Set instance count to 20 to allow scheduling of SLA aware jobs

Addtional changes include update settings to `10` instances and increasing
`go test` _timeout_ to `15m`.

.travis.yml

@@ -25,7 +25,7 @@ install:
   - docker-compose up -d
 
 script:
-  - go test -race -coverprofile=coverage.txt -covermode=atomic -v github.com/paypal/gorealis
+  - go test -race -coverprofile=coverage.txt -covermode=atomic -v github.com/paypal/gorealis -timeout 15m
 
 after_success:
   - bash <(curl -s https://codecov.io/bash)

docker-compose.yml

@@ -55,6 +55,74 @@ services:
     depends_on:
     - zk
 
+  agent-two:
+    image: rdelvalle/mesos-agent:1.5.1
+    pid: host
+    restart: on-failure
+    environment:
+      MESOS_MASTER: zk://192.168.33.2:2181/mesos
+      MESOS_CONTAINERIZERS: docker,mesos
+      MESOS_PORT: 5051
+      MESOS_HOSTNAME: localhost
+      MESOS_RESOURCES: ports(*):[11000-11999]
+      MESOS_SYSTEMD_ENABLE_SUPPORT: 'false'
+      MESOS_WORK_DIR: /tmp/mesos
+    networks:
+      aurora_cluster:
+        ipv4_address: 192.168.33.5
+
+    volumes:
+    - /sys/fs/cgroup:/sys/fs/cgroup
+    - /var/run/docker.sock:/var/run/docker.sock
+    depends_on:
+    - zk
+
+
+  agent-three:
+    image: rdelvalle/mesos-agent:1.5.1
+    pid: host
+    restart: on-failure
+    environment:
+      MESOS_MASTER: zk://192.168.33.2:2181/mesos
+      MESOS_CONTAINERIZERS: docker,mesos
+      MESOS_PORT: 5051
+      MESOS_HOSTNAME: localhost
+      MESOS_RESOURCES: ports(*):[11000-11999]
+      MESOS_SYSTEMD_ENABLE_SUPPORT: 'false'
+      MESOS_WORK_DIR: /tmp/mesos
+    networks:
+      aurora_cluster:
+        ipv4_address: 192.168.33.16
+
+    volumes:
+    - /sys/fs/cgroup:/sys/fs/cgroup
+    - /var/run/docker.sock:/var/run/docker.sock
+    depends_on:
+    - zk
+
+
+  agent-four:
+    image: rdelvalle/mesos-agent:1.5.1
+    pid: host
+    restart: on-failure
+    environment:
+      MESOS_MASTER: zk://192.168.33.2:2181/mesos
+      MESOS_CONTAINERIZERS: docker,mesos
+      MESOS_PORT: 5051
+      MESOS_HOSTNAME: localhost
+      MESOS_RESOURCES: ports(*):[11000-11999]
+      MESOS_SYSTEMD_ENABLE_SUPPORT: 'false'
+      MESOS_WORK_DIR: /tmp/mesos
+    networks:
+      aurora_cluster:
+        ipv4_address: 192.168.33.15
+
+    volumes:
+    - /sys/fs/cgroup:/sys/fs/cgroup
+    - /var/run/docker.sock:/var/run/docker.sock
+    depends_on:
+    - zk
+
   aurora-one:
     image: rdelvalle/aurora:0.21.0
     pid: host

job.go

@@ -56,7 +56,7 @@ type Job interface {
 	MaxFailure(maxFail int32) Job
 	Container(container Container) Job
 	PartitionPolicy(policy *aurora.PartitionPolicy) Job
-	Tier(tier *string) Job
+	Tier(tier string) Job
 	SlaPolicy(policy *aurora.SlaPolicy) Job
 }
 
@@ -326,8 +326,8 @@ func (j *AuroraJob) PartitionPolicy(policy *aurora.PartitionPolicy) Job {
 }
 
 // Set the Tier for the Job.
-func (j *AuroraJob) Tier(tier *string) Job {
+func (j *AuroraJob) Tier(tier string) Job {
-	j.jobConfig.TaskConfig.Tier = tier
+	j.jobConfig.TaskConfig.Tier = &tier
 
 	return j
 }

realis_e2e_test.go

@@ -393,6 +393,7 @@ func TestRealisClient_CreateService(t *testing.T) {
 
 	settings := realis.NewUpdateSettings()
 	settings.UpdateGroupSize = 2
+	settings.MinWaitInInstanceRunningMs = 10 * 1000
 	job.InstanceCount(3)
 	resp, result, err := r.CreateService(job, settings)
 
@@ -403,7 +404,7 @@ func TestRealisClient_CreateService(t *testing.T) {
 	var ok bool
 	var mErr error
 
-	if ok, mErr = monitor.JobUpdate(*result.GetKey(), 5, 180); !ok || mErr != nil {
+	if ok, mErr = monitor.JobUpdate(*result.GetKey(), 5, 240); !ok || mErr != nil {
 		// Update may already be in a terminal state so don't check for error
 		_, err := r.AbortJobUpdate(*result.GetKey(), "Monitor timed out.")
 
@@ -753,7 +754,6 @@ func TestAuroraJob_SlaPolicy(t *testing.T) {
 		},
 	}
 	role := "vagrant"
-	tier := "preferred"
 
 	var cpu float64 = 6.0
 	var disk int64 = 256
@@ -774,15 +774,16 @@ func TestAuroraJob_SlaPolicy(t *testing.T) {
 				ExecutorName(aurora.AURORA_EXECUTOR_NAME).
 				ExecutorData(string(thermosPayload)).
 				CPU(.1).
-				RAM(4).
+				RAM(2).
-				Disk(10).
+				Disk(5).
-				InstanceCount(3).
+				InstanceCount(20).
 				IsService(true).
 				SlaPolicy(&tt.args).
-				Tier(&tier)
+				Tier("preferred")
 
 			settings := realis.NewUpdateSettings()
-			settings.UpdateGroupSize = 1
+			settings.UpdateGroupSize = 10
+			settings.MinWaitInInstanceRunningMs = 15 * 1000
 			resp, result, err := r.CreateService(job, settings)
 
 			assert.NoError(t, err)
@@ -807,12 +808,12 @@ func TestAuroraJob_SlaPolicy(t *testing.T) {
 			_, err = r.KillJob(job.JobKey())
 
 			assert.NoError(t, err)
-
-			cpu = 0
-			ram = 0
-			disk = 0
-			r.SetQuota(role, &cpu, &ram, &disk)
 		})
+
+		cpu = 0
+		ram = 0
+		disk = 0
+		r.SetQuota(role, &cpu, &ram, &disk)
 	}
 
 }

runTestsMac.sh

@@ -1,4 +1,4 @@
 #!/bin/bash
 
 # Since we run our docker compose setup in bridge mode to be able to run on MacOS, we have to launch a Docker container within the bridge network in order to avoid any routing issues.
-docker run -t -v $(pwd):/go/src/github.com/paypal/gorealis --network gorealis_aurora_cluster golang:1.10.3-stretch  go test -v github.com/paypal/gorealis
+docker run -t -v $(pwd):/go/src/github.com/paypal/gorealis --network gorealis_aurora_cluster golang:1.10.3-stretch  go test -v github.com/paypal/gorealis -timeout 15m