incubator-openwhisk-deploy-kube.git
3 days agoTesting improvements (#332) master
David Grove [Sat, 8 Dec 2018 17:57:10 +0000 (12:57 -0500)] 
Testing improvements (#332)

* Support for running :tests:testSystemBasic as part of `helm test`.

* Enable running tests:testSystemBasic for DockerContainerFactory
  configurations (ping test still fails with KubernetesContainerFactory
  because need to fix upstream to drop CAP_NET_ADMIN on user action pod).

* Change error handling strategy in myTask.sh to being mostly
  explicit to get better control and allow retries of commands
  that might be vulnerable to CouchDB's eventual consistiency.

* Remove obsolete script for starting minikube in TravisCI

* Push openwhisk core git tag forward 2 commits to pick up fix
  for org.apache package rename in tests:testSystemBasic.

* Separate chart deployment and testing into separate scripts
  because we need to use travis_wait when running helm tests,
  but we want to be able to eagerly see the output of the deploy
  step to enable early aborts of the test run when it gets hung.

4 days agoswitch to ow-utils image (#386)
David Grove [Fri, 7 Dec 2018 17:01:21 +0000 (12:01 -0500)] 
switch to ow-utils image (#386)

6 days agoUpdate k8s-dind-cluster.md (#385)
Pavel Kravchenko [Wed, 5 Dec 2018 15:41:53 +0000 (17:41 +0200)] 
Update k8s-dind-cluster.md (#385)

Node name changed

6 days agobump nginx to 1.13 (match default version in core repo) (#384)
David Grove [Wed, 5 Dec 2018 04:58:44 +0000 (23:58 -0500)] 
bump nginx to 1.13 (match default version in core repo) (#384)

8 days agofix a broken link in document k8s-ibm-private (#383)
Ying Chun Guo [Mon, 3 Dec 2018 14:09:22 +0000 (22:09 +0800)] 
fix a broken link in document k8s-ibm-private (#383)

8 days agofix dangling links to docs/ingress.md (was split into multiple files). (#381)
David Grove [Mon, 3 Dec 2018 01:46:36 +0000 (20:46 -0500)] 
fix dangling links to docs/ingress.md (was split into multiple files). (#381)

Fixes #379.

8 days agocleanup specification of storageclasses and persistence (#378)
David Grove [Mon, 3 Dec 2018 01:44:54 +0000 (20:44 -0500)] 
cleanup specification of storageclasses and persistence (#378)

Rename persistence variables in values.yaml to better reflect
what they actually mean. Update docs to match.

Remove special instructions for persistence on
Docker Desktop 18.06 (not actually needed).

Fixes #376.

11 days agoDocument ICP and Dynamic DNS Provisioning (#380)
Jonathan Springer [Fri, 30 Nov 2018 16:55:46 +0000 (10:55 -0600)] 
Document ICP and Dynamic DNS Provisioning (#380)

* Document ICP and Dynamic DNS Provisioning

Edits to documentation to address two concerns:

1.  Directions for deploying OpenWhisk onto IBM Cloud Private,
    including configuration of the IBM Image Security engine.
2.  Further instructions on how to deploy an NFS-based dynamic
    storage provisioning engine (which can be the easiest way
    around the PersistentVolumeClaim needs of OpenWhisk).

12 days agoDocker for Mac 2.0 does not need additional configuration for PVs (#377)
David Grove [Thu, 29 Nov 2018 15:44:20 +0000 (10:44 -0500)] 
Docker for Mac 2.0 does not need additional configuration for PVs (#377)

12 days agoDocument PersistentVolume configuration in k8s-docker-for-mac.md (#375)
Ying Chun Guo [Thu, 29 Nov 2018 06:52:18 +0000 (14:52 +0800)] 
Document PersistentVolume configuration in k8s-docker-for-mac.md (#375)

13 days agoHelm chart providers (#372)
Neeraj Laad [Wed, 28 Nov 2018 15:04:06 +0000 (15:04 +0000)] 
Helm chart providers (#372)

Fixed some linting issue with helm chart and values.yaml

13 days agoImprove support for configuration of action container DNS (#370)
David Grove [Wed, 28 Nov 2018 05:22:19 +0000 (00:22 -0500)] 
Improve support for configuration of action container DNS (#370)

Use go template loops to allow specification of a colon
separated list of DNS servers for user action containers.
Also align the name/placement of the action networking
fields in values.yaml to better match PureConfig tree structure.

Fixes #160.

13 days agoPeriodic version bump for core, runtimes, cli, and providers. (#373)
David Grove [Tue, 27 Nov 2018 21:27:38 +0000 (16:27 -0500)] 
Periodic version bump for core, runtimes, cli, and providers. (#373)

2 weeks agoUpdate configurationChoices.md (#371)
xieweicarl2018 [Tue, 27 Nov 2018 05:19:27 +0000 (21:19 -0800)] 
Update configurationChoices.md (#371)

Use correct path helm/openwhisk/values.yaml instead previous path helm/values.yaml.

2 weeks agoHelm lint fixes (#369)
Neeraj Laad [Mon, 26 Nov 2018 22:13:55 +0000 (22:13 +0000)] 
Helm lint fixes (#369)

* helm lint issue resolution

2 weeks agoHelm chart changes (#367)
Neeraj Laad [Mon, 26 Nov 2018 21:40:17 +0000 (21:40 +0000)] 
Helm chart changes (#367)

Added readme for helm catalog
Updated chart description for helm catalog
updated notes.txt to include instruction on testing the release
included NodePort as default ingress

2 weeks agoingress cleanup: allow default value for httpsNodePort (#368)
David Grove [Wed, 21 Nov 2018 05:40:19 +0000 (00:40 -0500)] 
ingress cleanup: allow default value for httpsNodePort (#368)

Make usage of nginx.httpsNodePort templates conditional on
ingress.type being NodePort to enable a default value
for nginx.httpsNodePort to be defined.

3 weeks agoAdd discussion of actions resolving Kubernetes services via DNS (#366)
David Grove [Tue, 20 Nov 2018 04:38:26 +0000 (23:38 -0500)] 
Add discussion of actions resolving Kubernetes services via DNS (#366)

Fixes #315.

3 weeks agoAdd nodejs:10 to default set of runtimes for k8s (#365)
Carlos Santana [Mon, 19 Nov 2018 21:51:22 +0000 (16:51 -0500)] 
Add nodejs:10 to default set of runtimes for k8s (#365)

3 weeks agoPost install notes formatting (#363)
Neeraj Laad [Mon, 19 Nov 2018 14:40:00 +0000 (14:40 +0000)] 
Post install notes formatting (#363)

* consistent formatting of post install notes

3 weeks agoSwitch to pinned docker images and git clones (#362)
David Grove [Fri, 16 Nov 2018 22:35:57 +0000 (17:35 -0500)] 
Switch to pinned docker images and git clones (#362)

All docker images and git clone operations now take tags
that do not default to latest/master.  This will increase
the stability of the Helm chart "nightly builds" by allowing
us to have a more controlled following of upstream changes.

All version tags are defined in values.yaml to enable easy overrides.

Fixes #353

3 weeks agoAlso tag with githash when tagging with latest (#361)
David Grove [Fri, 16 Nov 2018 19:57:12 +0000 (14:57 -0500)] 
Also tag with githash when tagging with latest (#361)

If tagging an image with latest, also tag it with the short
githash of current head so we can have a stable tag on dockerhub.

Fixes #359

3 weeks agoSmall improvements to README (#358)
David Grove [Fri, 16 Nov 2018 19:56:24 +0000 (14:56 -0500)] 
Small improvements to README (#358)

3 weeks agoStreamline Helm install process + linting fixes (#352)
David Grove [Thu, 15 Nov 2018 15:49:51 +0000 (10:49 -0500)] 
Streamline Helm install process + linting fixes (#352)

1. Consolidate core and providers into a single chart.
   Structurally, the provider charts were not independent
   of the core chart so it is more in keeping with
   best practices to have a single Chart that can
   deploy both the core and the optionally enabled event
   providers.

2. Rework the TravisCI testing so that the tests are now
   available as part of the Helm chart. This enables them
   to be easily run by a user who has installed the chart
   via `helm test owdev` to verify their deployment was
   successful.

3. Install the catalog, routemgmt, and provider packages
   in a single Job.

4. Per Helm best practices, resources defined in the Chart
   should not explicitly define their namespace.

5. Helm will automatically create the namespace when it
   deploys the Chart, so we can remove that step from
   our README.md

6. Change minimum Tiller version from 2.10 back to 2.9

7. Uniform labeling of all resources created by the Chart.

8. A few misc linting fixes

4 weeks agoMove smoketest into chart so it can be used for `helm test` (#351)
David Grove [Mon, 12 Nov 2018 19:15:48 +0000 (14:15 -0500)] 
Move smoketest into chart so it can be used for `helm test` (#351)

Pull the logic for the simple hello action smoketest into the
Helm chart for OpenWhisk so it can be used to verify that the
Chart deployed correctly.

4 weeks agobump redis from 3.2 to 4.0 (upsteam PR#4103) (#349)
David Grove [Mon, 12 Nov 2018 02:08:42 +0000 (21:08 -0500)] 
bump redis from 3.2 to 4.0 (upsteam PR#4103) (#349)

4 weeks agoEnable persistence by default (#347)
David Grove [Mon, 12 Nov 2018 01:29:57 +0000 (20:29 -0500)] 
Enable persistence by default (#347)

Consolidate options around persistence and enable it by default.
Kubernetes on Docker for Mac has good support for dynamic provisioning of
persistent volumes (via the hostpath provisioner), so enabling this by
default is the less surprising option.  In particular, it allows an
OpenWhisk deployment to survive controlled host system reboots and still
be usable.

Major reorganization of setup/install instructions to group by the
different kinds of Kubernetes clusters.  I think this should result in
a simpler install experience, since the Kubernetes specific details are
better organized and distinct from the main flow which is common to all
clusters.

4 weeks agoUpdate for upstream changes in loadbalancer/invoker memory config (#348)
David Grove [Sat, 10 Nov 2018 10:55:40 +0000 (05:55 -0500)] 
Update for upstream changes in loadbalancer/invoker memory config (#348)

The name of invokerUserMemory was changed upstream in 5b3e0b6a3 and
deploy-kube was never updated to catch up with the change.

4 weeks agohelm linting (#343)
David Grove [Fri, 9 Nov 2018 13:16:25 +0000 (08:16 -0500)] 
helm linting (#343)

1. namespace all template defines with the chart name
  2. add LICENSE to all charts
  3. add tiller semver to all charts
  4. boilerplate labels linking resource back to Chart
  5. add icon URL

4 weeks agopackage rename: whisk ==> org.apache.openwhisk (#346)
David Grove [Thu, 8 Nov 2018 15:49:51 +0000 (10:49 -0500)] 
package rename: whisk ==> org.apache.openwhisk (#346)

4 weeks agofix typo in README (#344)
David Grove [Wed, 7 Nov 2018 23:42:45 +0000 (18:42 -0500)] 
fix typo in README (#344)

4 weeks agoFix YAML parse error (#341)
Patrick Löcker [Wed, 7 Nov 2018 14:31:34 +0000 (15:31 +0100)] 
Fix YAML parse error (#341)

A space was removed in commit #340. That causes a parse error when persistence is enabled.

4 weeks agoConform to Helm best practices for structure of templates/ (#340)
David Grove [Wed, 7 Nov 2018 09:14:24 +0000 (04:14 -0500)] 
Conform to Helm best practices for structure of templates/ (#340)

Specifically make changes to comply with the rules:
  1. Each resource definition should be in its own template file.
  2. Template file names should use dashed notation, not camelcase.
  3. Template file names should reflect the resource kind in the name.

See guidelines at https://github.com/helm/helm/blob/master/docs/chart_best_practices/templates.md.

4 weeks agoreorganize docs describing Kubernetes options and mycluster.yaml (#335)
David Grove [Tue, 6 Nov 2018 21:39:52 +0000 (16:39 -0500)] 
reorganize docs describing Kubernetes options and mycluster.yaml (#335)

* reorganize docs describing Kubernetes options and mycluster.yaml

* Expand dind documentation

* expand minikube docs

* expand docker-for-mac docs

5 weeks agoAdd php72 runtime (#334)
Carlos Santana [Mon, 5 Nov 2018 18:26:38 +0000 (13:26 -0500)] 
Add php72 runtime (#334)

* add php7.2 in runtimes.json

* add "sentinelledLogs": false, for java

* remove `"sentinelledLogs": false,` for java

5 weeks agoMaking the instructiosn for replacing port for docker-for-mac clearer (#333)
Neeraj Laad [Mon, 5 Nov 2018 17:04:34 +0000 (17:04 +0000)] 
Making the instructiosn for replacing port for docker-for-mac clearer (#333)

5 weeks agovariables in values.yaml should be in camelCase (#325)
David Grove [Fri, 2 Nov 2018 09:10:47 +0000 (05:10 -0400)] 
variables in values.yaml should be in camelCase (#325)

Per the Helm best practices document [1], variables in values.yaml
should be in camelCase

[1] https://github.com/helm/helm/blob/master/docs/chart_best_practices/values.md

5 weeks agorestore `set -x` that was removed by mistake (#326)
David Grove [Fri, 2 Nov 2018 02:18:09 +0000 (22:18 -0400)] 
restore `set -x` that was removed by mistake (#326)

5 weeks agofixes to include golang runtime (#327)
Sciabarra.com ltd [Thu, 1 Nov 2018 21:32:59 +0000 (22:32 +0100)] 
fixes to include golang runtime (#327)

5 weeks agoFix hostpath for /containers mount for dind cluster (#329)
David Grove [Thu, 1 Nov 2018 21:10:12 +0000 (17:10 -0400)] 
Fix hostpath for /containers mount for dind cluster (#329)

6 weeks agoUSE_HAIRPIN=true in dind-cluster command (#320)
David Grove [Wed, 24 Oct 2018 12:52:47 +0000 (08:52 -0400)] 
USE_HAIRPIN=true in dind-cluster command (#320)

6 weeks agoAdd an option include OpenWhisk system tests in TravisCI testing (#316)
David Grove [Wed, 24 Oct 2018 12:51:53 +0000 (08:51 -0400)] 
Add an option include OpenWhisk system tests in TravisCI testing (#316)

Unfortunately this additional level of testing is still disabled by
default because although it passes fairly reliably when run on a
development machine, it fails almost 100% of the time in the TravisCI
environment.

Also include a retry loop around the `wsk action list` smoketest in
buildHelm.sh and the `wsk package list` checks because there is
annoyingly often a delay between the entity being created successfully
and it showing up in `list` operations.

7 weeks agoAdd entry to values.yaml to configure Kubernetes DNS service (#319)
David Grove [Tue, 23 Oct 2018 12:30:16 +0000 (08:30 -0400)] 
Add entry to values.yaml to configure Kubernetes DNS service (#319)

Fixes #303
Fixes #311

7 weeks agoproper configuration of couchdb as a single node cluster (#318)
David Grove [Sat, 20 Oct 2018 22:01:39 +0000 (18:01 -0400)] 
proper configuration of couchdb as a single node cluster (#318)

7 weeks agoDocumentation enhancements (#317)
David Grove [Fri, 19 Oct 2018 12:30:48 +0000 (08:30 -0400)] 
Documentation enhancements (#317)

1. Document common development workflows when deploying on Kubernetes
2. Clarify selection of api-host for kubeadm-dind-cluster
3. Add some examples of good/bad controller logs to the troubleshooting
   section on the kubelet hairpin mode.

7 weeks agoSwitch from minikube to kubeadm-dind-cluster for TravisCI testing (#314)
David Grove [Thu, 18 Oct 2018 01:09:02 +0000 (21:09 -0400)] 
Switch from minikube to kubeadm-dind-cluster for TravisCI testing (#314)

A number of changes and improvements to TravisCI testing of Helm deployment:
  + Get our Kubernetes cluster via kubeadm-dind-cluster instead of minikube.
  + Now test Kubernetes versions 1.10 and 1.11 (drop 1.9; add 1.11).
  + Deploy providers in parallel to reduce testing latency
  + Enable apigateway sniff test (it reliably passes with kubeadm-dind-cluster)

Update main README and ingress and troubleshooting documentation to
reflect shift away from Minikube as the recommended environment.

2 months agofix typo in default db prefix in provider readmes (#312)
David Grove [Wed, 10 Oct 2018 02:34:16 +0000 (22:34 -0400)] 
fix typo in default db prefix in provider readmes (#312)

2 months agoingress instructions for deploying on AWS EKS (#306)
David Grove [Tue, 9 Oct 2018 21:18:34 +0000 (17:18 -0400)] 
ingress instructions for deploying on AWS EKS (#306)

2 months agoadd join-slack badge (#308)
David Grove [Fri, 5 Oct 2018 22:12:54 +0000 (18:12 -0400)] 
add join-slack badge (#308)

2 months agofixes for TravisCI runs (#309)
David Grove [Fri, 5 Oct 2018 20:56:52 +0000 (16:56 -0400)] 
fixes for TravisCI runs (#309)

2 months agoShuffle README to highlight Kubernetes on Docker option (#307)
David Grove [Wed, 3 Oct 2018 22:47:33 +0000 (18:47 -0400)] 
Shuffle README to highlight Kubernetes on Docker option (#307)

2 months agoDocument using the Kubernetes support available in Docker 18.06 (#304)
David Grove [Fri, 28 Sep 2018 23:16:07 +0000 (19:16 -0400)] 
Document using the Kubernetes support available in Docker 18.06 (#304)

2 months agoMake k8s domain name configurable in values.yaml (#299)
Ying Chun Guo [Thu, 20 Sep 2018 14:29:29 +0000 (22:29 +0800)] 
Make k8s domain name configurable in values.yaml (#299)

2 months agotravisCI improvements (#301)
David Grove [Thu, 20 Sep 2018 01:17:53 +0000 (21:17 -0400)] 
travisCI improvements (#301)

1. On minikube the default storageClass is called `standard`,
so we need to provide this value for those charts that require
PVs to deploy successfully.

2. Add a readiness check for a healthy invoker and gate the
catalog/routemgmt install jobs on that test.  This should cause
them to wait long enough to avoid whatever race condition is causing
them to often error out with 500 error code from couchdb.

2 months agoAdd travis test for providers (#296)
Ying Chun Guo [Fri, 14 Sep 2018 14:58:09 +0000 (22:58 +0800)] 
Add travis test for providers (#296)

2 months agoenable external db for alarm and kafka provider (#295)
Ying Chun Guo [Tue, 11 Sep 2018 16:00:32 +0000 (00:00 +0800)] 
enable external db for alarm and kafka provider (#295)

3 months agobug fix in wait-for-couchdb; improve logs of wait-for-X initContainers (#294)
David Grove [Thu, 6 Sep 2018 01:51:11 +0000 (21:51 -0400)] 
bug fix in wait-for-couchdb; improve logs of wait-for-X initContainers (#294)

Check for the presence of a synthetic marker database that is only created
after the ansible init_db and wipe_db jobs complete to avoid deciding
that CouchDB is ready while the wipe_db job is still actually running.

More descriptive messages from wait-for-X jobs while running
and always print a Success message when the job completes.

3 months ago(1) get utility image names from values.yaml; (2) change imagePullPolicy to Always...
David Grove [Mon, 3 Sep 2018 05:44:26 +0000 (01:44 -0400)] 
(1) get utility image names from values.yaml; (2) change imagePullPolicy to Always for all images with latest tag (#292)

3 months agoUse upstream script-runner and ansible-runner Docker images (#288)
David Grove [Wed, 29 Aug 2018 15:46:36 +0000 (11:46 -0400)] 
Use upstream script-runner and ansible-runner Docker images (#288)

3 months agoPlacate helm lint: Chart name must match directory name (#290)
David Grove [Wed, 29 Aug 2018 02:10:24 +0000 (22:10 -0400)] 
Placate helm lint: Chart name must match directory name (#290)

3 months agoenable node affinity for invoker-agent DaemonSet (#289)
David Grove [Wed, 29 Aug 2018 02:04:50 +0000 (22:04 -0400)] 
enable node affinity for invoker-agent DaemonSet (#289)

Upstream PR 3963 enhanced the KubernetesContainerFactory to
create user action pods with the proper node affinity, therefore
we can now restrict the invokerAgent to only run on invoker nodes.

3 months agoupdate for WorkloadAPI moving from beta to v1 in Kubernetes 1.9 (#287)
David Grove [Tue, 28 Aug 2018 13:58:58 +0000 (09:58 -0400)] 
update for WorkloadAPI moving from beta to v1 in Kubernetes 1.9 (#287)

Switch to stable v1 API for Workload objects (Deployment, DaemonSet,
StatefulSet, ReplicatSet).  This change requires Kubernetes version
1.9 or higher, so remove v1.8 from travisCI and documentation.

3 months agoenable cloudant provider package to use a different cloudant instance (#284)
Ying Chun Guo [Fri, 24 Aug 2018 14:35:56 +0000 (22:35 +0800)] 
enable cloudant provider package to use a different cloudant instance (#284)

3 months agoRevert the copyright to the content in the template (#285)
Vincent [Fri, 24 Aug 2018 01:06:37 +0000 (21:06 -0400)] 
Revert the copyright to the content in the template (#285)

3 months agoUpdate for upstream memory-based loadbalancing PR. (#286)
David Grove [Thu, 23 Aug 2018 21:12:21 +0000 (17:12 -0400)] 
Update for upstream memory-based loadbalancing PR. (#286)

3 months agoUse beta xenial dist to enable newer minikube versions in TravisCI (#282)
David Grove [Wed, 22 Aug 2018 18:42:54 +0000 (14:42 -0400)] 
Use beta xenial dist to enable newer minikube versions in TravisCI (#282)

3 months agoadd cloudant provider (#279)
Ying Chun Guo [Wed, 22 Aug 2018 13:25:03 +0000 (21:25 +0800)] 
add cloudant provider (#279)

3 months agobugfix in StatefulSet templates: use serviceName, not name (#280)
David Grove [Wed, 22 Aug 2018 01:25:14 +0000 (21:25 -0400)] 
bugfix in StatefulSet templates: use serviceName, not name (#280)

Fixes #278.

3 months agoRemove prototype log processing code from invoker agent (#275)
David Grove [Tue, 21 Aug 2018 07:15:24 +0000 (03:15 -0400)] 
Remove prototype log processing code from invoker agent (#275)

A better design is to focus on enhancing the upstream logging
infrastructure so that user container logs can be fed directly
into the platform logging service.  This prototype had unsolved
issues with log rotation (not supported), so removing the partially
completed functionality.

3 months agokafka provider readme (based on alarm provider readme). (#276)
David Grove [Tue, 21 Aug 2018 07:14:41 +0000 (03:14 -0400)] 
kafka provider readme (based on alarm provider readme). (#276)

3 months agoAdd Ruby 2.5 to the list of runtimes (#274)
Justin Halsall [Wed, 15 Aug 2018 01:01:55 +0000 (21:01 -0400)] 
Add Ruby 2.5 to the list of runtimes (#274)

4 months agoUse whisk.auth from secrets instead of the default auths in the git repo (#273)
David Grove [Fri, 10 Aug 2018 20:54:10 +0000 (16:54 -0400)] 
Use whisk.auth from secrets instead of the default auths in the git repo (#273)

A modification of the fix suggested in PR#271 that mounts the whisk.auth
secrets in a volume instead of putting them environment variables.

4 months agoset invoker's --uniqueName to $INVOKER_NAME; eliminate setting of COMPONENT_NAME...
David Grove [Mon, 6 Aug 2018 15:24:11 +0000 (11:24 -0400)] 
set invoker's --uniqueName to $INVOKER_NAME; eliminate setting of COMPONENT_NAME (#270)

4 months agoadd alarm provider (#265)
Ying Chun Guo [Thu, 2 Aug 2018 14:14:21 +0000 (22:14 +0800)] 
add alarm provider (#265)

4 months agoremove obsolete LIMITS_ACTIONS_INVOKES_CONCURRENTINSYSTEM (#269)
David Grove [Wed, 1 Aug 2018 01:31:39 +0000 (21:31 -0400)] 
remove obsolete LIMITS_ACTIONS_INVOKES_CONCURRENTINSYSTEM (#269)

This config values was made obsolete by upstream PR#3875.

4 months agoadd check for healthy invoker to build-helm.sh (#262)
David Grove [Wed, 1 Aug 2018 01:31:01 +0000 (21:31 -0400)] 
add check for healthy invoker to build-helm.sh (#262)

Use the controller's /invokers/healthy/count route to verify
that the controller has at least one healthy invoker before
attempting to proceed to the smoketest of invoking an action.

4 months agoFix for #266 and #267 (#268)
Carlos Duque [Tue, 31 Jul 2018 17:51:38 +0000 (18:51 +0100)] 
Fix for #266 and #267 (#268)

* fixed readiness probe for couch db issue: #266
* fixed invoker init.sh parameter needs uniqueName issue: #267

4 months agoExternal zookeeper kafka (#257)
Chris [Tue, 24 Jul 2018 00:35:33 +0000 (18:35 -0600)] 
External zookeeper kafka (#257)

* make kafka and zookeeper optional

* document optional kafka and zookeeper

* kafka.host to kafka.name, set zookeeper_connect and zookeeper_zero_host from zookeeper.name

* update docs with external kafka and zookeeper configuration examples

4 months agoMove db initialization into separate Job and reorder invoker init-containers (#261)
David Grove [Fri, 20 Jul 2018 15:10:33 +0000 (11:10 -0400)] 
Move db initialization into separate Job and reorder invoker init-containers (#261)

Restructure the db initialization into a separate Job using the
whisk-ansible-runner docker image that can be optionally run on both
"in cluster" and "external" database instances. In addition to
simplifying usage of an external database, this also allows the use of
a vanilla apache/couchdb image for in cluster CouchDB.

Also reorder invoker init-containers so we do the docker-pull-runtimes
before waiting for the controller to be up. This better overlaps the
docker pulls of the runtime containers with the inherent latency of
waiting for the controller to be ready.

4 months agorefactor of kube-deploy docker images (#249)
David Grove [Thu, 19 Jul 2018 13:32:04 +0000 (09:32 -0400)] 
refactor of kube-deploy docker images (#249)

Refactor kube docker images to reduce the number of unique docker
images by creating generic shell script and ansible playbook execution
images that will execute a script mounted into their /task volume by
the Job/Deployment.

Also switch the base image for these utility images from ubuntu to
alpine, which significantly decreases the image sizes.

4 months agoreorganize ingress documentation to group by ingress instead of cloud provider. ...
David Grove [Wed, 18 Jul 2018 14:50:08 +0000 (10:50 -0400)] 
reorganize ingress documentation to group by ingress instead of cloud provider. (#256)

4 months agoDocument expectations for PersistentVolume creation (#245)
David Grove [Wed, 18 Jul 2018 14:49:54 +0000 (10:49 -0400)] 
Document expectations for PersistentVolume creation (#245)

Fixes #244.

4 months agoingress.md information about google cloud DNS (#259)
Carlos Duque [Wed, 18 Jul 2018 14:18:55 +0000 (15:18 +0100)] 
ingress.md information about google cloud DNS (#259)

* Updated ingress.md documentation with link to google cloud documentation on DNS

4 months agoMissing documentation about ingress in previous pull request (#255)
Carlos Duque [Fri, 13 Jul 2018 22:00:38 +0000 (23:00 +0100)] 
Missing documentation about ingress in previous pull request (#255)

* Updated ingress.md documentation on how to configure the generic ingress values

4 months agoImprovement in default chart values and make #251 work by just pulling the chart...
Carlos Duque [Fri, 13 Jul 2018 20:41:14 +0000 (21:41 +0100)] 
Improvement in default chart values and make #251 work by just pulling the chart (#254)

* Changed ingress.yaml to be used in a generic way that instead of specific for ibm ingress
* updated default image pull policy to always for images using the lastest release tag
* updated gitignore to ignore myruntimes.json and.DS_Store

5 months agoSplit Helm chart for core OpenWhisk and providers (#248)
David Grove [Sat, 7 Jul 2018 17:20:22 +0000 (13:20 -0400)] 
Split Helm chart for core OpenWhisk and providers (#248)

Split the current single Helm chart for deploying OpenWhisk and its
providers into multiple Helm charts. In both production and CI/CD
environments it may be desirable to separately deploy the core
OpenWhisk system and the associated providers.

Fixes #246.

5 months agofix a typo in controller.yaml (#243)
Ying Chun Guo [Thu, 5 Jul 2018 12:08:07 +0000 (20:08 +0800)] 
fix a typo in controller.yaml (#243)

5 months agoadd explicit checks to ensure required variables are defined (#242)
David Grove [Wed, 4 Jul 2018 03:16:23 +0000 (23:16 -0400)] 
add explicit checks to ensure required variables are defined (#242)

Add explicit checks using the required operator for variables that
the user must define (ie, that have no sensible default). This enables
better error messages to be generated that can direct the user to the
relevant documentation on how to define the variable.

Fixes #240.

5 months agoHelm support for clustered zookeeper (#236)
David Grove [Wed, 4 Jul 2018 02:01:25 +0000 (22:01 -0400)] 
Helm support for clustered zookeeper (#236)

Allow zookeeper.replicaCount to be set to values > 1 by
properly configuring the zookeeper cluster and adjusting other
parts of the Helm chart so that dependent services can
properly communicate with the cluster.

Also document which of the deployments do not yet support
multiple replicas in docs/configurationChoices.md and
via comments in values.yaml.

5 months agoFinal removal of non-Helm based deployment (#239)
David Grove [Tue, 3 Jul 2018 13:34:39 +0000 (09:34 -0400)] 
Final removal of non-Helm based deployment (#239)

Helm-based deployment has strictly more capabilities than the non-Helm
based configuration files.  We already stopped testing the non-Helm
deployment and it has now become out-of-date with the upstream code
base.  Therefore it is time to remove the old style files to reduce
maintenance costs and avoid user confusion.

5 months agostore key OpenWhisk configuration values in ConfigMaps (#241)
David Grove [Tue, 3 Jul 2018 08:56:52 +0000 (04:56 -0400)] 
store key OpenWhisk configuration values in ConfigMaps (#241)

Go back to using ConfigMaps to store key cluster and database
configuration parameters and modify deployments to access the data
from those maps. Having explicit config maps makes it easier to
inspect the cluster configuration and enable separating the core
system and the providers into separate charts (subsequent PR after
this is merged).

Note: database and system authentication information continues to
be stored in secrets; that is unchanged by this PR.

5 months agoConfigure CLI's auth property directly from whisk.auth secret (#238)
David Grove [Thu, 28 Jun 2018 21:42:57 +0000 (17:42 -0400)] 
Configure CLI's auth property directly from whisk.auth secret (#238)

5 months agoUpdate for upstream move of version info to PureConfig (#234)
David Grove [Tue, 26 Jun 2018 10:57:26 +0000 (06:57 -0400)] 
Update for upstream move of version info to PureConfig (#234)

Update Helm chart for upstream PR 3617 that eliminated
whisk_version name and moved the version number and date
to PureConfig.

5 months agoTweaks to README based on user-feedback (#232)
David Grove [Sun, 24 Jun 2018 09:42:15 +0000 (05:42 -0400)] 
Tweaks to README based on user-feedback (#232)

Some improvements to helm instructions based on user comments
from issue #229.

5 months agoupdate docker-pull-runtimes for upstream change #3680 (#233)
David Grove [Fri, 22 Jun 2018 17:23:57 +0000 (13:23 -0400)] 
update docker-pull-runtimes for upstream change #3680 (#233)

Adapt the docker-pull-runtimes ansible script for the
upstream changes in the structure of runtimes.json made
in OpenWhisk PR #3680.

5 months agoBug fix for an issue in zookeeper volume mount (#231)
David Grove [Fri, 22 Jun 2018 07:52:06 +0000 (03:52 -0400)] 
Bug fix for an issue in zookeeper volume mount (#231)

Due to a typo in the pod specification, we were incorrectly mounting
the zk-data volume twice and not mounting the zk-datalog volume at
all.  This resulted on zookeeper crashing after restart because log
data had been written to the data directory.

Fixes #230.

5 months agoonly run deploy script for apache/incubator-openwhisk-deploy-kube:master (#228)
David Grove [Thu, 21 Jun 2018 05:07:27 +0000 (01:07 -0400)] 
only run deploy script for apache/incubator-openwhisk-deploy-kube:master (#228)

5 months agoUpdate runtimes for upstream changes. (#227)
rodric rabbah [Mon, 18 Jun 2018 21:27:51 +0000 (17:27 -0400)] 
Update runtimes for upstream changes. (#227)

5 months agoDeploy Kafka provider with Helm (#223)
Ying Chun Guo [Fri, 15 Jun 2018 15:09:49 +0000 (23:09 +0800)] 
Deploy Kafka provider with Helm (#223)

5 months agoreplace <release_name> with owdev in minikube commands (#222)
David Grove [Thu, 14 Jun 2018 09:01:42 +0000 (05:01 -0400)] 
replace <release_name> with owdev in minikube commands (#222)