4.3.4/4.3.5经常报错磁盘夯主

4.3.4/4.3.5经常报错磁盘夯主。ERROR [COMMON] inner_aio (ob_io_manager.cpp:1843) [690446][T1_MDS_MINI_MER][T1][YB420A01F59C-00062B2E50C5DBD6-0-0] [lt=117][errcode=-4392] disk is hung(msg=“disk has fatal error”)

ERROR [USING_LOG_PREFIX] detect_palf_hang_failure_ (ob_failure_detector.cpp:374) [154564][T1001_Occam][T1001][Y0-0000000000000000-0-0] [lt=18][errcode=-4392] disk is hung(msg=“clog disk may be hung, add failure event”, clog_disk_hang_event={type:PROCESS HANG, module:LOG, info:clog disk hang event, level:FATAL}, clog_disk_last_working_time=1736374877545223, hung time=73557821)

麻烦发一份observer日志看一下,434和435报错hung,改机器尝试安装其他版本没这个报错么

[2025-01-08 21:51:56.807626] ERROR [USING_LOG_PREFIX] detect_palf_hang_failure_ (ob_failure_detector.cpp:374) [243799][T1001_Occam][T1001][Y0-0000000000000000-0-0] [lt=23][errcode=-4392] disk is hung(msg=“clog disk may be hung, add failure event”, clog_disk_hang_event={type:PROCESS HANG, module:LOG, info:clog disk hang event, level:FATAL}, clog_disk_last_working_time=1736344306415118, hung time=10261405)

434/435都是这个报错,目前只尝试了这两个版本,服务器是企业VMware,里面别的数据库都没有磁盘夯住的情况。

麻烦发一份observer日志看一下,磁盘使用的是ssd么,ob推荐使用ssd磁盘

不是ssd,就是想要在非ssd的环境下运行,因为存储更换不容易。

2025-01-08 17:15:42.695095|INFO|SERVER|OB_SERVER_CHECK_USER_TENANT_SCHEMA_REFRESHED_FINISH|0|0|237434|observer|YB420A01F59A-00062B2E4E
A0A0E2-0-0|check_user_tenant_schema_refreshed|ob_server.cpp:3450|"[server_start 15/18] observer check user tenant schema refreshed fin
ish."
2025-01-08 17:15:42.695106|INFO|SERVER|OB_SERVER_CHECK_LOG_REPLAY_OVER_BEGIN|0|0|237434|observer|YB420A01F59A-00062B2E4EA0A0E2-0-0|che
ck_log_replay_over|ob_server.cpp:3457|"[server_start 16/18] observer check log replay over begin."
2025-01-08 17:15:42.695149|INFO|SERVER|OB_SERVER_CHECK_LOG_REPLAY_OVER_FINISH|0|0|237434|observer|YB420A01F59A-00062B2E4EA0A0E2-0-0|ch
eck_log_replay_over|ob_server.cpp:3483|"[server_start 17/18] observer check log replay over finish."
2025-01-08 17:15:42.695173|INFO|SERVER|OB_SERVER_START_SUCCESS|0|0|237434|observer|YB420A01F59A-00062B2E4EA0A0E2-0-0|start|ob_server.c
pp:1300|"[server_start 18/18] observer start success."
2025-01-08 17:15:42.695190|INFO|SERVER|OB_SERVER_WAIT_BEGIN|0|0|237434|observer|YB420A01F59A-00062B2E4EA0A0E2-0-0|wait|ob_server.cpp:1
784|“observer process wait begin.”
2025-01-08 21:51:56.807656|ERROR|USING_LOG_PREFIX|OB_FAILURE_LOG_DISK_HUNG|-4392|1001|243799|T1001_Occam|Y0-0000000000000000-0-0|detec
t_palf_hang_failure_|ob_failure_detector.cpp:375|“clog disk may be hung, add failure event”
2025-01-09 04:13:28.407952|ERROR|USING_LOG_PREFIX|OB_FAILURE_LOG_DISK_HUNG|-4392|1001|243799|T1001_Occam|Y0-0000000000000000-0-0|detec
t_palf_hang_failure_|ob_failure_detector.cpp:375|“clog disk may be hung, add failure event”
2025-01-09 04:13:29.233098|ERROR|USING_LOG_PREFIX|OB_FAILURE_DATA_DISK_HUNG|-4392|1001|243799|T1001_Occam|Y0-0000000000000000-0-0|dete
ct_data_disk_io_failure_|ob_failure_detector.cpp:412|“data disk may be hung, add failure event”
2025-01-09 07:55:28.767311|ERROR|USING_LOG_PREFIX|OB_FAILURE_LOG_DISK_HUNG|-4392|1001|243799|T1001_Occam|Y0-0000000000000000-0-0|detec
t_palf_hang_failure_|ob_failure_detector.cpp:375|“clog disk may be hung, add failure event”
2025-01-09 10:31:43.631971|ERROR|USING_LOG_PREFIX|OB_FAILURE_LOG_DISK_HUNG|-4392|1002|243584|T1002_Occam|Y0-0000000000000000-0-0|detec
t_palf_hang_failure_|ob_failure_detector.cpp:375|“clog disk may be hung, add failure event”

不是SSD是会经常报这个 :joy:

你好,非ssd,难道不能用国产吗,非ssd能否正常使用国产db?

还想上普通机械盘:joy:

别闹,我在问官方

普通磁盘在性能上与ssd会在数据库读写功能上体现出很大的差距。磁盘性能弱,建议你可以先做一个raid, 比如raid 0, 先提升磁盘性能。

就是因为更改存储不方便,所以才问有没别的办法能够让ob在普通磁盘上正常运行,例如修改参数。

可以参考下:【精打细算】利用混闪机型部署 OceanBase 集群的探索