提交 c0f26f76 编写于 作者: C chenwei 提交者: cw123

add test hardware config and fix readme invalid link

Signed-off-by: Nchenwei <461432360@qq.com>
上级 c92fad46
......@@ -28,22 +28,21 @@
✨ Contents
========
- [About Curve](#-about-curve)
- [Curve Architecture](#️-curve-architecture)
- [Curve Deployment](#-curve-deployment)
- [Curve Contrast](#️-curve-contrast)
- [Design Documentation](#-design-documentation)
- [CurveBS quick start](#-curvebs-quick-start)
- [Deploy an All-in-one experience environment](#-deploy-an-all-in-one-experience-environment)
- [About Curve](#about-curve)
- [Curve Architecture](#curve-architecture)
- [Design Documentation](#design-documentation)
- [CurveBS quick start](#curvebs-quick-start)
- [Deploy an All-in-one experience environment](#deploy-an-all-in-one-experience-environment)
- [FIO curve block storage engine](#fio-curve-block-storage-engine)
- [CurveFS quick start](#-curvefs-quick-start)
- [Governance](#-Governance)
- [Contribute us](#-contribute-us)
- [Code of Conduct](#-Code-of-Conduct)
- [LICENSE](#-license)
- [Release Cycle](#-release-cycle)
- [Branch](#-branch)
- [Feedback & Contact](#-feedback--contact)
- [CurveFS quick start](#curvefs-quick-start)
- [Test environment configuration](#test-environment-configuration)
- [Governance](#governance)
- [Contribute us](#contribute-us)
- [Code of Conduct](#code-of-conduct)
- [LICENSE](#license)
- [Release Cycle](#release-cycle)
- [Branch](#branch)
- [Feedback & Contact](#feedback--contact)
## About Curve
......@@ -195,7 +194,7 @@ Please refer to the [CurveBS cluster deployment steps](https://github.com/opencu
[The command tools' instructions](docs/cn/curve_ops_tool.md)
### FIO Curve block storage engine
Fio Curve engine is added, you can clone https://github.com/opencurve/fio and compile the fio tool with our engine(depend on nebd lib), fio command line example:
Fio Curve engine is added, you can clone https://github.com/opencurve/fio and compile the fio tool with our engine(depend on nebd lib), fio command line example:
```bash
$ ./fio --thread --rw=randwrite --bs=4k --ioengine=nebd --nebd=cbd:pool//pfstest_test_ --iodepth=10 --runtime=120 --numjobs=10 --time_based --group_reporting --name=curve-fio-test
```
......@@ -203,6 +202,10 @@ $ ./fio --thread --rw=randwrite --bs=4k --ioengine=nebd --nebd=cbd:pool//pfstest
## CurveFS quick start
Please use [CurveAdm](https://github.com/opencurve/curveadm/wiki) tool to deploy CurveFS,see [CurveFS Deployment Process](https://github.com/opencurve/curveadm/wiki/curvefs-cluster-deployment), and the [CurveFS Command Instructions](curvefs/src/tools#readme).
## Test environment configuration
Please refer to the [Test environment configuration](docs/cn/测试环境配置信息.md)
## Practical
- [CurveBS+NFS Build NFS Server](docs/practical/curvebs_nfs.md)
......
......@@ -28,22 +28,21 @@
✨ 目录
========
- [关于 Curve](#-关于-curve)
- [Curve Architecture](#️-curve-architecture)
- [Curve 部署](#-curve-部署)
- [测试对比](#️-测试对比)
- [设计文档](#-设计文档)
- [CurveBS 快速体验](#-curvebs-快速体验)
- [部署All-in-one体验环境](#-部署all-in-one体验环境)
- [关于 Curve](#关于-curve)
- [Curve 架构](#curve-架构)
- [设计文档](#设计文档)
- [CurveBS 快速体验](#curvebs-快速体验)
- [部署All-in-one体验环境](#部署all-in-one体验环境)
- [FIO curve块存储引擎](#fio-curve块存储引擎)
- [CurveFS 快速体验](#-curvefs-快速体验)
- [社区治理](#-社区治理)
- [贡献我们](#-贡献我们)
- [行为守则](#-行为守则)
- [LICENSE](#-license)
- [版本发布周期](#-版本发布周期)
- [分支规则](#-分支规则)
- [反馈及交流](#-反馈及交流)
- [CurveFS 快速体验](#curvefs-快速体验)
- [测试环境配置](#测试环境配置)
- [社区治理](#社区治理)
- [贡献我们](#贡献我们)
- [行为守则](#行为守则)
- [LICENSE](#license)
- [版本发布周期](#版本发布周期)
- [分支规则](#分支规则)
- [反馈及交流](#反馈及交流)
## 关于 Curve
......@@ -200,6 +199,9 @@ $ ./fio --thread --rw=randwrite --bs=4k --ioengine=nebd --nebd=cbd:pool//pfstest
## CurveFS 快速体验
请使用 [CurveAdm](https://github.com/opencurve/curveadm/wiki) 工具进行 CurveFS 的部署,具体流程见:[CurveFS部署流程](https://github.com/opencurve/curveadm/wiki/curvefs-cluster-deployment), 以及[CurveFS命令行工具说明](curvefs/src/tools#readme)
## 测试环境配置
请参考 [测试环境配置](docs/cn/测试环境配置信息.md)
## 社区治理
请参考[社区治理](https://github.com/opencurve/community/blob/master/GOVERNANCE.md)
......
# Curve 测试环境配置信息
Curve 当前测试的性能,如果没有特别说明,都是在基于以下配置的机器上测试的。该配置仅仅表示curve当前性能数据的测试环境,由于部分硬件版本比较旧,该环境的硬件不作为推荐硬件型号,但是对硬件的配置部分可做参考。
## 集群topo
### curvebs 块存储测试集群规模
mds和chunkserver服务混合部署。
mds : 3台机器
chunkserver : 9台机器 * 20块盘/每台机器
## 机器选型
| 计算/存储节点机型 |
| --- |
| Dell Inc. PowerEdge R730xd |
| Inspur NF5280M4 |
## 所有机器公共配置
| 模块 | 版本 | 配置 |
| --- | --- | --- |
| 系统盘 |$ sudo /usr/sbin/megacli -LDPDInfo -aAll <br>Number of Virtual Disks: 1<br>Virtual Drive: 0 (Target Id: 0)<br>Name :<br>RAID Level : Primary-1, Secondary-0, RAID Level Qualifier-0<br>Size : 1.090 TB<br>Sector Size : 512<br>Is VD emulated : No<br>Mirror Data : 1.090 TB<br>State : Optimal<br>Strip Size : 256 KB<br>Number Of Drives : 2<br>Span Depth : 1<br>Default Cache Policy: WriteBack, ReadAhead, Direct, Write Cache OK if<br>Bad BBU<br>Current Cache Policy: WriteBack, ReadAhead, Direct, Write Cache OK if<br>Bad BBU<br>Default Access Policy: Read/Write<br>Current Access Policy: Read/Write<br>Disk Cache Policy : Disabled<br>Encryption Type : None<br>Bad Blocks Exist: No<br>PI type: No PI |RAID0 |
## mds 节点
| 模块 | 版本 | 配置 |
| --- | --- | --- |
| 机型 | Dell PowerEdge R730xd |
| os | debain9 |
| 内核 | 4.9.65|
| cpu||设置performance,开启cpu性能最大化|
| 网卡| Bonding Mode: IEEE 802.3ad Dynamic link aggregation<br>Transmit Hash Policy: layer2+3 (2)<br>MII Status: up<br>MII Polling Interval (ms): 100<br>Up Delay (ms): 200<br>Down Delay (ms): 200 |万M网卡双网口组bond|
|系统盘||1.2T*2 SAS (RAID1模式)<br>初始化之后进行拷盘测试,提前发现并剔除坏盘、慢盘等异常|
|数据盘||1.8T*2 SSD (不需要RAID)<br>初始化之后进行拷盘测试,提前发现并剔除坏盘、慢盘等异常|
## chunkserver 节点
|模块|版本|配置|
|---|---|---|
|机型|inspur NF5280M4| |
|os|debian9| |
|内核|4.9.65| |
|cpu| |设置performance,开启cpu性能最大化|
|megaraid_sas| 06.811.02.00-rc4| |
|数据盘raid卡配置|1.RAID FW 版本:24.21.0-0061 <br>2.RAID卡驱动版本:06.811.02.00-rc4|1. 配置为JBOD模式<br> 2. RAID 卡关闭 consistent check<br>3.RAID卡缓存策略:WriteThrough|
|数据盘磁盘|数据盘类型:1.8TB * 20 Intel S4500<br>| 1.数据盘关闭磁盘缓存<br>2.磁盘调度策略,SSD盘noop,HDD盘deadline<br>// 初始化之后进行拷盘测试,提前发现并剔除坏盘、慢盘等异常 <br>3、 sudo tune2fs -m 0 /dev/sda8<br>释放文件系统预留的5%空间 <br>4、如果至少两个raid卡,每个raid卡上数据盘尽量均衡 |
| 网卡|bond配置(<br>Bonding Mode: IEEE 802.3ad Dynamic link aggregation<br>Transmit Hash Policy: layer2+3 (2)<br>MII Status: up<br>MII Polling Interval (ms): 100<br>Up Delay (ms): 200<br>Down Delay (ms): 200 )|万M网卡双网口组bond|
系统盘|1.2T*2 SAS | RAID1模式|
|fd数量||1.系统允许的最大fd数量至少为10000000<br> 2.单个进程允许的最大fd数量不低于250000|
|mmap地址数量||每一个bthread都需要使用mmap分配stack,mmap数量操作系统是有限制的,建议修改/proc/sys/vm/max_map_count 至少为5642720。|
## 一些相关的命令
### raid卡查询命令
1. RAID FW 版本 查询命令:`sudo /usr/sbin/megacli -AdpAllInfo -aALL | grep "FW Package Build"`
2. RAID 卡关闭 consistent check 查询命令:`sudo /usr/sbin/megacli -AdpCcSched -info -a0`
3. RAID卡缓存策略查询命令:`sudo megacli -LDGetProp -Cache -LALL -a0`
4. RAID卡驱动版本查询命令:`sudo modinfo megaraid_sas`
### 磁盘相关命令
1. 关闭磁盘缓存命令:`sudo /sbin/hdparm -W 0 /dev/sdh`
2. 磁盘缓存是否关闭,查询命令:`sudo /sbin/hdparm -W /dev/sdh`
3. 磁盘调度策略,查询命令:`cat /sys/block/sdxxx/queue/schedule`
4. 释放文件系统预留空空间命令:`sudo tune2fs -m 0 /dev/sda8`
### 修改单个进程的允许的最大fs数量
1. 查询系统允许的最大fd数量:`cat /proc/sys/fs/file-max`
2. 查询单个进程允许的最大fd数量:`ulimit -n`
3. 修改单个进程允许的最大fd数量:
```
修改/etc/sysctl.conf配置,然后执行sysctl -p 生效。
echo "fs.nr_open = 2500000" >> /etc/sysctl.conf
echo "fs.file-max = 40000000" >> /etc/sysctl.conf
```
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册