scottk/kubernetes
1
0
Fork 0
mirror of https://github.com/optim-enterprises-bv/kubernetes.git synced 2025-11-04 12:18:16 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #10314 from mikedanese/dont-sigkill-over-and-over

Browse Source 
remove repeat failure command option and increase timeouts in etcd failure e2e
This commit is contained in:
Maxwell Forbes
2015-06-24 16:30:52 -07:00
parent 6e3b9097c4 5d8ffbaad9
commit d1301d2b2a
1 changed files with 7 additions and 20 deletions
Download Patch File Download Diff File Expand all files Collapse all files

23
test/e2e/etcd_failure.go
View File

@@ -23,7 +23,6 @@ import (
"github.com/GoogleCloudPlatform/kubernetes/pkg/api"
"github.com/GoogleCloudPlatform/kubernetes/pkg/fields"
"github.com/GoogleCloudPlatform/kubernetes/pkg/labels"
"github.com/GoogleCloudPlatform/kubernetes/pkg/util"
"github.com/GoogleCloudPlatform/kubernetes/pkg/util/wait"
. "github.com/onsi/ginkgo"
@@ -53,7 +52,6 @@ var _ = Describe("Etcd failure", func() {
framework,
"sudo iptables -A INPUT -p tcp --destination-port 4001 -j DROP",
"sudo iptables -D INPUT -p tcp --destination-port 4001 -j DROP",
false,
)
})
@@ -62,12 +60,11 @@ var _ = Describe("Etcd failure", func() {
framework,
"pgrep etcd | xargs -I {} sudo kill -9 {}",
"echo 'do nothing. monit should restart etcd.'",
true,
)
})
})
func etcdFailTest(framework Framework, failCommand, fixCommand string, repeat bool) {
func etcdFailTest(framework Framework, failCommand, fixCommand string) {
// This test requires SSH, so the provider check should be identical to
// those tests.
if !providerIs("gce") {
@@ -75,7 +72,7 @@ func etcdFailTest(framework Framework, failCommand, fixCommand string, repeat bo
return
}
doEtcdFailure(failCommand, fixCommand, repeat)
doEtcdFailure(failCommand, fixCommand)
checkExistingRCRecovers(framework)
@@ -89,21 +86,11 @@ func etcdFailTest(framework Framework, failCommand, fixCommand string, repeat bo
// master and go on to assert that etcd and kubernetes components recover.
const etcdFailureDuration = 20 * time.Second
func doEtcdFailure(failCommand, fixCommand string, repeat bool) {
func doEtcdFailure(failCommand, fixCommand string) {
By("failing etcd")
if repeat {
stop := make(chan struct{}, 1)
go util.Until(func() {
defer GinkgoRecover()
masterExec(failCommand)
}, 1*time.Second, stop)
time.Sleep(etcdFailureDuration)
stop <- struct{}{}
} else {
masterExec(failCommand)
time.Sleep(etcdFailureDuration)
}
masterExec(fixCommand)
}
@@ -121,7 +108,7 @@ func checkExistingRCRecovers(f Framework) {
rcSelector := labels.Set{"name": "baz"}.AsSelector()
By("deleting pods from existing replication controller")
expectNoError(wait.Poll(time.Millisecond*500, time.Second*30, func() (bool, error) {
expectNoError(wait.Poll(time.Millisecond*500, time.Second*60, func() (bool, error) {
pods, err := podClient.List(rcSelector, fields.Everything())
if err != nil {
Logf("apiserver returned error, as expected before recovery: %v", err)
@@ -139,7 +126,7 @@ func checkExistingRCRecovers(f Framework) {
}))
By("waiting for replication controller to recover")
expectNoError(wait.Poll(time.Millisecond*500, time.Second*30, func() (bool, error) {
expectNoError(wait.Poll(time.Millisecond*500, time.Second*60, func() (bool, error) {
pods, err := podClient.List(rcSelector, fields.Everything())
Expect(err).NotTo(HaveOccurred())
for _, pod := range pods.Items {