gian/cubefs
1
0
Fork 0
mirror of https://github.com/cubefs/cubefs.git synced 2026-08-02 02:00:56 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

fix(docs): Remove the description of redundant flash nodes

Browse Source 
Signed-off-by: clinx <chenlin1@oppo.com>
This commit is contained in:
clinx 2025-05-27 15:35:48 +08:00 committed by 梁曟風
parent 3df63b1a3b
commit 19db2687fc
2 changed files with 0 additions and 69 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

34
docs-zh/source/feature/flash.md
View File

@ -14,37 +14,3 @@ CubeFS的分布式缓存由多个FlashGroup组成，每个FlashGroup负责管理
FlashGroup由缓存节点FlashNode组成，可以分布在不同的zone中。客户端读取缓存数据时，则会通过对FlashNode进行延时分析，选择访问延时最低的FlashNode进行读取。
**FlashNode**
下面为 FlashNode 进程启动所需的配置文件示例：
```text
{
"role": "flashnode",
"listen": "18510",
"prof": "18511",
"logDir": "./logs",
"masterAddr": [
"xxx",
"xxx",
"xxx"
],
"memTotal": 0,
"cachePercent": 0.8,
"readRps": 100000,
"disableTmpfs": true,
"diskDataPath": [
"/path/data1:0",
"/path/data2:0"
],
"zoneName":"default"
}
```
**Master**
Master 负责对整个集群中所有分布式缓存的拓扑信息进行持久化存储与统一管理。它接收 FlashNode 的注册信息和心跳消息，以此判断各个 FlashNode 节点的存活状态。master接受cli命令对FlashGroup的slot分配，客户端在读取数据时，会向 Master 请求获取最新的分布式缓存拓扑结构，用于实现数据读取的正确路由。
**Client**
为支持文件缓存读取，客户端将结合卷级别的缓存开关和集群缓存功能状态，决定是否从分布式缓存中获取数据。

35
docs/source/feature/flash.md
View File

@ -13,38 +13,3 @@ CubeFS's distributed cache is composed of multiple FlashGroups, each responsible
Clients calculate a unique value corresponding to the consistent hashing ring based on the volume ID, inode, and offset information of the data block to be cached. The distributed cache's routing algorithm then finds the first slot value on the ring that is greater than or equal to this calculated value. The FlashGroup owning that slot is responsible for persisting the data block and providing caching and read services for it.
A FlashGroup consists of cache nodes called FlashNodes, which can be deployed across different zones. When a client reads cached data, it performs latency analysis across the available FlashNodes and selects the one with the lowest access latency for the read operation.
**FlashNode**
The following shows an example of the configuration file used to launch the FlashNode process:
```text
{
"role": "flashnode",
"listen": "18510",
"prof": "18511",
"logDir": "./logs",
"masterAddr": [
"xxx",
"xxx",
"xxx"
],
"memTotal": 0,
"cachePercent": 0.8,
"readRps": 100000,
"disableTmpfs": true,
"diskDataPath": [
"/path/data1:0",
"/path/data2:0"
],
"zoneName":"default"
}
```
**Master**
The Master manages the global topology of the distributed cache across the cluster, including persistent storage and centralized control. It monitors the status of each FlashNode by receiving registration and heartbeat messages. Furthermore, the Master supports slot assignment to FlashGroups via CLI commands. Clients retrieve the most up-to-date topology information from the Master to enable accurate routing of data read operations.
**Client**
In order to support file cache retrieval, the client evaluates both the per-volume cache configuration and the global caching state of the cluster before deciding to fetch data from the distributed cache.