Translate NodeStatus into taint in order to fitting libscheduler (#192)

* fix error in libscheduler

Signed-off-by: ichAB <[email protected]>

* update readme

Signed-off-by: ichAB <[email protected]>

* translate Nodestatus to taint

Signed-off-by: ichAB <[email protected]>

* fmt

Signed-off-by: ichAB <[email protected]>

* modify readme

Signed-off-by: ichAB <[email protected]>

* fix clippy

Signed-off-by: ichAB <[email protected]>

---------

Signed-off-by: ichAB <[email protected]>

Author: ichAB

README.md

@@ -1,4 +1,3 @@
-
 # rk8s - A Lite Version of Kubernetes in Rust
 
 rk8s is a lightweight, Kubernetes-compatible container orchestration system built on top of [Youki](https://github.com/youki-dev/youki), implementing the Container Runtime Interface (CRI) with support for three primary workload types: single containers, Kubernetes-style pods, and Docker Compose-style multi-container applications.
@@ -180,85 +179,44 @@ sudo mv target/debug/libbridge /opt/cni/bin/
 mkdir -p rootfs
 docker export $(docker create busybox) | tar -C rootfs -xvf -
 ```
-4. **Start Xline:**
-```bash
-git clone https://github.com/xline-kv/Xline.git
-cd Xline
-docker pull ghcr.io/xline-kv/xline:latest
-cp fixtures/{private,public}.pem scripts/
-./scripts/quick_start.sh
-```
 ### Usage Examples
-
 **Single Container:**
 ```bash
-sudo rkl container run single.yaml
-sudo rkl container list
-sudo rkl container exec single-container-test /bin/sh
+sudo project/target/debug/rkl container run single.yaml
+sudo project/target/debug/rkl container list
+sudo project/target/debug/rkl container exec single-container-test /bin/sh
 ```
-
 **Pod Management(Standalone):**
+
 ```bash
-sudo rkl pod run pod.yaml
-sudo rkl pod state simple-container-task
-sudo rkl pod exec simple-container-task container-name /bin/sh
+sudo project/target/debug/rkl pod run pod.yaml
+sudo project/target/debug/rkl pod state simple-container-task
+sudo project/target/debug/rkl pod exec simple-container-task container-name /bin/sh
 ```
+
 **Pod Management(Cluster):**
 ```bash
-# Create pod via RKS control plane  
-sudo rkl pod create pod.yaml --cluster 127.0.0.1:50051  
-  
-# List pods from RKS  
-RKS_ADDRESS=127.0.0.1:50051 sudo rkl pod list  
-  
-# Delete pod via RKS  
-sudo rkl pod delete simple-container-task --cluster 127.0.0.1:50051
+# Create pod via RKS control plane
+sudo project/target/debug/rkl pod create pod.yaml --cluster 127.0.0.1:50051
+# List pods from RKS
+sudo project/target/debug/rkl pod list --cluster 127.0.0.1:50051
+# Delete pod via RKS
+sudo project/target/debug/rkl pod delete simple-container-task --cluster 127.0.0.1:50051
 ```
 
 **Compose Applications:**
 ```bash
-sudo rkl compose up
-sudo rkl compose ps
-sudo rkl compose down
+sudo project/target/debug/rkl compose up
+sudo project/target/debug/rkl compose ps
+sudo project/target/debug/rkl compose down
 ```
 
 **Daemon Mode:**
 ```bash
-export RKL_ADDRESS=127.0.0.1:50051
-sudo rkl pod daemon  # Monitors /etc/rk8s/manifests/ and acts as work node of rks
-```
-**Start RKS:**
-```bash
-sudo rks start --config config.yaml
-```
-Example config.yaml:
-```yaml
-addr: "127.0.0.1:50051"
-
-xline_config:
-
-  endpoints:
-
-    - "http://172.20.0.3:2379"
-
-    - "http://172.20.0.4:2379"
-
-    - "http://172.20.0.5:2379"
-
-  prefix: "/coreos.com/network"
-
-  subnet_lease_renew_margin: 60
-
-network_config:
-
-  Network: "10.1.0.0/16"
-
-  SubnetMin: "10.1.1.0"
-
-  SubnetMax: "10.1.254.0"
-
-  SubnetLen: 24
+sudo RKL_ADDRESS=127.0.0.1:50051 project/target/debug/rkl pod daemon # Monitors /etc/rk8s/manifests/ and acts as work node of rks
 ```
+**RKS:**
+For detail about building a cluster with RKS and RKL, please refer to [RKS–RKL usage](./project/rks/readme.md).
 ## Key Features
 
 - **CRI Compliance** - Full Container Runtime Interface implementation

project/common/src/lib.rs

@@ -38,6 +38,8 @@ pub struct PodSpec {
     pub containers: Vec<ContainerSpec>,
     #[serde(default)]
     pub init_containers: Vec<ContainerSpec>,
+    #[serde(default)]
+    pub tolerations: Vec<Toleration>,
 }
 
 #[derive(Debug, Serialize, Deserialize, Clone)]
@@ -95,6 +97,66 @@ pub struct PodStatus {
     #[serde(rename = "podIP")]
     pub pod_ip: Option<String>,
 }
+
+#[derive(Debug, Serialize, Deserialize, Clone, Default)]
+pub struct Toleration {
+    /// Empty means match all taint keys.
+    pub key: Option<TaintKey>,
+
+    /// Operator represents a key's relationship to the value.
+    #[serde(default)]
+    pub operator: TolerationOperator,
+
+    /// Effect indicates the taint effect to match. None means match all.
+    pub effect: Option<TaintEffect>,
+
+    #[serde(default)]
+    pub value: String,
+}
+impl Toleration {
+    pub fn tolerate(&self, taint: &Taint) -> bool {
+        if self.effect.is_some() && self.effect.as_ref().unwrap() != &taint.effect {
+            return false;
+        }
+        if self.key.is_some() && self.key.as_ref().unwrap() != &taint.key {
+            return false;
+        }
+        match self.operator {
+            TolerationOperator::Equal => self.value == taint.value,
+            TolerationOperator::Exists => true,
+        }
+    }
+}
+#[derive(Debug, Serialize, Deserialize, Clone)]
+pub enum TolerationOperator {
+    Exists,
+    Equal,
+}
+
+impl Default for TolerationOperator {
+    fn default() -> Self {
+        Self::Equal
+    }
+}
+
+#[derive(Debug, Serialize, Deserialize, Clone, PartialEq, Eq)]
+pub enum TaintEffect {
+    NoSchedule,
+    PreferNoSchedule,
+    NoExecute,
+}
+
+#[derive(Debug, Serialize, Deserialize, Clone, PartialEq, Eq)]
+pub enum TaintKey {
+    NodeNotReady,
+    NodeUnreachable,
+    NodeUnschedulable,
+    NodeMemoryPressure,
+    NodeDiskPressure,
+    NodeNetworkUnavailable,
+    NodeOutOfService,
+}
+
 #[derive(Debug, Serialize, Deserialize, Clone)]
 pub enum RksMessage {
     //request
@@ -107,7 +169,7 @@ pub enum RksMessage {
     UserRequest(String),
     Heartbeat {
         node_name: String,
-        conditions: Vec<NodeCondition>,
+        status: NodeStatus,
     },
     SetNetwork(Box<NodeNetworkConfig>),
     UpdateRoutes(String, Vec<Route>),
@@ -126,8 +188,25 @@ pub enum RksMessage {
 pub struct NodeSpec {
     #[serde(rename = "podCIDR")]
     pub pod_cidr: String, // Pod network CIDR assigned to this node
+    #[serde(default)]
+    pub taints: Vec<Taint>,
+}
+#[derive(Debug, Serialize, Deserialize, Clone)]
+pub struct Taint {
+    pub key: TaintKey,
+    #[serde(default)]
+    pub value: String,
+    pub effect: TaintEffect,
+}
+impl Taint {
+    pub fn new(key: TaintKey, effect: TaintEffect) -> Self {
+        Self {
+            key,
+            effect,
+            value: String::new(),
+        }
+    }
 }
-
 /// Node status
 #[derive(Debug, Serialize, Deserialize, Clone)]
 pub struct NodeStatus {

project/libscheduler/src/cache.rs

@@ -46,13 +46,14 @@ impl Cache {
         } else {
             return false;
         };
+
         pod_info.scheduled = Some(node_name.to_owned());
         node.requested.cpu += pod_info.spec.resources.cpu;
         node.requested.memory += pod_info.spec.resources.memory;
+
         true
     }
 
-    /// Un assume a pod, if the pod is not scheduled, do nothing.
     pub fn unassume(&mut self, pod_name: &str) -> Option<PodInfo> {
         let pod_info = self.pods.get_mut(pod_name)?;
         let node_name_opt = pod_info.scheduled.clone();
@@ -61,9 +62,11 @@ impl Cache {
         } else {
             return None;
         };
+
         pod_info.scheduled = None;
         node.requested.cpu -= pod_info.spec.resources.cpu;
         node.requested.memory -= pod_info.spec.resources.memory;
+
         Some(pod_info.clone())
     }
 
@@ -74,16 +77,28 @@ impl Cache {
     pub fn remove_pod(&mut self, pod_name: &str) -> Option<PodInfo> {
         if let Some(p) = self.pods.get(pod_name)
             && let Some(n) = &p.scheduled
+            && let Some(node) = self.nodes.get_mut(n)
         {
-            let node = self.nodes.get_mut(n);
-            if let Some(node) = node {
-                node.requested.cpu -= p.spec.resources.cpu;
-                node.requested.memory -= p.spec.resources.memory;
-            }
+            node.requested.cpu -= p.spec.resources.cpu;
+            node.requested.memory -= p.spec.resources.memory;
         }
         self.pods.remove(pod_name)
     }
 
+    pub fn update_node(&mut self, new_node: NodeInfo) -> Option<NodeInfo> {
+        let name = new_node.name.clone();
+        if let Some(old_node) = self.nodes.get_mut(&name) {
+            let requested = old_node.requested.clone();
+            old_node.labels = new_node.labels;
+            old_node.spec = new_node.spec;
+            old_node.allocatable = new_node.allocatable;
+            old_node.requested = requested;
+            Some(old_node.clone())
+        } else {
+            self.nodes.insert(name.clone(), new_node)
+        }
+    }
+
     pub fn pop_pod_on_node(&mut self, node_name: &str) -> Vec<PodNameWithPriority> {
         let mut res = Vec::new();
         self.pods
@@ -97,10 +112,6 @@ impl Cache {
         res
     }
 
-    pub fn update_node(&mut self, node: NodeInfo) -> Option<NodeInfo> {
-        self.nodes.insert(node.name.clone(), node)
-    }
-
     pub fn remove_node(&mut self, node_name: &str) {
         self.nodes.remove(node_name);
     }

project/libscheduler/src/models.rs

@@ -2,6 +2,7 @@ use std::{cmp::Ordering, collections::HashMap};
 
 use tokio::time::Instant;
 
+use common::*;
 #[derive(Clone, Default, Debug)]
 pub struct ResourcesRequirements {
     /// CPU resource limits, measured in millicores.
@@ -196,7 +197,6 @@ pub struct PodInfo {
     pub queued_info: QueuedInfo,
     pub scheduled: Option<String>,
 }
-
 impl PartialEq for PodInfo {
     fn eq(&self, other: &Self) -> bool {
         self.spec.priority == other.spec.priority
@@ -264,81 +264,3 @@ pub struct Assignment {
     pub pod_name: String,
     pub node_name: String,
 }
-
-/// The pod this Toleration is attached to tolerates any taint that matches
-/// the triple <key,value,effect> using the matching operator <operator>.
-#[derive(Default, Clone, Debug)]
-pub struct Toleration {
-    /// Key is the taint key that the toleration applies to. Empty means match all taint keys.
-    /// If the key is empty, operator must be Exists; this combination means to match all values and all keys.
-    pub key: Option<TaintKey>,
-    /// Operator represents a key's relationship to the value.
-    /// Valid operators are Exists and Equal. Defaults to Equal.
-    pub operator: TolerationOperator,
-    /// Effect indicates the taint effect to match. None means match all taint effects.
-    /// When specified, allowed values are NoSchedule, PreferNoSchedule and NoExecute.
-    pub effect: Option<TaintEffect>,
-    pub value: String,
-}
-
-impl Toleration {
-    pub fn tolerate(&self, taint: &Taint) -> bool {
-        if self.effect.is_some() && self.effect.as_ref().unwrap() != &taint.effect {
-            return false;
-        }
-        if self.key.is_some() && self.key.as_ref().unwrap() != &taint.key {
-            return false;
-        }
-        match self.operator {
-            TolerationOperator::Equal => self.value == taint.value,
-            TolerationOperator::Exists => true,
-        }
-    }
-}
-
-#[derive(Clone, Debug)]
-pub enum TolerationOperator {
-    Exists,
-    Equal,
-}
-
-impl Default for TolerationOperator {
-    fn default() -> Self {
-        Self::Equal
-    }
-}
-
-#[derive(Clone, PartialEq, Eq, Debug)]
-pub enum TaintEffect {
-    NoSchedule,
-    PreferNoSchedule,
-    NoExecute,
-}
-
-#[derive(Clone, Debug)]
-pub struct Taint {
-    pub key: TaintKey,
-    pub value: String,
-    pub effect: TaintEffect,
-}
-
-impl Taint {
-    pub fn new(key: TaintKey, effect: TaintEffect) -> Self {
-        Self {
-            key,
-            effect,
-            value: String::new(),
-        }
-    }
-}
-
-#[derive(Clone, PartialEq, Eq, Debug)]
-pub enum TaintKey {
-    NodeNotReady,
-    NodeUnreachable,
-    NodeUnschedulable,
-    NodeMemoryPressure,
-    NodeDiskPressure,
-    NodeNetworkUnavailable,
-    NodeOutOfService,
-}