[GH-ISSUE #157] 如何修改storage_main的监听端口?我修改后还是端口冲突8000 #91

Closed
opened 2026-10-26 09:44:57 +00:00 by bsakaguchi · 2 comments
Owner

Originally created by @xiaoyaoyouyue on GitHub (Mar 12, 2025).
Original GitHub issue: https://github.com/deepseek-ai/3FS/issues/157

Issue Report

标题

storage_main 进程在初始化时因端口冲突崩溃

问题描述

storage_main 进程在初始化时崩溃,并生成了coredump文件。崩溃发生在尝试绑定到端口8000时,因为该端口已经被占用。

环境信息

操作系统: Ubuntu
程序路径: /opt/3fs/bin/storage_main

配置文件路径:

/opt/3fs/etc/storage_main_launcher.toml
/opt/3fs/etc/storage_main.toml

崩溃信息

信号: SIGABRT (Aborted)
进程ID: 99874
线程ID: 0x7f7c290f8b80 (LWP 99874)

调用栈信息

@ 00000000011e5aef _ZN5folly10symbolizer12_GLOBAL__N_113signalHandlerEiP9siginfo_tPv /root/3fs/3fs/third_party/folly/folly/experimental/symbolizer/SignalHandler.cpp:449 @ 000000000004251f (unknown) @ 00000000000969fc pthread_kill @ 0000000000042475 raise @ 00000000000287f2 abort @ 000000000124c220 _ZNK5folly11LogCategory12admitMessageERKNS_10LogMessageE /root/3fs/3fs/third_party/folly/folly/logging/LogCategory.cpp:71 @ 000000000124b6d6 _ZN5folly18LogStreamProcessor6logNowEv /root/3fs/3fs/third_party/folly/folly/logging/LogStreamProcessor.cpp:190 @ 000000000124b929 _ZN5folly16LogStreamVoidifyILb1EEanERSo /root/3fs/3fs/third_party/folly/folly/logging/LogStreamProcessor.cpp:222 @ 00000000022a7c48 _ZN5hf3fs19TwoPhaseApplicationINS_7storage13StorageServerEE15initApplicationEv /root/3fs/3fs/src/common/app/TwoPhaseApplication.h:59 -> /root/3fs/3fs/src/storage/storage.cpp @ 0000000000f79956 _ZN5hf3fs15ApplicationBase3runEiPPc /root/3fs/3fs/src/common/app/ApplicationBase.cc:65 @ 00000000022a4a43 main /root/3fs/3fs/src/storage/storage.cpp:7 @ 0000000000029d8f (unknown) @ 0000000000029e3f __libc_start_main @ 00000000007eb964 _start ### 日志信息 [2025-03-12T11:02:58.636580924+08:00 storage_main:99874 Listener.cc:102 ERROR] create socket failed: std::system_error: failed to bind to async server socket: 192.168.223.128:8000: Address already in use [2025-03-12T11:02:58.636735964+08:00 storage_main:99874 ServiceGroup.cc:26 ERROR] error: RPC::ListenFailed(2011) [2025-03-12T11:02:58.636880029+08:00 storage_main:99874 Server.cc:27 ERROR] Setup group (StorageSerde) failed: RPC::ListenFailed(2011) [2025-03-12T11:02:58.637110823+08:00 storage_main:99874 Server.cc:31 ERROR] Server::setup failed: RPC::ListenFailed(2011) [2025-03-12T11:02:58.637139597+08:00 storage_main:99874 TwoPhaseApplication.h:59 FATAL] Init server failed: RPC::ListenFailed(2011) ### 配置信息 mgmtd_server_addresses = ["RDMA://192.168.223.128:8000"] (出现在多个配置文件中) listen_port = 18000 (出现在某些配置文件中) listen_port = 9000 (出现在某些配置文件中) ### 详细信息 崩溃点: hf3fs::TwoPhaseApplication::initApplication 方法中 ### 相关代码路径: /root/3fs/3fs/src/common/app/TwoPhaseApplication.h:59 /root/3fs/3fs/src/storage/storage.cpp /root/3fs/3fs/src/common/app/ApplicationBase.cc:65 /root/3fs/3fs/src/storage/storage.cpp:7 ### 可能的原因 端口冲突: 端口8000已经被其他进程占用,导致storage_main无法绑定到该端口。 配置不一致: 配置文件中存在多个不同的listen_port设置,可能导致端口冲突。 ### 重现步骤 启动 storage_main 进程,使用配置文件 /opt/3fs/etc/storage_main_launcher.toml 和 /opt/3fs/etc/storage_main.toml。
Originally created by @xiaoyaoyouyue on GitHub (Mar 12, 2025). Original GitHub issue: https://github.com/deepseek-ai/3FS/issues/157 ### Issue Report ### 标题 storage_main 进程在初始化时因端口冲突崩溃 ### 问题描述 storage_main 进程在初始化时崩溃,并生成了coredump文件。崩溃发生在尝试绑定到端口8000时,因为该端口已经被占用。 ### 环境信息 操作系统: Ubuntu 程序路径: /opt/3fs/bin/storage_main ### 配置文件路径: /opt/3fs/etc/storage_main_launcher.toml /opt/3fs/etc/storage_main.toml ### 崩溃信息 信号: SIGABRT (Aborted) 进程ID: 99874 线程ID: 0x7f7c290f8b80 (LWP 99874) ### 调用栈信息 <TEXT> @ 00000000011e5aef _ZN5folly10symbolizer12_GLOBAL__N_113signalHandlerEiP9siginfo_tPv /root/3fs/3fs/third_party/folly/folly/experimental/symbolizer/SignalHandler.cpp:449 @ 000000000004251f (unknown) @ 00000000000969fc pthread_kill @ 0000000000042475 raise @ 00000000000287f2 abort @ 000000000124c220 _ZNK5folly11LogCategory12admitMessageERKNS_10LogMessageE /root/3fs/3fs/third_party/folly/folly/logging/LogCategory.cpp:71 @ 000000000124b6d6 _ZN5folly18LogStreamProcessor6logNowEv /root/3fs/3fs/third_party/folly/folly/logging/LogStreamProcessor.cpp:190 @ 000000000124b929 _ZN5folly16LogStreamVoidifyILb1EEanERSo /root/3fs/3fs/third_party/folly/folly/logging/LogStreamProcessor.cpp:222 @ 00000000022a7c48 _ZN5hf3fs19TwoPhaseApplicationINS_7storage13StorageServerEE15initApplicationEv /root/3fs/3fs/src/common/app/TwoPhaseApplication.h:59 -> /root/3fs/3fs/src/storage/storage.cpp @ 0000000000f79956 _ZN5hf3fs15ApplicationBase3runEiPPc /root/3fs/3fs/src/common/app/ApplicationBase.cc:65 @ 00000000022a4a43 main /root/3fs/3fs/src/storage/storage.cpp:7 @ 0000000000029d8f (unknown) @ 0000000000029e3f __libc_start_main @ 00000000007eb964 _start ### 日志信息 <TEXT> [2025-03-12T11:02:58.636580924+08:00 storage_main:99874 Listener.cc:102 ERROR] create socket failed: std::system_error: failed to bind to async server socket: 192.168.223.128:8000: Address already in use [2025-03-12T11:02:58.636735964+08:00 storage_main:99874 ServiceGroup.cc:26 ERROR] error: RPC::ListenFailed(2011) [2025-03-12T11:02:58.636880029+08:00 storage_main:99874 Server.cc:27 ERROR] Setup group (StorageSerde) failed: RPC::ListenFailed(2011) [2025-03-12T11:02:58.637110823+08:00 storage_main:99874 Server.cc:31 ERROR] Server::setup failed: RPC::ListenFailed(2011) [2025-03-12T11:02:58.637139597+08:00 storage_main:99874 TwoPhaseApplication.h:59 FATAL] Init server failed: RPC::ListenFailed(2011) ### 配置信息 mgmtd_server_addresses = ["RDMA://192.168.223.128:8000"] (出现在多个配置文件中) listen_port = 18000 (出现在某些配置文件中) listen_port = 9000 (出现在某些配置文件中) ### 详细信息 崩溃点: hf3fs::TwoPhaseApplication<hf3fs::storage::StorageServer>::initApplication 方法中 ### 相关代码路径: /root/3fs/3fs/src/common/app/TwoPhaseApplication.h:59 /root/3fs/3fs/src/storage/storage.cpp /root/3fs/3fs/src/common/app/ApplicationBase.cc:65 /root/3fs/3fs/src/storage/storage.cpp:7 ### 可能的原因 端口冲突: 端口8000已经被其他进程占用,导致storage_main无法绑定到该端口。 配置不一致: 配置文件中存在多个不同的listen_port设置,可能导致端口冲突。 ### 重现步骤 启动 storage_main 进程,使用配置文件 /opt/3fs/etc/storage_main_launcher.toml 和 /opt/3fs/etc/storage_main.toml。
Author
Owner

@yuchengkang commented on GitHub (Mar 12, 2025):

修改配置后需要重新执行 /opt/3fs/bin/admin_cli -cfg /opt/3fs/etc/admin_cli.toml "set-config --type STORAGE --file /opt/3fs/etc/storage_main.toml"

<!-- gh-comment-id:2716323320 --> @yuchengkang commented on GitHub (Mar 12, 2025): 修改配置后需要重新执行 /opt/3fs/bin/admin_cli -cfg /opt/3fs/etc/admin_cli.toml "set-config --type STORAGE --file /opt/3fs/etc/storage_main.toml"
Author
Owner

@xiaoyaoyouyue commented on GitHub (Mar 12, 2025):

修改配置后需要重新执行 /opt/3fs/bin/admin_cli -cfg /opt/3fs/etc/admin_cli.toml "set-config --type STORAGE --file /opt/3fs/etc/storage_main.toml"

多谢,

<!-- gh-comment-id:2716346271 --> @xiaoyaoyouyue commented on GitHub (Mar 12, 2025): > 修改配置后需要重新执行 /opt/3fs/bin/admin_cli -cfg /opt/3fs/etc/admin_cli.toml "set-config --type STORAGE --file /opt/3fs/etc/storage_main.toml" 多谢,
Sign in to join this conversation.
No labels
pull-request
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
deepseek-ai/3FS#91
No description provided.