systemd方式安装的oceanbase服务重启异常

【 系统环境 】
kylinos v10 sp3
【 使用版本 】
使用rpm包部署了systemd版本的oceanbase
oceanbase-ce.x86_64 4.3.5.1-101000042025031818.el8
【问题描述】
使用systemctl restart oceanbase命令手动重启服务失败,服务长时间卡住

【附件及日志】
obshell.log日志片段:
2025-08-19T10:53:13.362 INFO [4061242] [LS00000000000000] [scheduler/dag_handler.go:35] advance dag 6753

2025-08-19T10:53:13.362 INFO [4061242] [LS00000000000000] [scheduler/node_handler.go:25] advance node 37130

2025-08-19T10:53:13.363 INFO [4061242] [LS00000000000000] [scheduler/task_handler.go:39] advanceTask: node 37130 operator 1

2025-08-19T10:53:13.363 INFO [4061242] [LS00000000000000] [scheduler/task_handler.go:79] sub task 37130 state 3

2025-08-19T10:53:13.363 INFO [4061242] [LS00000000000000] [scheduler/task_handler.go:52] ready Task num 0, isFinished false, isSucceed true

2025-08-19T10:53:13.456 INFO [4061242] [57f48b5b0c9baacf] [common/middleware.go:189] API request: [GET /api/v1/task/dag/20175344868028866753, client=, traceId=57f48b5b0c9baacf, body={}]

2025-08-19T10:53:13.510 INFO [4061242] [F000000000000000] [ob/start_obsvr.go:518] get db instance failed: oceanbase db is nil , retry [19/600]

2025-08-19T10:53:13.525 INFO [4061242] [57f48b5b0c9baacf] [common/middleware.go:271] API response OK: [GET /api/v1/task/dag/20175344868028866753, client=, traceId=57f48b5b0c9baacf, duration=68, status=200, data=&{GenericDTO:0xc0014d14e0 DagDetail:0xc0014caa00}]

2025-08-19T10:53:14.022 ERROR [4061242] [F000000000000000] [runtime/asm_amd64.s:1598] get ocs db connection failed fields:, error=“oceanbase db is nil”

2025-08-19T10:53:14.526 INFO [4061242] [F000000000000000] [ob/start_obsvr.go:518] get db instance failed: oceanbase db is nil , retry [20/600]

2025-08-19T10:53:14.527 INFO [4061242] [6e896cfe7c9f1461] [common/middleware.go:189] API request: [GET /api/v1/task/dag/20175344868028866753, client=, traceId=6e896cfe7c9f1461, body={}]

1 个赞

需要提供一份observer日志

observer日志片段截图如下,是否需要更多的内容?

restart一下 oceanbase不行么,你这个报错,像是有冲突,正在运行一个,又开启一个,或者有两个程序试图用一个PID,看看这个PID被哪个程序占用了

1 个赞

尝试过停止服务并kill掉所有observer进程和obshell进程,再systemctl restart oceanbase,仍然无法启动

1 个赞

需要提供一份附件。截图里的没有启动报错

这个不是报错,这个是检测不到 cgroup。随便在 cgroup 里面设置一个 oceanbase 的软链就可以了,甚至不用配置 ob ,如果你不用 cgroup 控制的话。这样,就不会提示这个警告了。

https://www.oceanbase.com/docs/common-oceanbase-database-cn-1000000003381940