Skip to content

Conversation

@akram
Copy link
Contributor

@akram akram commented Apr 21, 2015

  • Uses an TEST_MORE environment variable to run addition tests
  • additional tests are expected to be found under hack/more-end-to-end-tests/ directory
  • spawn a replication controller in testrc namespace using hello-openshift image
  • increase the number of replicas
  • check that the number of pods are increased
  • decrease the number of replicas
  • check that the number of pods are decreased

@mfojtik
Copy link
Contributor

mfojtik commented Apr 21, 2015

@akram is there a reason to have both Template and Config JSON in the repo? I mean you can create the Config by processing the Template.

@ncdc
Copy link
Contributor

ncdc commented Apr 21, 2015

Why is there a test more var? Are these tests that aren't valuable for normal e2e?

Sent from my iPhone

On Apr 21, 2015, at 6:29 AM, Akram Ben Aissi [email protected] wrote:

Uses an TEST_MORE environment variable to run addition tests
additional tests are expected to be found under hack/more-end-to-end-tests/ directory
spawn a replication controller in testrc namespace using hello-openshift image
increase the number of replicas
check that the number of pods are increased
decrease the number of replicas
check that the number of pods are decreased
You can view, comment on, or merge this pull request online at:

#1832

Commit Summary

Adds kube resize end-to-end test
File Changes

A examples/hello-openshift/hello-replication-controller.json (66)
A examples/hello-openshift/hello-template.json (61)
A hack/more-end-to-end-tests/test-kube-resize-command.sh (50)
M hack/test-end-to-end.sh (11)
Patch Links:

https://github.com/openshift/origin/pull/1832.patch
https://github.com/openshift/origin/pull/1832.diff

Reply to this email directly or view it on GitHub.

@akram
Copy link
Contributor Author

akram commented Apr 21, 2015

Hi
@ncdc
I wanted to reach consensus on the question if those tests should be run everytime or not

If you think so, I have no problem removing the variable.

Sent from mobile

On 21 avr. 2015, at 12:56, Andy Goldstein [email protected] wrote:

Why is there a test more var? Are these tests that aren't valuable for normal e2e?

Sent from my iPhone

On Apr 21, 2015, at 6:29 AM, Akram Ben Aissi [email protected] wrote:

Uses an TEST_MORE environment variable to run addition tests
additional tests are expected to be found under hack/more-end-to-end-tests/ directory
spawn a replication controller in testrc namespace using hello-openshift image
increase the number of replicas
check that the number of pods are increased
decrease the number of replicas
check that the number of pods are decreased
You can view, comment on, or merge this pull request online at:

#1832

Commit Summary

Adds kube resize end-to-end test
File Changes

A examples/hello-openshift/hello-replication-controller.json (66)
A examples/hello-openshift/hello-template.json (61)
A hack/more-end-to-end-tests/test-kube-resize-command.sh (50)
M hack/test-end-to-end.sh (11)
Patch Links:

https://github.com/openshift/origin/pull/1832.patch
https://github.com/openshift/origin/pull/1832.diff

Reply to this email directly or view it on GitHub.


Reply to this email directly or view it on GitHub.

@akram
Copy link
Contributor Author

akram commented Apr 21, 2015

@mfojtik
For my use case, I wanted to avoid a additional processing of the template to only create the config. I didn't want to rely on processing the template to do these checks.

Thinking about it, it could be useful to add an assertion on the processing, that would ensure that processing remains consistent.

If you agree, I can rely on processing, but I want to add an additional test on the resulting processing compared to the Config that I added. Any opinion ?

@smarterclayton
Copy link
Contributor

These tests should be part of hack/test-cmd, not hack/test-end-to-end. There's no reason we need a kubelet running to test this.

@akram
Copy link
Contributor Author

akram commented Apr 22, 2015

@smarterclayton: these tests check that pods are actually spawned+destroyed+respwaned and hence requires docker to be running (and also a kubelet I presume).
If not, can you please point to an example of an equivalent test that does not require kubelet in test-cmd, I can't see any.

@smarterclayton
Copy link
Contributor

I don't want to verify that here. Upstream verifies all that - there should be the primary test (CLI -> API) and then a maybe a deeper test (extended) that verifies pods get recreated. The two shouldn't be lumped together. The latter test is something you could very simply in e2e but I don't want e2e verifying duplicate things from what hack/test-cmd can verify.

----- Original Message -----

@smarterclayton: these tests check that pods are actually
spawned+destroyed+respwaned and hence requires docker to be running (and
also a kubelet I presume).
If not, can you please point to an example of an equivalent test that does
not require kubelet in test-cmd, I can't see any.


Reply to this email directly or view it on GitHub:
#1832 (comment)

@akram
Copy link
Contributor Author

akram commented Apr 22, 2015

For the kube resize test, this is not doing duplicate verification:

  • test-cmd.sh only check that the command runs and return with exit code 0
  • test-e2e.sh checks that upsizing deploys new pods up to the specified
    replicas count, and down sizing destroy pods until reaching the
    specified replica count.

This is possibly checked upstream, but the openshift wrapping is not e2e
tested, even if in this case, that is probably a straight forward passing.

Le 22/04/15 17:10, Clayton Coleman a écrit :

I don't want to verify that here. Upstream verifies all that - there
should be the primary test (CLI -> API) and then a maybe a deeper test
(extended) that verifies pods get recreated. The two shouldn't be
lumped together. The latter test is something you could very simply in
e2e but I don't want e2e verifying duplicate things from what
hack/test-cmd can verify.

----- Original Message -----

@smarterclayton: these tests check that pods are actually
spawned+destroyed+respwaned and hence requires docker to be running (and
also a kubelet I presume).
If not, can you please point to an example of an equivalent test
that does
not require kubelet in test-cmd, I can't see any.


Reply to this email directly or view it on GitHub:
#1832 (comment)


Reply to this email directly or view it on GitHub
#1832 (comment).

@smarterclayton
Copy link
Contributor

test-cmd has the replication controllers running - there's no reason it can't test the controller behavior.

    - Uses an TEST_MORE environment variable to run addition tests
    - additional tests are expected to be found under hack/more-end-to-end-tests/ directory
    - spawn a replication controller in testrc namespace using hello-openshift image
    - increase the number of replicas
    - check that the number of pods are increased
    - decrease the number of replicas
    - check that the number of pods are decreased
@smarterclayton smarterclayton modified the milestone: 1.1.0 May 27, 2015
@danmcp danmcp closed this Jun 4, 2015
jboyd01 pushed a commit to jboyd01/origin that referenced this pull request Mar 26, 2018
…service-catalog/' changes from c3e3071633..231772fcc0

231772fcc0 origin build: add origin tooling
98af588 v0.1.11 release changes
01e2f90 v0.1.10 release changes
49af948 clear polling queue before starting new operation (openshift#1855)
252958e Refactor common serviceclass validations (openshift#1858)
68f55c6 Catalog Controller should listen on https and serve metrics over TLS secured channel (openshift#1851)
5d0f773 Refactor common broker validations (openshift#1865)
eeaf285 Add NamespacedServiceBroker switch to helm chart (openshift#1864)
d2c960c Add NamespacedServiceBroker Feature (openshift#1863)
ef15310 Fix NamespaceScoped doc text for ns types (openshift#1862)
8d0a637 fix async deprovision retry (openshift#1832)
a918a16 Update registry code from serviceclass to clusterserviceclass (openshift#1852)
4dfd13c Bump dependency on go-open-service-broker-client to 0.0.6 (openshift#1856)
958b7cd Rename SharedServicePlanSpec to CommonServicePlanSpec (openshift#1850)
426aec3 pick a better random port to listen on in integration tests (openshift#1844)
6a59ada OrphanMitigation condition and different handling of retry timeout (openshift#1789)
c9b8f60 Extracting common broker spec elements into embeddable struct (openshift#1841)
3f8fab6 Extract common service class spec fields (openshift#1834)
93dab13 Support for OSB [PR#452](cloudfoundry/servicebroker#452). (openshift#1849)
5e1e90d [WIP] Pass correct plan ID in deprovision request (for both deleting and orphan mitigation) (openshift#1847)
74f73c0 disable tests for deployment stage (openshift#1845)
82fc6e4 Revert "Pass correct plan ID in deprovision request (for both deleting and orphan mitigation) (openshift#1803)" (openshift#1843)
94b5795 gitignore integration.test binary (openshift#1840)
014c468 Extracting common plan spec into embeddable struct (openshift#1833)
5d7041b Use k8s NewUUID method exclusively  (openshift#1836)
eac3f96 A new test was added after prechecks happened for last pr. (openshift#1838)
4b5d159 Pass correct plan ID in deprovision request (for both deleting and orphan mitigation) (openshift#1803)
cc02f0e [svcat] Adding a filter to get plan. (openshift#1758)
70afb56 reset RemovedFromBroker flag on plans that are re-added by broker (openshift#1824)
712dd4a Add behavior to print deleted instance name (openshift#1806)
55505be Update catalog charts README configuration (openshift#1823)
6426c98 Controller reconciliation rework - part 2 (ServiceBinding) (openshift#1819)
c606560 Integrate svcat docs with Service Catalog's (openshift#1784)
e9aeeb0 Synchronize some generated code that was missed along the way (openshift#1801)
a63ebf7 Fix failing test: TestReconcileServiceInstanceWithFailedCondition (openshift#1813)
07ef743 Controller reconciliation rework - part 1 (ServiceInstance) (openshift#1779)
a7d602b Export the touch instance command (openshift#1809)
bddb9a7 Allow retries for instances with Failed condition after spec changes (openshift#1751)
a777af5 Add enhanced parameter options to provision (openshift#1785)
fd1a0b9 Print deleted bindings (openshift#1799)
36d437a Adding the ability to sync a service instance (openshift#1762)
1f60676 Remove Failed condition if there was no terminal failure (openshift#1788)
cd831de allow brokers to respond to getCatalog() with no services (openshift#1772) (openshift#1781)
e5c37ad Add ObservedGeneration and Provisioned into ServiceInstanceStatus (openshift#1748)
9021d8b Add carolynvs to OWNERS (openshift#1780)
b7643d6 Add all-namespaces flag to svcat (openshift#1782)
01e652f Use docker to interact with files made by docker (openshift#1777)
REVERT: c3e3071633 origin build: add origin tooling

git-subtree-dir: cmd/service-catalog/go/src/github.com/kubernetes-incubator/service-catalog
git-subtree-split: 231772fcc00be08b6b2665a39c4a3bacb0b2271f
jboyd01 pushed a commit to jboyd01/origin that referenced this pull request Mar 27, 2018
…service-catalog/' changes from c3e3071633..231772fcc0

231772fcc0 origin build: add origin tooling
98af588 v0.1.11 release changes
01e2f90 v0.1.10 release changes
49af948 clear polling queue before starting new operation (openshift#1855)
252958e Refactor common serviceclass validations (openshift#1858)
68f55c6 Catalog Controller should listen on https and serve metrics over TLS secured channel (openshift#1851)
5d0f773 Refactor common broker validations (openshift#1865)
eeaf285 Add NamespacedServiceBroker switch to helm chart (openshift#1864)
d2c960c Add NamespacedServiceBroker Feature (openshift#1863)
ef15310 Fix NamespaceScoped doc text for ns types (openshift#1862)
8d0a637 fix async deprovision retry (openshift#1832)
a918a16 Update registry code from serviceclass to clusterserviceclass (openshift#1852)
4dfd13c Bump dependency on go-open-service-broker-client to 0.0.6 (openshift#1856)
958b7cd Rename SharedServicePlanSpec to CommonServicePlanSpec (openshift#1850)
426aec3 pick a better random port to listen on in integration tests (openshift#1844)
6a59ada OrphanMitigation condition and different handling of retry timeout (openshift#1789)
c9b8f60 Extracting common broker spec elements into embeddable struct (openshift#1841)
3f8fab6 Extract common service class spec fields (openshift#1834)
93dab13 Support for OSB [PR#452](cloudfoundry/servicebroker#452). (openshift#1849)
5e1e90d [WIP] Pass correct plan ID in deprovision request (for both deleting and orphan mitigation) (openshift#1847)
74f73c0 disable tests for deployment stage (openshift#1845)
82fc6e4 Revert "Pass correct plan ID in deprovision request (for both deleting and orphan mitigation) (openshift#1803)" (openshift#1843)
94b5795 gitignore integration.test binary (openshift#1840)
014c468 Extracting common plan spec into embeddable struct (openshift#1833)
5d7041b Use k8s NewUUID method exclusively  (openshift#1836)
eac3f96 A new test was added after prechecks happened for last pr. (openshift#1838)
4b5d159 Pass correct plan ID in deprovision request (for both deleting and orphan mitigation) (openshift#1803)
cc02f0e [svcat] Adding a filter to get plan. (openshift#1758)
70afb56 reset RemovedFromBroker flag on plans that are re-added by broker (openshift#1824)
712dd4a Add behavior to print deleted instance name (openshift#1806)
55505be Update catalog charts README configuration (openshift#1823)
6426c98 Controller reconciliation rework - part 2 (ServiceBinding) (openshift#1819)
c606560 Integrate svcat docs with Service Catalog's (openshift#1784)
e9aeeb0 Synchronize some generated code that was missed along the way (openshift#1801)
a63ebf7 Fix failing test: TestReconcileServiceInstanceWithFailedCondition (openshift#1813)
07ef743 Controller reconciliation rework - part 1 (ServiceInstance) (openshift#1779)
a7d602b Export the touch instance command (openshift#1809)
bddb9a7 Allow retries for instances with Failed condition after spec changes (openshift#1751)
a777af5 Add enhanced parameter options to provision (openshift#1785)
fd1a0b9 Print deleted bindings (openshift#1799)
36d437a Adding the ability to sync a service instance (openshift#1762)
1f60676 Remove Failed condition if there was no terminal failure (openshift#1788)
cd831de allow brokers to respond to getCatalog() with no services (openshift#1772) (openshift#1781)
e5c37ad Add ObservedGeneration and Provisioned into ServiceInstanceStatus (openshift#1748)
9021d8b Add carolynvs to OWNERS (openshift#1780)
b7643d6 Add all-namespaces flag to svcat (openshift#1782)
01e652f Use docker to interact with files made by docker (openshift#1777)
REVERT: c3e3071633 origin build: add origin tooling

git-subtree-dir: cmd/service-catalog/go/src/github.com/kubernetes-incubator/service-catalog
git-subtree-split: 231772fcc00be08b6b2665a39c4a3bacb0b2271f
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants