summaryrefslogtreecommitdiff
path: root/docs
diff options
context:
space:
mode:
authorcl2017 <chenlin.liu@daocloud.io>2023-05-22 11:03:34 +0800
committercl2017 <chenlin.liu@daocloud.io>2023-05-22 11:03:34 +0800
commitff67b930775f57cc875e5423d487cdcf41a125e3 (patch)
treee0342c74542238b10e8b0db723d2c962a196138b /docs
parentMerge pull request #4712 from Shelley-BaoYue/goprocess (diff)
downloadkubeedge-ff67b930775f57cc875e5423d487cdcf41a125e3.tar.gz
fix doc error
Signed-off-by: cl2017 <chenlin.liu@daocloud.io>
Diffstat (limited to 'docs')
-rw-r--r--docs/proposals/cloudcore-ha-design.md12
-rw-r--r--docs/proposals/cri.md14
-rw-r--r--docs/proposals/edgemesh-design.md12
3 files changed, 19 insertions, 19 deletions
diff --git a/docs/proposals/cloudcore-ha-design.md b/docs/proposals/cloudcore-ha-design.md
index 90bf4a865..e6c5bb7a9 100644
--- a/docs/proposals/cloudcore-ha-design.md
+++ b/docs/proposals/cloudcore-ha-design.md
@@ -51,7 +51,7 @@ At present, CloudCore is essentially an edge-resource distribution center, which
```
- UpStream (in EdgeController)
- When edge-node gets the pod from cloud, it will check if there are related configmaps and secrets in edge-store. If not it will send a "Query" message(with node information) to cloud, and Upstream will response.
+ When edge-node gets the pod from cloud, it will check if there are related configmaps and secrets in edge-store. If not it will send a "Query" message(with node information) to cloud, and Upstream will response.
- DeviceController
@@ -72,11 +72,11 @@ At present, CloudCore is essentially an edge-resource distribution center, which
### HA in hot standby mode
#### The lifecycle of CloudCore
1. LeaderElection block
- 1. Become the leader
- 1. Start each controller and serve for edge-nodes
- 1. Notify the LoadBalance cloudcore is ready
- 1. Lose the leadership position due to network fluctuation or other reason, **shutdown immediately**;At the same time, one of the standby cloudcores becomes leader.
- 1. Restart by an external operation, downgrade to be a standby cloudcore also known as follower.
+ 2. Become the leader
+ 3. Start each controller and serve for edge-nodes
+ 4. Notify the LoadBalance cloudcore is ready
+ 5. Lose the leadership position due to network fluctuation or other reason, **shutdown immediately**;At the same time, one of the standby cloudcores becomes leader.
+ 6. Restart by an external operation, downgrade to be a standby cloudcore also known as follower.
The process of 3(start the controllers) also contains the cache initialization, such as the above-mentioned LocationCache.Ideally cache should initialize before the Leader election block, but it means the deconstruction of controllers as well as a lot of work. If subsequent discoveries show that cache initialization time led to the switch time growth greatly, then consider to deconstruct the controllers. At present, the cache initialization time is in milliseconds.
diff --git a/docs/proposals/cri.md b/docs/proposals/cri.md
index 17e08cc68..c26b2b979 100644
--- a/docs/proposals/cri.md
+++ b/docs/proposals/cri.md
@@ -72,13 +72,13 @@ for docker runtime using dockershim is not considered in edged
The following configuration parameters need to be added
-No | Parameter | Type | Values | Description
----|---|---|---|---
-1 | runtimeType | string | docker/remote | Runtime name
-2 | remoteRuntimeEndpoint | string | /var/run/*.sock | Endpoint of remote runtime service
-3 | remoteImageEndpoint | string | same as remoteRuntimeEndpoint if not specified | Endpoint of remote image service
-4 | RuntimeRequestTimeout | Duration | time value | timeout for all runtime request
-5 | PodSandboxImage | string | image name | Image used for pause container in sandbox
+| No | Parameter | Type | Values | Description |
+|-----|-----------------------|----------|------------------------------------------------|-------------------------------------------|
+| 1 | runtimeType | string | docker/remote | Runtime name |
+| 2 | remoteRuntimeEndpoint | string | /var/run/*.sock | Endpoint of remote runtime service |
+| 3 | remoteImageEndpoint | string | same as remoteRuntimeEndpoint if not specified | Endpoint of remote image service |
+| 4 | RuntimeRequestTimeout | Duration | time value | timeout for all runtime request |
+| 5 | PodSandboxImage | string | image name | Image used for pause container in sandbox |
```go
type Config struct {
diff --git a/docs/proposals/edgemesh-design.md b/docs/proposals/edgemesh-design.md
index 86c29f73a..7013fd074 100644
--- a/docs/proposals/edgemesh-design.md
+++ b/docs/proposals/edgemesh-design.md
@@ -70,12 +70,12 @@ ServiceBus is in the same process with EdgeHub as a module
### Router
Add two endpoints definition:
-|Endpoint|Type|Capability|
-|---|---|---|
-|ClusterIP|Target|Request Forward|
-|ServiceBus|Source|Request Forward, Service Discovery, Load Balance|
-|ServiceBus|Target|Request Forward|
-|Router|Source|Request Forward, Service Discovery, Load Balance|
+| Endpoint | Type | Capability |
+|------------|--------|--------------------------------------------------|
+| ClusterIP | Target | Request Forward |
+| ServiceBus | Source | Request Forward, Service Discovery, Load Balance |
+| ServiceBus | Target | Request Forward |
+| Router | Source | Request Forward, Service Discovery, Load Balance |
<img src="../images/proposals/router-design.png">