" "$@"

[liouxiao]

怀疑遇到类似的问题，但Alice/Bob的报错信息和楼主的不太一样：

ValueError: Failed to look up actor with name 'SenderProxyActor'. This could because 1. You are trying to look up a named actor you didn't create. 2. The named actor died. 3. You did not use a namespace matching the namespace of the actor.

Carol异常退出：

/opt/occlum/build/bin/occlum: lin 354: 528 Killed RUST_BACKTRACE=1 "$instance_dir/build/bin/occlum-run" "$@"

看来是我的内存太小了，导致TEEU容器异常退出。

[Chrisdehe]

hi，@liouxiao 问题是否解决了呢