[Enhancement]: 租户unit_num扩容，评估节点剩余磁盘空间是否足够 || [Enhancement]: Tenant unit_num expansion, evaluate whether the remaining disk space of the node is sufficient

[ob-lzq]

Description

扩容时检查租户的ls日志流大小，如果ls日志流太大，同zone下的节点可能有其他租户占用磁盘出现扩容时磁盘不足，扩容卡住的问题。

需要考虑unit_num 不同个数扩容时对磁盘大小的要求。

---

Description

Check the ls log stream size of the tenant when expanding capacity. If the ls log stream is too large, other tenants in the same zone may have insufficient disks during expansion, and the expansion is stuck.

It is necessary to consider the disk size requirements for unit_num when expanding capacity by different numbers.

[wayyoungboy]

了解当前资源使用情况:

在扩容前，首先需要了解当前每个节点上的资源使用情况，特别是磁盘空间的使用率。可以通过OceanBase提供的监控工具或SQL查询来获取这些信息。
查询示例：
SELECT svr_ip, svr_port, count(1)
FROM oceanbase.dba_ob_tablet_replicas
GROUP BY svr_ip, svr_port
ORDER BY svr_ip, svr_port;
这条SQL语句可以帮助您了解每个OBServer节点上的tablet数量，从而间接评估磁盘使用情况。
计算扩容后的资源需求:

根据扩容后所需的逻辑资源，计算扩容后单个节点所需的物理资源量。可以使用DBMS_OB_LIMIT_CALCULATOR.CALCULATE_MIN_PHY_RES_NEEDED_BY_LOGIC_RES存储过程来进行计算。
示例：
CALL DBMS_OB_LIMIT_CALCULATOR.CALCULATE_MIN_PHY_RES_NEEDED_BY_LOGIC_RES("LS:14, TABLET:40000");
计算结果将显示每个节点所需的最小物理资源量，包括内存、数据磁盘和clog磁盘等。

评估节点剩余磁盘空间:
根据上述计算结果，评估每个节点的剩余磁盘空间是否足够支持扩容后的资源需求。

如果某个节点的剩余磁盘空间不足，可能需要考虑以下几种解决方案：
增加磁盘容量：通过物理手段增加节点的磁盘容量。
迁移Unit：将部分Unit迁移到其他节点，以释放当前节点的磁盘空间。
调整资源分配：优化现有资源分配，减少不必要的资源占用。

监控和应急措施:
扩容后，持续监控节点的磁盘使用情况，确保磁盘空间不会达到临界值。

如果遇到磁盘空间满的情况，可以采取以下应急措施：
调大磁盘最大空间百分比：
ALTER SYSTEM SET datafile_disk_percentage = 98;
清理不必要的数据：删除部分可删除的数据、表、冗余索引等，并清空回收站中的数据。

---

Understand current resource usage:

Before expansion, you first need to understand the current resource usage on each node, especially the disk space usage. This information can be obtained through the monitoring tools or SQL queries provided by OceanBase.
Query example:
SELECT svr_ip, svr_port, count(1)
FROM oceanbase.dba_ob_tablet_replicas
GROUP BY svr_ip, svr_port
ORDER BY svr_ip, svr_port;
This SQL statement can help you understand the number of tablets on each OBServer node, thereby indirectly assessing disk usage.
Calculate the resource requirements after expansion:

Based on the logical resources required after expansion, calculate the amount of physical resources required by a single node after expansion. The calculation can be performed using the DBMS_OB_LIMIT_CALCULATOR.CALCULATE_MIN_PHY_RES_NEEDED_BY_LOGIC_RES stored procedure.
Example:
CALL DBMS_OB_LIMIT_CALCULATOR.CALCULATE_MIN_PHY_RES_NEEDED_BY_LOGIC_RES("LS:14, TABLET:40000");
The calculation results will show the minimum amount of physical resources required by each node, including memory, data disk, clog disk, etc.

Evaluate the remaining disk space of the node:
Based on the above calculation results, evaluate whether the remaining disk space of each node is sufficient to support the resource requirements after expansion.

If a node is low on remaining disk space, you may want to consider the following solutions:
Increase disk capacity: Increase the disk capacity of the node through physical means.
Migrate Units: Migrate some Units to other nodes to release disk space on the current node.
Adjust resource allocation: Optimize existing resource allocation and reduce unnecessary resource occupation.

Monitoring and emergency measures:
After expansion, continuously monitor the disk usage of the node to ensure that the disk space does not reach a critical value.

If you encounter a situation where the disk space is full, you can take the following emergency measures:
Increase the maximum disk space percentage:
ALTER SYSTEM SET datafile_disk_percentage = 98;
Clean up unnecessary data: Delete some deletable data, tables, redundant indexes, etc., and empty the data in the recycle bin.