flagos-ai / flagos-ai/FlagPerf

单机BENCHMARK 测试基础规格的inferconnect相关用例时会报错

Open
#797 0 comments 0 reactions 0 assignees View on GitHub
Dominant language
Python
Stars
379
Forks
124
PR merge metrics
No merged PRs in 30d

Description

报错详细信息日志:
2025-05-15 11:26:08,036 [DEBUG] [cluster_manager.py,306]Scp from 10.6.28.80:/home/metax/yiyu/flagperf/FlagPerf/base/result/run20250515112346 to /home/metax/yiyu/flagperf/FlagPerf/base/result/run20250515112346 [FAILED]. Output: /home/metax/yiyu/flagperf/FlagPerf/base/result/run20250515112346//container_main.log.txt: Permission denied
/home/metax/yiyu/flagperf/FlagPerf/base/result/run20250515112346//env.log.txt: Permission denied
/home/metax/yiyu/flagperf/FlagPerf/base/result/run20250515112346//hello.log.txt: Permission denied
/home/metax/yiyu/flagperf/FlagPerf/base/result/run20250515112346//interconnect-P2P_intraserver:C550/10.6.28.80_noderank0/benchmark.log.txt: Permission denied
/home/metax/yiyu/flagperf/FlagPerf/base/result/run20250515112346//interconnect-P2P_intraserver:C550/10.6.28.80_noderank0/container_main.log.txt: Permission denied
/home/metax/yiyu/flagperf/FlagPerf/base/result/run20250515112346//pip_install.log.txt: Permission denied
/home/metax/yiyu/flagperf/FlagPerf/base/result/run20250515112346//start_base_task.pid: Permission denied

[ERROR] [run.py,341]Case interconnect-P2P_intraserver:C550, log dir: /home/metax/yiyu/flagperf/FlagPerf/base/result/run20250515112346/interconnect-P2P_intraserver:C550 collect log failed on hosts: 10.6.28.80 ]2025-05-15 11:26:08,037 [WARNING] [run.py,352]Sorry! Not all logs have been collected in /home/metax/yiyu/flagperf/FlagPerf/base/result/run20250515112346

采用的是 :BENCHMARKS_OR_TOOLKITS: "BENCHMARK", 另外这个错误并不影响测试,并不会终止测试, 单机并不需要执行scp命令, 请

Contributor guide

No contributing guide indexed for this repository

Research direction

Start by reading cluster_manager.py around line 306 and run.py around lines 341-352, then trace log collection for a single-host BENCHMARK run. Reproduce the inferconnect case using the logged command path and permission errors; done means single-machine testing no longer attempts unnecessary SCP and the test continues with logs handled correctly.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
performance, testing-qa
Issue type
Bug
Difficulty
3/5
Estimated time
1-2 days
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.