OB 4.2.4_CE_HF1版本出现inner tables are unmatched错误

select * from DBA_OB_SERVERS; 用这个视图查一下

不是oms的问题 应该是ob的问题 你先提供一下信息 同学排查一下

在sys租户查询下,下面两个也查询下

select * from dba_ob_servers;

select * from dba_ob_zones;

obclient [oceanbase]> select * from dba_ob_servers;
±-------------±---------±—±------±---------±----------------±-------±---------------------------±----------±----------------------±---------------------------±---------------------------±------------------------------------------------------------------------------------------±------------------+
| SVR_IP | SVR_PORT | ID | ZONE | SQL_PORT | WITH_ROOTSERVER | STATUS | START_SERVICE_TIME | STOP_TIME | BLOCK_MIGRATE_IN_TIME | CREATE_TIME | MODIFY_TIME | BUILD_VERSION | LAST_OFFLINE_TIME |
±-------------±---------±—±------±---------±----------------±-------±---------------------------±----------±----------------------±---------------------------±---------------------------±------------------------------------------------------------------------------------------±------------------+
| 172.16.xx.xx | 2882 | 1 | zone1 | 2881 | YES | ACTIVE | 2024-09-13 17:29:19.454893 | NULL | NULL | 2024-09-13 17:12:56.276053 | 2024-09-13 17:29:25.224687 | 4.2.4.0_100010022024091012-0e8ca8d9363eb5d5fbb56e9ed0159b949c21dc80(Sep 10 2024 13:57:03) | NULL |
| 172.16.xx.xx | 2882 | 2 | zone2 | 2881 | NO | ACTIVE | 2024-09-13 17:30:17.290526 | NULL | NULL | 2024-09-13 17:12:56.307070 | 2024-09-13 17:30:19.399047 | 4.2.4.0_100010022024091012-0e8ca8d9363eb5d5fbb56e9ed0159b949c21dc80(Sep 10 2024 13:57:03) | NULL |
| 172.16.xx.xx | 2882 | 3 | zone3 | 2881 | NO | ACTIVE | 2024-09-13 17:31:08.623663 | NULL | NULL | 2024-09-13 17:12:56.343110 | 2024-09-13 17:31:09.574727 | 4.2.4.0_100010022024091012-0e8ca8d9363eb5d5fbb56e9ed0159b949c21dc80(Sep 10 2024 13:57:03) | NULL |
±-------------±---------±—±------±---------±----------------±-------±---------------------------±----------±----------------------±---------------------------±---------------------------±------------------------------------------------------------------------------------------±------------------+

obclient [oceanbase]> select * from dba_ob_zones;
±------±---------------------------±---------------------------±-------±----±-----------±----------+
| ZONE | CREATE_TIME | MODIFY_TIME | STATUS | IDC | REGION | TYPE |
±------±---------------------------±---------------------------±-------±----±-----------±----------+
| zone1 | 2024-09-13 17:12:55.537426 | 2024-09-13 17:29:21.239638 | ACTIVE | | sys_region | ReadWrite |
| zone2 | 2024-09-13 17:12:55.537426 | 2024-09-13 17:30:22.680200 | ACTIVE | | sys_region | ReadWrite |
| zone3 | 2024-09-13 17:12:55.537426 | 2024-09-13 17:31:10.308711 | ACTIVE | | sys_region | ReadWrite |
±------±---------------------------±---------------------------±-------±----±-----------±----------+
3 rows in set (0.227 sec)

这个问题已联系研发同学排查中,有进展会及时回复你

这个集群是升级上来的吗?

查看下集群升级情况
select * from DBA_OB_CLUSTER_EVENT_HISTORY where module like ‘%upgrade%’;

该语句没有返回数据:
obclient [oceanbase]> select * from DBA_OB_CLUSTER_EVENT_HISTORY where module like ‘%upgrade%’;
Empty set (0.002 sec)

我这个应该是升级上来的,部署没法指定安装包的md5值,
这么升级的:
obd cluster upgrade ob_poc -c oceanbase-ce -V 4.2.4.0 --usable=e6c0a15b9aba27db858c1d336b898c57fce93c0b

同版本有两个安装包md5不同,部署的时候没有选择HF1版本,升级的时候指定md5升级的。

是从OB 4.2.4_CE升级到OB 4.2.4_CE_HF1吗?

是的

这个bug促使的升级

麻烦找下升级日志:

upgrade_checker.log
upgrade_cluster_health_checker.log
upgrade_post.log
upgrade_pre.log

upgrade_checker.log (7.1 KB)
upgrade_cluster_health_checker.log (2.6 KB)
只找到这两个日志

select * from __all_cluster_event_history; 麻烦查下这个 发下结果

升级时间对应的 obd 日志附件也发下

obclient [oceanbase]> select * from __all_cluster_event_history\G
*************************** 1. row ***************************
gmt_create: 2024-09-13 17:13:15.833555
module: BOOTSTRAP
event: BOOTSTRAP_SUCCESS
name1: cluster_version
value1: 4.2.4.0
name2: build_version
value2: 4.2.4.0_100000082024070810-556a8f594436d32a23ee92289717913cf503184b(Jul 8 2024 11:07:07)
name3:
value3:
name4:
value4:
name5:
value5:
name6:
value6:
extra_info:
1 row in set (0.002 sec)

obd.log (16.4 KB)

obd.2024-09-13.log (754.9 KB)

看日志是OBD跳过了这两项,应该是不符合预期的,我们排查下,

可以先执行 alter system run upgrade job ‘UPGRADE_ALL’; 就不会产生告警了。

[2024-09-13 17:27:53.748] [54979fb4-71b2-11ef-adfe-fa163e7493bc] [DEBUG] – skip oceanbase-ce-4.2.4.0-100010022024091012.el7-e6c0a15b9aba27db858c1d336b898c57fce93c0b upgrade_pre.py

[2024-09-13 17:31:10.340] [54979fb4-71b2-11ef-adfe-fa163e7493bc] [INFO] Exec upgrade_post.py
[2024-09-13 17:31:10.341] [54979fb4-71b2-11ef-adfe-fa163e7493bc] [DEBUG] – skip oceanbase-ce-4.2.4.0-100010022024091012.el7-e6c0a15b9aba27db858c1d336b898c57fce93c0b upgrade_post.py

已确认是OBD缺陷,计划在OBD3.0版本修复
绕过方法:执行 alter system run upgrade job ‘UPGRADE_ALL’; 会自动完成后续升级步骤,消除告警

感谢反馈!

执行该命令,已解决