广州做网站哪家公司好汕头网站建设平台
内存结构
从官网的结构图可以看出，一台observer可使用的总内存(memory_limit)包括 系统内存(system_memory) 和 租户内存(sys租户与普通租户)
系统内存
系统内存system_memory 属于 observer 的内部内存，允许其它租户共享使用该内存资源
(root@10.0.0.61:2883) [oceanbase]> show parameters like 'system_memory'; +-------+----------+-----------+----------+---------------+-----------+-------+-------------------------------------------------------------------------------------------------------------------------------------------------------------------------+----------+---------+---------+-------------------+ | zone | svr_type | svr_ip | svr_port | name | data_type | value | info | section | scope | source | edit_level | +-------+----------+-----------+----------+---------------+-----------+-------+-------------------------------------------------------------------------------------------------------------------------------------------------------------------------+----------+---------+---------+-------------------+ | zone1 | observer | 10.0.0.61 | 2882 | system_memory | NULL | 1G | the memory reserved for internal use which cannot be allocated to any outer-tenant, and should be determined to guarantee every server functions normally. Range: [0M,) | OBSERVER | CLUSTER | DEFAULT | DYNAMIC_EFFECTIVE | +-------+----------+-----------+----------+---------------+-----------+-------+-------------------------------------------------------------------------------------------------------------------------------------------------------------------------+----------+---------+---------+-------------------+ 1 row in set (0.018 sec)
租户内存
指 sys租户内存 和 创建的普通租户内存，OceanBase集群的租户 真正可分配使用的内存 为 可使用的总内存(memory_limit) — 系统内存(system_memory)，每个租户内存 包含 增量数据的 MemStore（tenant memstore）以及 KVCache 缓存（tenant cache memory）。
注:以下实验环节使用OceanBase 3.2.4.5企业版本，均使用sys租户登录查看
查看OBserver内存
可以查看当前 OBserver 总内存 mem_total 为15G，已经分配了10G，总的CPU为14，目前CPU已经全部分配
(root@10.0.0.61:2883) [oceanbase]> select zone,cpu_total,cpu_assigned,mem_total/1024/1024/1024,mem_assigned/1024/1024/1024 from __all_virtual_server_stat; +-------+-----------+--------------+--------------------------+-----------------------------+ | zone | cpu_total | cpu_assigned | mem_total/1024/1024/1024 | mem_assigned/1024/1024/1024 | +-------+-----------+--------------+--------------------------+-----------------------------+ | zone1 | 14 | 14 | 15.000000000000 | 10.000000000000 | +-------+-----------+--------------+--------------------------+-----------------------------+ 1 row in set (0.015 sec)(root@10.0.0.61:2883) [oceanbase]>
注意
mem_total值(租户可使用内存) 为 总内存(memory_limit) — 系统内存(system_memory)
(root@10.0.0.61:2883) [oceanbase]> show parameters like 'system_memory'; +-------+----------+-----------+----------+---------------+-----------+-------+-------------------------------------------------------------------------------------------------------------------------------------------------------------------------+----------+---------+---------+-------------------+ | zone | svr_type | svr_ip | svr_port | name | data_type | value | info | section | scope | source | edit_level | +-------+----------+-----------+----------+---------------+-----------+-------+-------------------------------------------------------------------------------------------------------------------------------------------------------------------------+----------+---------+---------+-------------------+ | zone1 | observer | 10.0.0.61 | 2882 | system_memory | NULL | 1G | the memory reserved for internal use which cannot be allocated to any outer-tenant, and should be determined to guarantee every server functions normally. Range: [0M,) | OBSERVER | CLUSTER | DEFAULT | DYNAMIC_EFFECTIVE | +-------+----------+-----------+----------+---------------+-----------+-------+-------------------------------------------------------------------------------------------------------------------------------------------------------------------------+----------+---------+---------+-------------------+ 1 row in set (0.011 sec)(root@10.0.0.61:2883) [oceanbase]> (root@10.0.0.61:2883) [oceanbase]> (root@10.0.0.61:2883) [oceanbase]> show parameters like 'memory_limit'; +-------+----------+-----------+----------+--------------+-----------+-------+-------------------------------------------------------------------------------------+----------+---------+---------+-------------------+ | zone | svr_type | svr_ip | svr_port | name | data_type | value | info | section | scope | source | edit_level | +-------+----------+-----------+----------+--------------+-----------+-------+-------------------------------------------------------------------------------------+----------+---------+---------+-------------------+ | zone1 | observer | 10.0.0.61 | 2882 | memory_limit | NULL | 16G | the size of the memory reserved for internal use(for testing purpose). Range: [0M,) | OBSERVER | CLUSTER | DEFAULT | DYNAMIC_EFFECTIVE | +-------+----------+-----------+----------+--------------+-----------+-------+-------------------------------------------------------------------------------------+----------+---------+---------+-------------------+ 1 row in set (0.019 sec)(root@10.0.0.61:2883) [oceanbase]>
查看租户内存
查看每个租户可用内存
这里查看的是 每个租户可使用的最大内存，包含memstore。 其实就是创建租户时指定的max_memory的值
(root@10.0.0.61:2883) [oceanbase]> select tenant_id,tenant_name,max_cpu,min_cpu,max_memory/1024/1024/1024 from v$unit; +-----------+-------------+---------+---------+---------------------------+ | tenant_id | tenant_name | max_cpu | min_cpu | max_memory/1024/1024/1024 | +-----------+-------------+---------+---------+---------------------------+ | NULL | NULL | 2 | 2 | 2.000000000000 | | 1 | sys | 4 | 4 | 2.000000000000 | | 1001 | obmysql | 2 | 2 | 4.000000000000 | | 1002 | oboracle | 2 | 2 | 2.000000000000 | | 1006 | tnt_pay | 2 | 2 | 1.000000000000 | | 1009 | oboms | 2 | 2 | 1.000000000000 | +-----------+-------------+---------+---------+---------------------------+ 6 rows in set (0.032 sec)(root@10.0.0.61:2883) [oceanbase]>
查看每个租户 memstore 内存
通过 tenant_id(租户标识字段) 与 memstore_limit(可使用的memstore内存) 来查看租户内存情况。500租户(系统内存存放的租户)可使用内存最大，为4G
(root@10.0.0.61:2883) [oceanbase]> select tenant_id,round(active/1024/1024/1024,2) active,round(total/1024/1024/1024,2) total,round(freeze_trigger/1024/1024/1024,2)-> freeze_trigger,round(mem_limit/1024/1024/1024,2) memstore_limit from v$memstore; +-----------+--------+-------+----------------+----------------+ | tenant_id | active | total | freeze_trigger | memstore_limit | +-----------+--------+-------+----------------+----------------+ | 1 | 0.20 | 0.20 | 0.30 | 1.00 | | 500 | 0.00 | 0.00 | 1288490188.80 | 4294967296.00 | | 1001 | 0.13 | 0.13 | 0.60 | 2.00 | | 1002 | 0.11 | 0.11 | 0.30 | 1.00 | | 1006 | 0.10 | 0.10 | 0.15 | 0.50 | | 1009 | 0.10 | 0.28 | 0.15 | 0.50 | +-----------+--------+-------+----------------+----------------+ 6 rows in set (0.010 sec)(root@10.0.0.61:2883) [oceanbase]>
修改OBserver内存
可以看到 memory_limit_percentage(可使用服务器物理内存最大百分比) 与 memory_limit(OBserver总内存) 分别为 90% 与 16G
(root@10.0.0.61:2883) [oceanbase]> show parameters like 'memory_limit%'; +-------+----------+-----------+----------+-------------------------+-----------+-------+----------------------------------------------------------------------------------------+----------+---------+---------+-------------------+ | zone | svr_type | svr_ip | svr_port | name | data_type | value | info | section | scope | source | edit_level | +-------+----------+-----------+----------+-------------------------+-----------+-------+----------------------------------------------------------------------------------------+----------+---------+---------+-------------------+ | zone1 | observer | 10.0.0.61 | 2882 | memory_limit_percentage | NULL | 90 | the size of the memory reserved for internal use(for testing purpose). Range: [10, 90] | OBSERVER | CLUSTER | DEFAULT | DYNAMIC_EFFECTIVE | | zone1 | observer | 10.0.0.61 | 2882 | memory_limit | NULL | 16G | the size of the memory reserved for internal use(for testing purpose). Range: [0M,) | OBSERVER | CLUSTER | DEFAULT | DYNAMIC_EFFECTIVE | +-------+----------+-----------+----------+-------------------------+-----------+-------+----------------------------------------------------------------------------------------+----------+---------+---------+-------------------+ 2 rows in set (0.005 sec)(root@10.0.0.61:2883) [oceanbase]>
修改 memory_limit_percentage
将百分比值改为60
(root@10.0.0.61:2883) [oceanbase]> alter system set memory_limit_percentage=60; Query OK, 0 rows affected (0.389 sec)(root@10.0.0.61:2883) [oceanbase]> (root@10.0.0.61:2883) [oceanbase]> show parameters like 'memory_limit%'; +-------+----------+-----------+----------+-------------------------+-----------+-------+----------------------------------------------------------------------------------------+----------+---------+---------+-------------------+ | zone | svr_type | svr_ip | svr_port | name | data_type | value | info | section | scope | source | edit_level | +-------+----------+-----------+----------+-------------------------+-----------+-------+----------------------------------------------------------------------------------------+----------+---------+---------+-------------------+ | zone1 | observer | 10.0.0.61 | 2882 | memory_limit_percentage | NULL | 60 | the size of the memory reserved for internal use(for testing purpose). Range: [10, 90] | OBSERVER | CLUSTER | DEFAULT | DYNAMIC_EFFECTIVE | | zone1 | observer | 10.0.0.61 | 2882 | memory_limit | NULL | 16G | the size of the memory reserved for internal use(for testing purpose). Range: [0M,) | OBSERVER | CLUSTER | DEFAULT | DYNAMIC_EFFECTIVE | +-------+----------+-----------+----------+-------------------------+-----------+-------+----------------------------------------------------------------------------------------+----------+---------+---------+-------------------+ 2 rows in set (0.011 sec)
再次查看OBserver总内存
可以发现 即使修改 memory_limit_percentage 的值，但是mem_total总内存依然不变
(root@10.0.0.61:2883) [oceanbase]> select zone,cpu_total,cpu_assigned,mem_total/1024/1024/1024,mem_assigned/1024/1024/1024 from __all_virtual_server_stat; +-------+-----------+--------------+--------------------------+-----------------------------+ | zone | cpu_total | cpu_assigned | mem_total/1024/1024/1024 | mem_assigned/1024/1024/1024 | +-------+-----------+--------------+--------------------------+-----------------------------+ | zone1 | 14 | 14 | 15.000000000000 | 10.000000000000 | +-------+-----------+--------------+--------------------------+-----------------------------+ 1 row in set (0.021 sec)(root@10.0.0.61:2883) [oceanbase]>
修改 memory_limit
将 memory_limit 值改为13
(root@10.0.0.61:2883) [oceanbase]> alter system set memory_limit='13G'; Query OK, 0 rows affected (0.175 sec)(root@10.0.0.61:2883) [oceanbase]> (root@10.0.0.61:2883) [oceanbase]> (root@10.0.0.61:2883) [oceanbase]> show parameters like 'memory_limit%'; +-------+----------+-----------+----------+-------------------------+-----------+-------+----------------------------------------------------------------------------------------+----------+---------+---------+-------------------+ | zone | svr_type | svr_ip | svr_port | name | data_type | value | info | section | scope | source | edit_level | +-------+----------+-----------+----------+-------------------------+-----------+-------+----------------------------------------------------------------------------------------+----------+---------+---------+-------------------+ | zone1 | observer | 10.0.0.61 | 2882 | memory_limit_percentage | NULL | 60 | the size of the memory reserved for internal use(for testing purpose). Range: [10, 90] | OBSERVER | CLUSTER | DEFAULT | DYNAMIC_EFFECTIVE | | zone1 | observer | 10.0.0.61 | 2882 | memory_limit | NULL | 13G | the size of the memory reserved for internal use(for testing purpose). Range: [0M,) | OBSERVER | CLUSTER | DEFAULT | DYNAMIC_EFFECTIVE | +-------+----------+-----------+----------+-------------------------+-----------+-------+----------------------------------------------------------------------------------------+----------+---------+---------+-------------------+ 2 rows in set (0.006 sec)(root@10.0.0.61:2883) [oceanbase]>
再次查看OBserver总内存
修改 memory_limit 的值，mem_total总内存发生变化
(root@10.0.0.61:2883) [oceanbase]> select zone,cpu_total,cpu_assigned,mem_total/1024/1024/1024,mem_assigned/1024/1024/1024 from __all_virtual_server_stat; +-------+-----------+--------------+--------------------------+-----------------------------+ | zone | cpu_total | cpu_assigned | mem_total/1024/1024/1024 | mem_assigned/1024/1024/1024 | +-------+-----------+--------------+--------------------------+-----------------------------+ | zone1 | 14 | 14 | 12.000000000000 | 10.000000000000 | +-------+-----------+--------------+--------------------------+-----------------------------+ 1 row in set (0.003 sec)
总结
1.当 memory_limit_percentage(可使用服务器物理内存最大百分比) 值 与 memory_limit(OBserver总内存) 值 均不为空时，OBserver总内存以 memory_limit为主，不受 memory_limit_percentage 影响。如果 memory_limit值为0，OBserver总内存 受
memory_limit_percentage 改变影响。
2.每个租户可用内存 包含 增量数据的 MemStore（tenant memstore）以及 KVCache 缓存（tenant cache memory）
3.OceanBase集群的租户 真正可分配使用的内存 为 可使用的总内存(memory_limit) — 系统内存(system_memory)，但系统内部内存允许其它租户共享使用该内存资源