DOC: Add EAC doc (fluid-cloudnative#2432)

* eac charts doc Signed-off-by: mayingchun <[email protected]> * eac_runtime.md Signed-off-by: mayingchun <[email protected]> * eac_runtime.md Signed-off-by: mayingchun <[email protected]> Signed-off-by: mayingchun <[email protected]>
shiwenwenya · Dec 25, 2022 · 7498a79 · 7498a79
1 parent c29fe08
commit 7498a79
Show file tree

Hide file tree

Showing 3 changed files with 294 additions and 0 deletions.
diff --git a/charts/eac/CHANGELOG.md b/charts/eac/CHANGELOG.md
@@ -0,0 +1,3 @@
+0.1.0
+
+- Support Kubernetes Orchestration via Fluid
diff --git a/charts/eac/README.md b/charts/eac/README.md
@@ -0,0 +1,2 @@
+# EAC (Elastic Accelerate Client)
+EAC 是一款由阿里云文件存储团队研发的针对 NAS 的用户态客户端，并在提供分布式缓存的同时，保证多客户端之间的缓存一致性。
diff --git a/docs/zh/samples/eac_runtime.md b/docs/zh/samples/eac_runtime.md
@@ -0,0 +1,289 @@
+# 示例 - 如何在 Fluid 中使用 EAC
+
+## 背景介绍
+
+EAC 是一款针对 NAS 的用户态客户端，并在提供分布式缓存的同时，保证多客户端之间的缓存一致性。
+
+如何在 ACK（Alibaba Cloud Container Service for Kubernetes）场景中使用 EAC 可以参考文档 [开启 CNFS 文件存储客户端加速特性](https://help.aliyun.com/document_detail/440307.html)。
+
+## 安装
+
+您可以从 [Fluid Releases](https://github.com/fluid-cloudnative/fluid/releases) 下载最新的 Fluid 安装包。
+
+在 Fluid 的安装 chart values.yaml 中将 `runtime.eac.enable` 设置为 `true` ，再参考 [安装文档](../userguide/install.md) 完成安装。并检查Fluid各组件正常运行：
+
+```shell
+$ kubectl get po -n fluid-system
+NAME                                     READY   STATUS    RESTARTS   AGE
+csi-nodeplugin-fluid-4m2rq               2/2     Running   0          81s
+csi-nodeplugin-fluid-8l6nr               2/2     Running   0          81s
+csi-nodeplugin-fluid-t7hl2               2/2     Running   0          81s
+dataset-controller-99bc4dcc8-sl6h7       1/1     Running   0          81s
+eacruntime-controller-6fd48c77fc-k2hhr   1/1     Running   0          81s
+fluid-webhook-d8c4dcc7-whq5k             1/1     Running   0          81s
+fluidapp-controller-78c7ccd7fd-blw6w     1/1     Running   0          81s
+```
+
+确保 `eacruntime-controller`、`dataset-controller`、`fluid-webhook` 的 pod 以及若干 `csi-nodeplugin` pod 正常运行。
+
+## 新建工作环境
+
+```shell
+$ mkdir <any-path>/eac
+$ cd <any-path>/eac
+```
+
+## 运行示例
+
+在使用 EAC 之前，您需要拥有一个[通用型 NAS](https://www.aliyun.com/product/nas?spm=5176.19720258.J_3207526240.80.e93976f4Ps3XxX)，以及一个和 NAS 处在同一 [VPC](https://www.aliyun.com/product/vpc?spm=5176.59209.J_3207526240.35.253f76b9hZAU4x) 的 [ACK](https://www.aliyun.com/product/kubernetes?spm=5176.7937172.J_3207526240.54.7f51751avPxHwi) 集群，并确保集群节点操作系统使用 Alibaba Cloud Linux 2.1903。
+
+**查看待创建的 `Dataset` 资源对象**
+
+```yaml
+$ cat<<EOF >dataset.yaml
+apiVersion: data.fluid.io/v1alpha1
+kind: Dataset
+metadata:
+  name: mydemo
+spec:
+  mounts:
+    - mountPoint: "eac://nas-mount-point-address:/sub/path"
+EOF
+```
+
+其中：
+
+- `mountPoint`：指的是 EAC 的子目录，是用户在 NAS 文件系统中存储数据的目录，以 `eac://` 开头；如 `eac://nas-mount-point-address:/sub/path` 为 `nas-mount-point-address` 文件系统的 `/sub/path` 子目录。
+
+**创建 `Dataset` 资源对象**
+```shell
+$ kubectl create -f dataset.yaml
+dataset.data.fluid.io/mydemo created
+```
+
+**查看 `Dataset` 资源对象状态**
+```shell
+$ kubectl get dataset mydemo
+NAME     UFS TOTAL SIZE   CACHED   CACHE CAPACITY   CACHED PERCENTAGE   PHASE      AGE
+mydemo                                                                  NotBound   14s
+```
+
+如上所示，`status` 中的 `phase` 属性值为 `NotBound`，这意味着该 `Dataset` 资源对象目前还未与任何 `EACRuntime` 资源对象绑定，接下来，我们将创建一个 `EACRuntime` 资源对象。
+
+**查看待创建的 `EACRuntime` 资源对象**
+
+```yaml
+$ cat<<EOF >runtime.yaml
+apiVersion: data.fluid.io/v1alpha1
+kind: EACRuntime
+metadata:
+  name: mydemo
+spec:
+  replicas: 1
+  tieredstore:
+    levels:
+      - mediumtype: MEM
+        volumeType: emptyDir
+        path: /mnt/eac-worker-cache-path
+        quota: 2Gi
+  fuse:
+    properties:
+      g_unas_InodeAttrExpireTimeoutSec: "100"
+      g_unas_InodeEntryExpireTimeoutSec: "100"
+EOF
+```
+
+**创建 `EACRuntime` 资源对象**
+
+```shell
+$ kubectl create -f runtime.yaml
+eacruntime.data.fluid.io/mydemo created
+```
+
+**检查 `EACRuntime` 资源对象是否已经创建**
+```shell
+$ kubectl get eacruntime
+NAME     MASTER PHASE   WORKER PHASE   FUSE PHASE   AGE
+mydemo   NotReady                                   23s
+```
+
+等待一段时间，让 `EACRuntime` 资源对象中的各个组件得以顺利启动，你会看到类似以下状态：
+
+```shell
+$ kubectl get po | grep mydemo
+mydemo-master-0   2/2     Running   0          81s
+mydemo-worker-0   1/1     Running   0          61s
+```
+
+```shell
+$ kubectl get eacruntime
+NAME     MASTER PHASE   WORKER PHASE   FUSE PHASE   AGE
+mydemo   Ready          Ready          Ready        55s
+```
+
+`EACRuntime` 的 FUSE 组件实现了懒启动，会在 pod 使用时再创建。
+
+然后，再查看 `Dataset` 状态，发现已经与 `EACRuntime` 绑定。
+
+```shell
+$ kubectl get dataset mydemo
+NAME     UFS TOTAL SIZE   CACHED   CACHE CAPACITY   CACHED PERCENTAGE   PHASE   AGE
+mydemo                                                                  Bound   106s
+```
+
+**查看待创建的 StatefulSet 资源对象**，其中 StatefulSet 使用上面创建的 `Dataset` 的方式为指定同名的 PVC。
+
+```yaml
+$ cat<<EOF >app.yaml
+apiVersion: apps/v1
+kind: StatefulSet
+metadata:
+  name: mydemo-app
+  labels:
+    app: busybox
+spec:
+  serviceName: "busybox"
+  replicas: 2
+  selector:
+    matchLabels:
+      app: busybox
+  template:
+    metadata:
+      labels:
+        app: busybox
+    spec:
+      containers:
+        - name: busybox
+          image: busybox
+          command: ["/bin/sh"]
+          args: ["-c", "sleep 3600"]
+          volumeMounts:
+            - mountPath: "/data"
+              name: demo
+      volumes:
+        - name: demo
+          persistentVolumeClaim:
+            claimName: mydemo
+EOF
+```
+
+**创建 StatefulSet 资源对象**
+
+```shell
+$ kubectl create -f app.yaml
+statefulset.apps/mydemo-app created
+```
+
+**检查 StatefulSet 资源对象是否已经创建**
+```shell
+$ kubectl get po | grep mydemo
+mydemo-app-0        1/1     Running   0          64s
+mydemo-app-1        1/1     Running   0          20s
+mydemo-fuse-dfswf   2/2     Running   0          20s
+mydemo-fuse-gb4lm   2/2     Running   0          63s
+mydemo-master-0     2/2     Running   0          3m12s
+mydemo-worker-0     1/1     Running   0          2m52s
+```
+
+可以看到 StatefulSet 已经创建成功，同时 EAC 的 FUSE 组件也启动成功。
+
+**测试缓存加速效果**
+```shell
+$ kubectl exec -it mydemo-app-0 -- /bin/sh -c  'ls -hl /data'
+total 1G
+-rw-r--r--    1 root     root          15 Dec 14 08:24 test.txt
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp1.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp10.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp11.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp12.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp13.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp14.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp15.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp16.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp17.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp18.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp19.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp2.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp20.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp21.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp22.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp23.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp24.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp25.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp26.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp27.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp28.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp29.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp3.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp30.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp31.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp32.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp33.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp34.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp35.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp36.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp37.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp38.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp39.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp4.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp40.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp5.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp6.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp7.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp8.32M
+-rw-r--r--    1 root     root       32.0M Dec 14 08:24 tmp9.32M
+```
+可以看到被挂载目录下存在多个 32M 的普通文件，现通过在两个不同业务 pod 中进行 cp 操作来验证缓存加速效果。
+
+```shell
+$ kubectl exec -it mydemo-app-0 -- /bin/sh -c  'time cp /data/* /'
+real	0m 15.25s
+user	0m 0.00s
+sys	0m 1.20s
+```
+先在 mydemo-app-0 pod 中执行 cp 操作耗时 15.25s。
+
+```shell
+$ kubectl exec -it mydemo-app-1 -- /bin/sh -c  'time cp /data/* /'
+real	0m 5.27s
+user	0m 0.00s
+sys	0m 1.27s
+```
+后在 mydemo-app-1 pod 中执行 cp 操作耗时 5.27s，加速效果明显。
+
+```shell
+$ kubectl logs mydemo-worker-0
+2022/12/14 17:23:57|INFO |th=0000000002B70450|photon/syncio/epoll.cpp:336|fd_events_epoll_init:init event engine: epoll
+2022/12/14 17:23:57|INFO |th=0000000002B70450|server/server.cpp:323|Server:[cachedfs->get_pool()->defaultQuota()=-1]
+2022/12/14 17:23:57|INFO |th=0000000002B70450|server/server.cpp:353|Server:Server Start.
+2022/12/14 17:34:30|INFO |th=00007F05981F8300|server/server.cpp:335|operator():Accept 192.168.0.185:45966
+2022/12/14 17:34:30|INFO |th=00007F05969E7280|server/server.cpp:335|operator():Accept 192.168.0.185:45968
+2022/12/14 17:34:30|INFO |th=00007F05979F16C0|server/cache-adapter.cpp:50|add_throughPutCount:server current READ Throughput = 0MB/s, WRITE Throughput = 0MB/s
+2022/12/14 17:34:30|INFO |th=00007F05961E4340|server/server.cpp:335|operator():Accept 192.168.0.185:45970
+2022/12/14 17:34:31|INFO |th=00007F05959E0740|server/server.cpp:335|operator():Accept 192.168.0.185:45972
+2022/12/14 17:34:31|INFO |th=00007F05930B7280|server/cache-adapter.cpp:50|add_throughPutCount:server current READ Throughput = 71MB/s, WRITE Throughput = 37MB/s
+2022/12/14 17:34:32|INFO |th=00007F05926AEF40|server/cache-adapter.cpp:50|add_throughPutCount:server current READ Throughput = 177MB/s, WRITE Throughput = 82MB/s
+2022/12/14 17:34:33|INFO |th=00007F05926AEF40|server/cache-adapter.cpp:50|add_throughPutCount:server current READ Throughput = 212MB/s, WRITE Throughput = 89MB/s
+2022/12/14 17:34:34|INFO |th=00007F05926AEF40|server/cache-adapter.cpp:50|add_throughPutCount:server current READ Throughput = 204MB/s, WRITE Throughput = 92MB/s
+2022/12/14 17:34:35|INFO |th=00007F05951D9AC0|server/cache-adapter.cpp:50|add_throughPutCount:server current READ Throughput = 191MB/s, WRITE Throughput = 86MB/s
+2022/12/14 17:34:37|INFO |th=00007F05916A4B00|server/cache-adapter.cpp:50|add_throughPutCount:server current READ Throughput = 199MB/s, WRITE Throughput = 90MB/s
+2022/12/14 17:34:38|INFO |th=00007F05938C0EC0|server/cache-adapter.cpp:50|add_throughPutCount:server current READ Throughput = 196MB/s, WRITE Throughput = 88MB/s
+2022/12/14 17:34:39|INFO |th=00007F05926AEF40|server/cache-adapter.cpp:50|add_throughPutCount:server current READ Throughput = 191MB/s, WRITE Throughput = 87MB/s
+2022/12/14 17:34:40|INFO |th=00007F058F57DA80|server/cache-adapter.cpp:50|add_throughPutCount:server current READ Throughput = 154MB/s, WRITE Throughput = 58MB/s
+2022/12/14 17:34:41|INFO |th=00007F05979F4B40|server/cache-adapter.cpp:50|add_throughPutCount:server current READ Throughput = 157MB/s, WRITE Throughput = 60MB/s
+2022/12/14 17:34:42|INFO |th=00007F0590D99EC0|server/cache-adapter.cpp:50|add_throughPutCount:server current READ Throughput = 210MB/s, WRITE Throughput = 77MB/s
+2022/12/14 17:34:43|INFO |th=00007F05951D82C0|server/cache-adapter.cpp:50|add_throughPutCount:server current READ Throughput = 213MB/s, WRITE Throughput = 109MB/s
+2022/12/14 17:34:44|INFO |th=00007F0590D99EC0|server/cache-adapter.cpp:50|add_throughPutCount:server current READ Throughput = 278MB/s, WRITE Throughput = 118MB/s
+2022/12/14 17:34:45|INFO |th=00007F05979F3B00|server/cache-adapter.cpp:50|add_throughPutCount:server current READ Throughput = 282MB/s, WRITE Throughput = 113MB/s
+2022/12/14 17:34:58|INFO |th=00007F05951D6680|server/server.cpp:335|operator():Accept 192.168.0.181:55794
+2022/12/14 17:34:58|INFO |th=00007F05940C9EC0|server/cache-adapter.cpp:50|add_throughPutCount:server current READ Throughput = 5MB/s, WRITE Throughput = 2MB/s
+2022/12/14 17:34:58|INFO |th=00007F05979F3300|server/server.cpp:335|operator():Accept 192.168.0.181:55796
+2022/12/14 17:34:58|INFO |th=00007F05938C26C0|server/server.cpp:335|operator():Accept 192.168.0.181:55798
+2022/12/14 17:34:58|INFO |th=00007F05930BCB00|server/server.cpp:335|operator():Accept 192.168.0.181:55800
+2022/12/14 17:34:59|INFO |th=00007F0590D9AEC0|server/cache-adapter.cpp:50|add_throughPutCount:server current READ Throughput = 475MB/s, WRITE Throughput = 0MB/s
+2022/12/14 17:35:00|INFO |th=00007F05926AF740|server/cache-adapter.cpp:50|add_throughPutCount:server current READ Throughput = 463MB/s, WRITE Throughput = 0MB/s
+2022/12/14 17:35:01|INFO |th=00007F0590592EC0|server/cache-adapter.cpp:50|add_throughPutCount:server current READ Throughput = 493MB/s, WRITE Throughput = 0MB/s
+2022/12/14 17:35:02|INFO |th=00007F058F585700|server/cache-adapter.cpp:50|add_throughPutCount:server current READ Throughput = 477MB/s, WRITE Throughput = 0MB/s
+2022/12/14 17:35:03|INFO |th=00007F0591EA6280|server/cache-adapter.cpp:50|add_throughPutCount:server current READ Throughput = 562MB/s, WRITE Throughput = 0MB/s
+```
+观察 worker 的日志，可以发现首次操作会将缓存写入 worker，第二次操作将读取缓存。
Original file line number	Diff line number	Diff line change
		@@ -0,0 +1,3 @@
		0.1.0

		- Support Kubernetes Orchestration via Fluid
Original file line number	Diff line number	Diff line change
		@@ -0,0 +1,2 @@
		# EAC (Elastic Accelerate Client)
		EAC 是一款由阿里云文件存储团队研发的针对 NAS 的用户态客户端，并在提供分布式缓存的同时，保证多客户端之间的缓存一致性。