memstore占用高长时间不释放

OceanBase 技术问题
#1

result_details.txt (144.8 KB)
这里是诊断日志，observer.log中出现大量[errcode=-4184] The data file disk space is exhausted. Please expand the capacity by resizing datafile!!!(ret=-4184)

13 个赞
#3

查看data_dir所在目录的磁盘占用情况，实际上还有很多空间可使用

6 个赞
#4

应该和你设置的参数有关系吧

– 查看各节点数据盘使用情况
SELECT svr_ip, svr_port, zone,
data_disk_capacity / 1024 / 1024 / 1024 AS data_cap_gb,
data_disk_in_use / 1024 / 1024 / 1024 AS data_used_gb,
(data_disk_capacity - data_disk_in_use) / 1024 / 1024 / 1024 AS data_free_gb,
data_disk_health_status
FROM oceanbase.GV$OB_SERVERS;

– 查看 datafile 相关参数
SELECT name, value, svr_ip, svr_port
FROM oceanbase.GV$OB_PARAMETERS
WHERE name IN (‘datafile_size’, ‘datafile_maxsize’, ‘datafile_next’,
‘datafile_disk_percentage’, ‘data_disk_usage_limit_percentage’,
‘data_disk_write_limit_percentage’)
ORDER BY svr_ip, name;
– 查看 datafile 相关参数
SELECT name, value, svr_ip, svr_port
FROM oceanbase.GV$OB_PARAMETERS
WHERE name IN (‘datafile_size’, ‘datafile_maxsize’, ‘datafile_next’,
‘datafile_disk_percentage’, ‘data_disk_usage_limit_percentage’,
‘data_disk_write_limit_percentage’)
ORDER BY svr_ip, name;

df -h查看一下信息

5 个赞
#5

多谢分享

2 个赞
#6

学到了

1 个赞
#7

obclient(root@sys)[oceanbase]> SELECT svr_ip, svr_port, zone,
→ data_disk_capacity / 1024 / 1024 / 1024 AS data_cap_gb,
→ data_disk_in_use / 1024 / 1024 / 1024 AS data_used_gb,
→ (data_disk_capacity - data_disk_in_use) / 1024 / 1024 / 1024 AS data_free_gb,
→ data_disk_health_status
→ FROM oceanbase.GV$OB_SERVERS;
±--------------±---------±------±---------------±---------------±---------------±------------------------+
| svr_ip | svr_port | zone | data_cap_gb | data_used_gb | data_free_gb | data_disk_health_status |
±--------------±---------±------±---------------±---------------±---------------±------------------------+
| 172.17.32.178 | 2882 | zone1 | 8.000000000000 | 7.996093750000 | 0.003906250000 | NORMAL |
±--------------±---------±------±---------------±---------------±---------------±------------------------+
1 row in set (0.007 sec)

obclient(root@sys)[oceanbase]> SELECT name, value, svr_ip, svr_port
→ FROM oceanbase.GV$OB_PARAMETERS
→ WHERE name IN (‘datafile_size’, ‘datafile_maxsize’, ‘datafile_next’,
→ ‘datafile_disk_percentage’, ‘data_disk_usage_limit_percentage’,
→ ‘data_disk_write_limit_percentage’)
→ ORDER BY svr_ip, name;
ERROR 1054 (42S22): Unknown column ‘‘datafile_size’’ in ‘where clause’
obclient(root@sys)[oceanbase]> SELECT name, value, svr_ip, svr_port
→ FROM oceanbase.GV$OB_PARAMETERS
→ WHERE name IN (‘datafile_size’, ‘datafile_maxsize’, ‘datafile_next’,
→ ‘datafile_disk_percentage’, ‘data_disk_usage_limit_percentage’,
→ ‘data_disk_write_limit_percentage’)
→ ORDER BY svr_ip, name;
ERROR 1054 (42S22): Unknown column ‘‘datafile_size’’ in ‘where clause’

[root@xx web]# df -h /home/max/obdata/1
文件系统 容量 已用 可用 已用% 挂载点
/dev/mapper/cl-home 300G 185G 116G 62% /home

2 个赞
#8

SELECT name, value, svr_ip, svr_port
→ FROM oceanbase.GV$OB_PARAMETERS
→ WHERE name IN (‘datafile_size’, ‘datafile_maxsize’, ‘datafile_next’,
→ ‘datafile_disk_percentage’, ‘data_disk_usage_limit_percentage’,
→ ‘data_disk_write_limit_percentage’)
→ ORDER BY svr_ip, name;
±---------------------------------±------±--------------±---------+
| name | value | svr_ip | svr_port |
±---------------------------------±------±--------------±---------+
| datafile_disk_percentage | 0 | 172.17.32.178 | 2882 |
| datafile_maxsize | 8G | 172.17.32.178 | 2882 |
| datafile_next | 2G | 172.17.32.178 | 2882 |
| datafile_size | 2G | 172.17.32.178 | 2882 |
| data_disk_usage_limit_percentage | 90 | 172.17.32.178 | 2882 |
| data_disk_write_limit_percentage | 0 | 172.17.32.178 | 2882 |
±---------------------------------±------±--------------±---------+
6 rows in set (0.006 sec)

2 个赞
#9

datafile_maxsize = 8G 限制了继续扩展，因此报 -4184 同时使用率 99.95% > data_disk_usage_limit_percentage = 90 ，也会触发保护。

把datafile_maxsize参数设置大点

4 个赞
#10

重新看一下OB相关的容量参数。

#11

:+1: :grinning:

#12

看来的多了解了解ob磁盘管理知识 :grinning:

#13

读完关于memstore的内容，感触很多。

#14

针对memstore的分析很透彻，受教了。

#15

6666

#16

写得很详细