Implement different driver modes and AWS Region override for controller service #438

rfranzke · 2020-01-10T09:28:38Z

Is this a bug fix or adding new feature?
Feature

Fixes #378

What is this PR about? / Why do we need it?
This PR allows to only run the controller service of the driver (and not the node service). Also, it allows to configure the AWS region instead of always looking it up via the AWS EC2 metadata service.

This enables the use-case of running the CSI controllers (csi-provisioner, csi-attacher, etc., + the driver controller) separately outside of the cluster, and not necessarily on an AWS EC2 instance.

What testing is done?
I ran the CSI controllers + EBS driver separately outside of the Kubernetes cluster (with the newly introduced flags of this PR). Also, I ran the EBS driver daemon set inside the cluster (together with the registrar) without the flags (i.e., with standard mode). I was successfully able to provision and deprovision EBS volumes.

k8s-ci-robot · 2020-01-10T09:28:46Z

Welcome @rfranzke!

It looks like this is your first PR to kubernetes-sigs/aws-ebs-csi-driver 🎉. Please refer to our pull request process documentation to help your PR have a smooth ride to approval.

You will be prompted by a bot to use commands during the review process. Do not be afraid to follow the prompts! It is okay to experiment. Here is the bot commands documentation.

You can also check if kubernetes-sigs/aws-ebs-csi-driver has its own contribution guidelines.

You may want to refer to our testing guide if you run into trouble with your tests not passing.

If you are having difficulty getting your pull request seen, please follow the recommended escalation practices. Also, for tips and tricks in the contribution process you may want to read the Kubernetes contributor cheat sheet. We want to make sure your contribution gets all the attention it needs!

Thank you, and welcome to Kubernetes. 😃

k8s-ci-robot · 2020-01-10T09:28:46Z

Hi @rfranzke. Thanks for your PR.

I'm waiting for a kubernetes-sigs member to verify that this patch is reasonable to test. If it is, they should reply with /ok-to-test on its own line. Until that is done, I will not automatically test new commits in this PR, but the usual testing commands by org members will still work. Regular contributors should join the org to skip this step.

Once the patch is verified, the new status will be reflected by the ok-to-test label.

I understand the commands that are listed here.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.

coveralls · 2020-01-10T09:31:41Z

Pull Request Test Coverage Report for Build 992

15 of 45 (33.33%) changed or added relevant lines in 4 files are covered.
No unchanged relevant lines lost coverage.
Overall coverage increased (+0.2%) to 70.953%

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
pkg/driver/fakes.go	0	1	0.0%
pkg/driver/driver.go	0	29	0.0%

Totals
Change from base Build 991:	0.2%
Covered Lines:	1258
Relevant Lines:	1773

💛 - Coveralls

rfranzke · 2020-01-10T09:37:42Z

/assign @leakingtapan

leakingtapan · 2020-01-11T00:54:58Z

@rfranzke This is a nice features! I like the way this PR is heading towards. Several comments:

Regarding adding the --region flag, IIUC, this is similar to what's asked
in Support create pvc in other regions #378. I think a better way to handle this is letting controller service to figure out the region through volume topology instead of either passing the region as cli flag or reading from IMDS. The way it will works is during provisioning in the CreateVolume method, controller service will read the region from the zone here, and create a new cloud client using the region, the volume will be created with the per region client.
Regarding the --run-as-controller flag, there are three modes to run the driver 1) controller service only 2) node service only and 3) both (the current mode). I would like a more expressive way to define the three modes than --run-as-controller. One way I'm thinking is we use subcommand to describe what service to run. To run just the controller service, we do:
/bin/aws-ebs-csi-driver controller. To run just the node service, we do: /bin/aws-ebs-csi-driver node. The benefit is we can cleanly separate the flags that matters to each service independently. Such that extra-volume-tags will only be a flag that is valid to the aws-ebs-csi-driver controller subcommand. And if no subcommand is provided, we will run the driver as current mode (both controller and node service).

leakingtapan · 2020-01-11T00:55:07Z

/ok-to-test

rfranzke · 2020-01-11T17:33:02Z

@leakingtapan Thanks for your feedback, I’ve updated the PR based on your suggestions wrt driver mode (all vs. controller vs. node) and implemented the subcommands.

Regarding the first suggestion wrt region:

The way it will works is during provisioning in the CreateVolume method, controller service will read the region from the zone here, and create a new cloud client using the region, the volume will be created with the per region client.

I don’t think this is possible, unfortunately. Checking the link you referred, I could certainly move the statement up and create a new cloud client with the region, but this only covers the CreateVolume case. What if a volume shall be deleted? As far as I can see no region information is provided in the DeleteVolumeRequest object but only the volume id. Where would I get the correct region from?

Another concern was that, even if the DeleteVolume case is solvable somehow, then we would have to remove the default cloud client as it tries to talk to the metadata service (which will not work if the controller does not run on an AWS EC2 instance). Instead, we would have to check in every function where d.cloud is used if it was initialized already, and if not yet then initialize it. (Yes, this is not a problem/argument against it, just something I noticed and wanted to mention).

Though, I have changed it from being a command line parameter to an environment variable (analogous to this code and as AWS_REGION is a pretty well-known environment variable for the use-case of overwriting the region).

Do you see an option to get rid of the “static” region overwriting via the env variable? I’m no CSI expert and not too familiar with the code overall, so chances are high that I’m missing something.

Happy to hear your feedback now!

PS: I don’t think the failing tests were related to my changes, were they?

rfranzke · 2020-01-11T18:29:22Z

/retest

leakingtapan · 2020-01-14T07:57:39Z

I don’t think this is possible, unfortunately. Checking the link you referred, I could certainly move the statement up and create a new cloud client with the region, but this only covers the CreateVolume case. What if a volume shall be deleted? As far as I can see no region information is provided in the DeleteVolumeRequest object but only the volume id. Where would I get the correct region from?

It's a good point for DeleteVolume. In my opinion, the CSI spec should have a way to pass in topology information (eg. zone) into DeleteVolume call. So that volume deletion could be handled correctly for volumes across different zones. Symmetrically, this seems necessary as a reverse operation of CreateVolume.

After a second thought, I think we don't need to block this by the above issue. Using AWS_REGION as an override to the IMDS region is acceptable to me. As we can solve the above problem later (if CSI spec) can support this. In the ideal world, the region resolution ordering will be 1) automatic resolution using topology 2) using AWS_REGION 3) using IMDS. And with this PR, we are adding option 2) which sounds good to me :)

leakingtapan · 2020-01-14T08:09:45Z

Created an issue in CSI: container-storage-interface/spec#408

aws-ebs-csi-driver/templates/manifest.yaml

aws-ebs-csi-driver/values.yaml

cmd/options.go

pkg/driver/driver.go

leakingtapan · 2020-01-14T08:29:42Z

Since coverage dropped, could you add some unit tests? thx for the contribution :)

rfranzke · 2020-01-26T09:08:02Z

@leakingtapan can you have another look, please?

leakingtapan · 2020-01-27T03:57:50Z

@rfranzke I see the coverage is still drop. Could you check where the coverage is missing? Once the coverage is back, you are good to go. Thx

rfranzke · 2020-01-27T07:41:34Z

@leakingtapan please check now :)

rfranzke · 2020-01-27T08:33:19Z

/retest

leakingtapan · 2020-01-27T17:26:16Z

/lgtm
/approve

k8s-ci-robot · 2020-01-27T17:26:32Z

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: leakingtapan, rfranzke

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

~~OWNERS~~ [leakingtapan]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

@jieyu

@jieyu

@jieyu

@jieyu

* chore: update aws-ebs-csi-driver from 0.4.0 to 0.5.0 [Documentation](https://github.com/kubernetes-sigs/aws-ebs-csi-driver/blob/v0.5.0/docs/README.md) filename | sha512 hash --------- | ------------ [v0.5.0.zip](https://github.com/kubernetes-sigs/aws-ebs-csi-driver/archive/v0.5.0.zip) | `c53327e090352a7f79ee642dbf8c211733f4a2cb78968ec688a1eade55151e65f1f97cd228d22168317439f1db9f3d2f07dcaa2873f44732ad23aaf632cbef3a` [v0.5.0.tar.gz](https://github.com/kubernetes-sigs/aws-ebs-csi-driver/archive/v0.5.0.tar.gz) | `ec4963d34c601cdf718838d90b8aa6f36b16c9ac127743e73fbe76118a606d41aced116aaaab73370c17bcc536945d5ccd735bc5a4a00f523025c8e41ddedcb8` * Add a cmdline option to add extra volume tags ([#353](kubernetes-sigs/aws-ebs-csi-driver#353), [@jieyu](https://github.com/jieyu)) * Switch to use kustomize for manifest ([#360](kubernetes-sigs/aws-ebs-csi-driver#360), [@leakingtapan](https://github.com/leakingtapan)) * enable users to set ec2-endpoint for nonstandard regions ([#369](kubernetes-sigs/aws-ebs-csi-driver#369), [@amdonov](https://github.com/amdonov)) * Add standard volume type ([#379](kubernetes-sigs/aws-ebs-csi-driver#379), [@leakingtapan](https://github.com/leakingtapan)) * Update aws sdk version to enable EKS IAM for SA ([#386](kubernetes-sigs/aws-ebs-csi-driver#386), [@leakingtapan](https://github.com/leakingtapan)) * Implement different driver modes and AWS Region override for controller service ([#438](kubernetes-sigs/aws-ebs-csi-driver#438), [@rfranzke](https://github.com/rfranzke)) * Add manifest files for snapshotter 2.0 ([#452](kubernetes-sigs/aws-ebs-csi-driver#452), [@leakingtapan](https://github.com/leakingtapan)) * Return success if instance or volume are not found ([#375](kubernetes-sigs/aws-ebs-csi-driver#375), [@bertinatto](https://github.com/bertinatto)) * Patch k8scsi sidecars CVE-2019-11255 ([#413](kubernetes-sigs/aws-ebs-csi-driver#413), [@jnaulty](https://github.com/jnaulty)) * Handle mount flags in NodeStageVolume ([#430](kubernetes-sigs/aws-ebs-csi-driver#430), [@bertinatto](https://github.com/bertinatto)) * Run upstream e2e test suites with migration ([#341](kubernetes-sigs/aws-ebs-csi-driver#341), [@wongma7](https://github.com/wongma7)) * Use new test framework for test orchestration ([#359](kubernetes-sigs/aws-ebs-csi-driver#359), [@leakingtapan](https://github.com/leakingtapan)) * Update to use 1.16 cluster with inline test enabled ([#362](kubernetes-sigs/aws-ebs-csi-driver#362), [@leakingtapan](https://github.com/leakingtapan)) * Enable leader election ([#380](kubernetes-sigs/aws-ebs-csi-driver#380), [@leakingtapan](https://github.com/leakingtapan)) * Update go mod and mount library ([#388](kubernetes-sigs/aws-ebs-csi-driver#388), [@leakingtapan](https://github.com/leakingtapan)) * Refactor NewCloud by pass in region ([#394](kubernetes-sigs/aws-ebs-csi-driver#394), [@leakingtapan](https://github.com/leakingtapan)) * helm: provide an option to set extra volume tags ([#396](kubernetes-sigs/aws-ebs-csi-driver#396), [@jieyu](https://github.com/jieyu)) * Allow override for csi-provisioner image ([#401](kubernetes-sigs/aws-ebs-csi-driver#401), [@gliptak](https://github.com/gliptak)) * Enable volume expansion e2e test for CSI migration ([#407](kubernetes-sigs/aws-ebs-csi-driver#407), [@leakingtapan](https://github.com/leakingtapan)) * Swith to use kops 1.16 ([#409](kubernetes-sigs/aws-ebs-csi-driver#409), [@leakingtapan](https://github.com/leakingtapan)) * Added tolerations for node support ([#420](kubernetes-sigs/aws-ebs-csi-driver#420), [@zerkms](https://github.com/zerkms)) * Update helm chart to better match available values and add the ability to add annotations ([#423](kubernetes-sigs/aws-ebs-csi-driver#423), [@krmichel](https://github.com/krmichel)) * [helm] Also add toleration support to controller ([#433](kubernetes-sigs/aws-ebs-csi-driver#433), [@jyaworski](https://github.com/jyaworski)) * Add ec2:ModifyVolume action ([#434](kubernetes-sigs/aws-ebs-csi-driver#434), [@zodiac12k](https://github.com/zodiac12k)) * Schedule the EBS CSI DaemonSet on all nodes by default ([#441](kubernetes-sigs/aws-ebs-csi-driver#441), [@pcfens](https://github.com/pcfens)) * bump chart version * chore: bump liveness probe from 1.1.0 to 2.0.0 - Introduce V(5) on the health check begin/success log lines to allow filtering of these entries from logs. If you would like to retain these log entries the action required would be to set `-v==5` or higher for the livenessprobe container. ([#57](kubernetes-csi/livenessprobe#57), [@stefansedich](https://github.com/stefansedich)) - Deprecated "--connection-timeout" argument has been removed. ([#59](kubernetes-csi/livenessprobe#59), [@msau42](https://github.com/msau42)) - Fix nil pointer bug when driver responds with not ready ([#58](kubernetes-csi/livenessprobe#58), [@scuzhanglei](https://github.com/scuzhanglei)) - Migrated to Go modules, so the source builds also outside of GOPATH. ([#53](kubernetes-csi/livenessprobe#53), [@pohly](https://github.com/pohly)) * chore: bump csi external-provisioner from 1.3.0 to 1.4.0 All external-provisioner versions < 1.4.0 are deprecated and will stop functioning in Kubernetes v1.20. See [#323](kubernetes-csi/external-provisioner#323) and [k/k#80978](kubernetes/kubernetes#80978) for more details. Upgrade your external-provisioner to v1.4+ before Kubernetes v1.20. None - Fixes migration scenarios for Topology, fstype, and accessmodes for the kubernetes.io/gce-pd in-tree plugin ([#277](kubernetes-csi/external-provisioner#277), [@davidz627](https://github.com/davidz627)) - Checks if volume content source is populated if creating a volume from a snapshot source. ([#283](kubernetes-csi/external-provisioner#283), [@zhucan](https://github.com/zhucan)) - Fixes issue when SelfLink removal is turned on in Kubernetes. ([#323](kubernetes-csi/external-provisioner#323), [@msau42](https://github.com/msau42)) - CSI driver can return `CreateVolumeResponse` with size 0, which means unknown volume size. In this case, Provisioner will use PVC requested size as PV size rather than 0 bytes ([#271](kubernetes-csi/external-provisioner#271), [@hoyho](https://github.com/hoyho)) - Fixed potential leak of volumes after CSI driver timeouts. ([#312](kubernetes-csi/external-provisioner#312), [@jsafrane](https://github.com/jsafrane)) - Fixes issue where provisioner provisions volumes for in-tree PVC's which have not been migrated ([#341](kubernetes-csi/external-provisioner#341), [@davidz627](https://github.com/davidz627)) - Send the CSI volume_id instead of PVC Name to the csi-driver in volumeCreate when datasource is PVC ([#310](kubernetes-csi/external-provisioner#310), [@Madhu-1](https://github.com/Madhu-1)) - Fixes nil pointer derefence in log when migration turned on ([#342](kubernetes-csi/external-provisioner#342), [@davidz627](https://github.com/davidz627)) - Handle deletion of CSI migrated volumes ([#273](kubernetes-csi/external-provisioner#273), [@ddebroy](https://github.com/ddebroy)) - Reduced logging noise of unrelated PVCs. Emit event on successful provisioning. ([#351](kubernetes-csi/external-provisioner#351), [@jsafrane](https://github.com/jsafrane)) - Added extra verification of source Snapshot and PersistentVolumeClaim before provisioning. ([#352](kubernetes-csi/external-provisioner#352), [@jsafrane](https://github.com/jsafrane)) * chore: bump attacher * Fixed handling of ControllerUnpublish errors. The attacher will retry to ControllerUnpublish a volume after any error except for NotFound. (#168, @jsafrane) * bump external-snapshotter from 1.1.0 to 1.2.2 Breaking Changes * Changes the API group name for the fake VolumeSnapshot object to "snapshot.storage.k8s.io" to be in-sync with the group name of the real VolumeSnapshot object. As a result, the generated interfaces for clientset and informers of VolumeSnapshot are also changed from "VolumeSnapshot" to "Snapshot". (#123, @xing-yang) New Features * Adds Finalizer on the snapshot source PVC to prevent it from being deleted when a snapshot is being created from it. (#47, @xing-yang) Other Notable Changes * Add Status subresource for VolumeSnapshot. (#121, @zhucan) * Cherry picks PR #138: Prebound snapshots will work correctly with CSI drivers that does not support ListSnasphots.(#156, @hakanmemisoglu) * Cherry picks PR #172: Added extra verification of source PersistentVolumeClaim before creating snapshot.(#173, @xing-yang) * bump external-resizer from 0.2.0 to 0.4.0 New Features * Add prometheus metrics to CSI external-resizer under the /metrics endpoint. This can be enabled via the "--metrics-address" and "--metrics-path" options. (#67, @saad-ali) Bug Fixes * Avoid concurrent processing of same PVCs (#6, @mlmhl) * Exit on CSI gRPC conn loss (#55, @ggriffiths) * Verify claimref associated with PVs before resizing (#57, @gnufied) Other Notable Changes * Migrated to Go modules, so the source builds also outside of GOPATH. (#60, @pohly) * feat(awsebscsiprovisioner): updated awsebscsiprovisioner flags - updated args as mentioned in comments - updated container versions as mentioned in the comments D2IQ-64990 #comment updated awsebscsiprovisioner pod arg * feat(awsebscsiprovisioner): added podAnnotations - added statefulSet.podAnnotations feature - added new roles and snapshotter-controller - added more values to be setable D2IQ-64992 #comment updated awsebscsiprovisioner to include statefulSet.podAnnotations * fix: added replacing system-x-critial replacement - this was added for being able to run the ct install / upgrade behaviour so that we also can test with that priorityClassName set pods, that normally get a system-node critial or system-cluster-critical priorityClassName set. These only will be allowed to run in namespace kube-system and that for we need to drop the priorityClassName here to null for our tests. - separated lint and install, as otherwise lint would fail because of the sed changes - exclude gcp-csi-driver [D2IQ-65765] [D2IQ-65765]: https://jira.d2iq.com/browse/D2IQ-65765 Co-authored-by: Sebastian Brandt <793580+sebbrandt87@users.noreply.github.com>

@jieyu

…bs-csi-driver from 0.4.0 to 0.5.0 [Documentation](https://github.com/kubernetes-sigs/aws-ebs-csi-driver/blob/v0.5.0/docs/README.md) filename | sha512 hash --------- | ------------ [v0.5.0.zip](https://github.com/kubernetes-sigs/aws-ebs-csi-driver/archive/v0.5.0.zip) | [v0.5.0.tar.gz](https://github.com/kubernetes-sigs/aws-ebs-csi-driver/archive/v0.5.0.tar.gz) | * Add a cmdline option to add extra volume tags ([#353](kubernetes-sigs/aws-ebs-csi-driver#353), [@jieyu](https://github.com/jieyu)) * Switch to use kustomize for manifest ([#360](kubernetes-sigs/aws-ebs-csi-driver#360), [@leakingtapan](https://github.com/leakingtapan)) * enable users to set ec2-endpoint for nonstandard regions ([#369](kubernetes-sigs/aws-ebs-csi-driver#369), [@amdonov](https://github.com/amdonov)) * Add standard volume type ([#379](kubernetes-sigs/aws-ebs-csi-driver#379), [@leakingtapan](https://github.com/leakingtapan)) * Update aws sdk version to enable EKS IAM for SA ([#386](kubernetes-sigs/aws-ebs-csi-driver#386), [@leakingtapan](https://github.com/leakingtapan)) * Implement different driver modes and AWS Region override for controller service ([#438](kubernetes-sigs/aws-ebs-csi-driver#438), [@rfranzke](https://github.com/rfranzke)) * Add manifest files for snapshotter 2.0 ([#452](kubernetes-sigs/aws-ebs-csi-driver#452), [@leakingtapan](https://github.com/leakingtapan)) * Return success if instance or volume are not found ([#375](kubernetes-sigs/aws-ebs-csi-driver#375), [@bertinatto](https://github.com/bertinatto)) * Patch k8scsi sidecars CVE-2019-11255 ([#413](kubernetes-sigs/aws-ebs-csi-driver#413), [@jnaulty](https://github.com/jnaulty)) * Handle mount flags in NodeStageVolume ([#430](kubernetes-sigs/aws-ebs-csi-driver#430), [@bertinatto](https://github.com/bertinatto)) * Run upstream e2e test suites with migration ([#341](kubernetes-sigs/aws-ebs-csi-driver#341), [@wongma7](https://github.com/wongma7)) * Use new test framework for test orchestration ([#359](kubernetes-sigs/aws-ebs-csi-driver#359), [@leakingtapan](https://github.com/leakingtapan)) * Update to use 1.16 cluster with inline test enabled ([#362](kubernetes-sigs/aws-ebs-csi-driver#362), [@leakingtapan](https://github.com/leakingtapan)) * Enable leader election ([#380](kubernetes-sigs/aws-ebs-csi-driver#380), [@leakingtapan](https://github.com/leakingtapan)) * Update go mod and mount library ([#388](kubernetes-sigs/aws-ebs-csi-driver#388), [@leakingtapan](https://github.com/leakingtapan)) * Refactor NewCloud by pass in region ([#394](kubernetes-sigs/aws-ebs-csi-driver#394), [@leakingtapan](https://github.com/leakingtapan)) * helm: provide an option to set extra volume tags ([#396](kubernetes-sigs/aws-ebs-csi-driver#396), [@jieyu](https://github.com/jieyu)) * Allow override for csi-provisioner image ([#401](kubernetes-sigs/aws-ebs-csi-driver#401), [@gliptak](https://github.com/gliptak)) * Enable volume expansion e2e test for CSI migration ([#407](kubernetes-sigs/aws-ebs-csi-driver#407), [@leakingtapan](https://github.com/leakingtapan)) * Swith to use kops 1.16 ([#409](kubernetes-sigs/aws-ebs-csi-driver#409), [@leakingtapan](https://github.com/leakingtapan)) * Added tolerations for node support ([#420](kubernetes-sigs/aws-ebs-csi-driver#420), [@zerkms](https://github.com/zerkms)) * Update helm chart to better match available values and add the ability to add annotations ([#423](kubernetes-sigs/aws-ebs-csi-driver#423), [@krmichel](https://github.com/krmichel)) * [helm] Also add toleration support to controller ([#433](kubernetes-sigs/aws-ebs-csi-driver#433), [@jyaworski](https://github.com/jyaworski)) * Add ec2:ModifyVolume action ([#434](kubernetes-sigs/aws-ebs-csi-driver#434), [@zodiac12k](https://github.com/zodiac12k)) * Schedule the EBS CSI DaemonSet on all nodes by default ([#441](kubernetes-sigs/aws-ebs-csi-driver#441), [@pcfens](https://github.com/pcfens)) * bump chart version * chore: bump liveness probe from 1.1.0 to 2.0.0 - Introduce V(5) on the health check begin/success log lines to allow filtering of these entries from logs. If you would like to retain these log entries the action required would be to set or higher for the livenessprobe container. ([#57](kubernetes-csi/livenessprobe#57), [@stefansedich](https://github.com/stefansedich)) - Deprecated --connection-timeout argument has been removed. ([#59](kubernetes-csi/livenessprobe#59), [@msau42](https://github.com/msau42)) - Fix nil pointer bug when driver responds with not ready ([#58](kubernetes-csi/livenessprobe#58), [@scuzhanglei](https://github.com/scuzhanglei)) - Migrated to Go modules, so the source builds also outside of GOPATH. ([#53](kubernetes-csi/livenessprobe#53), [@pohly](https://github.com/pohly)) * chore: bump csi external-provisioner from 1.3.0 to 1.4.0 All external-provisioner versions < 1.4.0 are deprecated and will stop functioning in Kubernetes v1.20. See [#323](kubernetes-csi/external-provisioner#323) and [k/k#80978](kubernetes/kubernetes#80978) for more details. Upgrade your external-provisioner to v1.4+ before Kubernetes v1.20. None - Fixes migration scenarios for Topology, fstype, and accessmodes for the kubernetes.io/gce-pd in-tree plugin ([#277](kubernetes-csi/external-provisioner#277), [@davidz627](https://github.com/davidz627)) - Checks if volume content source is populated if creating a volume from a snapshot source. ([#283](kubernetes-csi/external-provisioner#283), [@zhucan](https://github.com/zhucan)) - Fixes issue when SelfLink removal is turned on in Kubernetes. ([#323](kubernetes-csi/external-provisioner#323), [@msau42](https://github.com/msau42)) - CSI driver can return with size 0, which means unknown volume size. In this case, Provisioner will use PVC requested size as PV size rather than 0 bytes ([#271](kubernetes-csi/external-provisioner#271), [@hoyho](https://github.com/hoyho)) - Fixed potential leak of volumes after CSI driver timeouts. ([#312](kubernetes-csi/external-provisioner#312), [@jsafrane](https://github.com/jsafrane)) - Fixes issue where provisioner provisions volumes for in-tree PVC's which have not been migrated ([#341](kubernetes-csi/external-provisioner#341), [@davidz627](https://github.com/davidz627)) - Send the CSI volume_id instead of PVC Name to the csi-driver in volumeCreate when datasource is PVC ([#310](kubernetes-csi/external-provisioner#310), [@Madhu-1](https://github.com/Madhu-1)) - Fixes nil pointer derefence in log when migration turned on ([#342](kubernetes-csi/external-provisioner#342), [@davidz627](https://github.com/davidz627)) - Handle deletion of CSI migrated volumes ([#273](kubernetes-csi/external-provisioner#273), [@ddebroy](https://github.com/ddebroy)) - Reduced logging noise of unrelated PVCs. Emit event on successful provisioning. ([#351](kubernetes-csi/external-provisioner#351), [@jsafrane](https://github.com/jsafrane)) - Added extra verification of source Snapshot and PersistentVolumeClaim before provisioning. ([#352](kubernetes-csi/external-provisioner#352), [@jsafrane](https://github.com/jsafrane)) * chore: bump attacher * Fixed handling of ControllerUnpublish errors. The attacher will retry to ControllerUnpublish a volume after any error except for NotFound. (#168, @jsafrane) * bump external-snapshotter from 1.1.0 to 1.2.2 Breaking Changes * Changes the API group name for the fake VolumeSnapshot object to snapshot.storage.k8s.io to be in-sync with the group name of the real VolumeSnapshot object. As a result, the generated interfaces for clientset and informers of VolumeSnapshot are also changed from VolumeSnapshot to Snapshot. (#123, @xing-yang) New Features * Adds Finalizer on the snapshot source PVC to prevent it from being deleted when a snapshot is being created from it. (#47, @xing-yang) Other Notable Changes * Add Status subresource for VolumeSnapshot. (#121, @zhucan) * Cherry picks PR #138: Prebound snapshots will work correctly with CSI drivers that does not support ListSnasphots.(#156, @hakanmemisoglu) * Cherry picks PR #172: Added extra verification of source PersistentVolumeClaim before creating snapshot.(#173, @xing-yang) * bump external-resizer from 0.2.0 to 0.4.0 New Features * Add prometheus metrics to CSI external-resizer under the /metrics endpoint. This can be enabled via the --metrics-address and --metrics-path options. (#67, @saad-ali) Bug Fixes * Avoid concurrent processing of same PVCs (#6, @mlmhl) * Exit on CSI gRPC conn loss (#55, @ggriffiths) * Verify claimref associated with PVs before resizing (#57, @gnufied) Other Notable Changes * Migrated to Go modules, so the source builds also outside of GOPATH. (#60, @pohly) * feat(awsebscsiprovisioner): updated awsebscsiprovisioner flags - updated args as mentioned in comments - updated container versions as mentioned in the comments D2IQ-64990 #comment updated awsebscsiprovisioner pod arg * feat(awsebscsiprovisioner): added podAnnotations - added statefulSet.podAnnotations feature - added new roles and snapshotter-controller - added more values to be setable D2IQ-64992 #comment updated awsebscsiprovisioner to include statefulSet.podAnnotations * fix: added replacing system-x-critial replacement - this was added for being able to run the ct install / upgrade behaviour so that we also can test with that priorityClassName set pods, that normally get a system-node critial or system-cluster-critical priorityClassName set. These only will be allowed to run in namespace kube-system and that for we need to drop the priorityClassName here to null for our tests. - separated lint and install, as otherwise lint would fail because of the sed changes - exclude gcp-csi-driver [D2IQ-65765] [D2IQ-65765]: https://jira.d2iq.com/browse/D2IQ-65765 Co-authored-by: Sebastian Brandt <793580+sebbrandt87@users.noreply.github.com>

k8s-ci-robot added the cncf-cla: yes Indicates the PR's author has signed the CNCF CLA. label Jan 10, 2020

k8s-ci-robot added the needs-ok-to-test Indicates a PR that requires an org member to verify it is safe to test. label Jan 10, 2020

k8s-ci-robot requested review from bertinatto and d-nishi January 10, 2020 09:28

k8s-ci-robot added the size/M Denotes a PR that changes 30-99 lines, ignoring generated files. label Jan 10, 2020

rfranzke force-pushed the feature/run-as-controller branch from 79cb969 to 2159cdf Compare January 10, 2020 09:36

k8s-ci-robot assigned leakingtapan Jan 10, 2020

k8s-ci-robot added ok-to-test Indicates a non-member PR verified by an org member that is safe to test. and removed needs-ok-to-test Indicates a PR that requires an org member to verify it is safe to test. labels Jan 11, 2020

rfranzke force-pushed the feature/run-as-controller branch from 2159cdf to ccd4753 Compare January 11, 2020 17:32

k8s-ci-robot added size/L Denotes a PR that changes 100-499 lines, ignoring generated files. and removed size/M Denotes a PR that changes 30-99 lines, ignoring generated files. labels Jan 11, 2020

rfranzke changed the title ~~Allow running the controller service separately~~ Implement different driver modes Jan 11, 2020

leakingtapan reviewed Jan 14, 2020

View reviewed changes

aws-ebs-csi-driver/templates/manifest.yaml Outdated Show resolved Hide resolved

aws-ebs-csi-driver/values.yaml Outdated Show resolved Hide resolved

cmd/options.go Show resolved Hide resolved

pkg/driver/driver.go Show resolved Hide resolved

leakingtapan changed the title ~~Implement different driver modes~~ Implement different driver modes and AWS Region override for controller service Jan 14, 2020

rfranzke force-pushed the feature/run-as-controller branch from ccd4753 to 2a049e4 Compare January 19, 2020 12:03

k8s-ci-robot added size/XL Denotes a PR that changes 500-999 lines, ignoring generated files. and removed size/L Denotes a PR that changes 100-499 lines, ignoring generated files. labels Jan 19, 2020

rfranzke requested a review from leakingtapan January 19, 2020 12:03

rfranzke force-pushed the feature/run-as-controller branch 3 times, most recently from f86c6e2 to 25e4df2 Compare January 19, 2020 12:33

Implement operating modes for the CSI driver

3f65bd5

rfranzke force-pushed the feature/run-as-controller branch from 25e4df2 to 3f65bd5 Compare January 27, 2020 07:35

k8s-ci-robot added the lgtm "Looks good to me", indicates that a PR is ready to be merged. label Jan 27, 2020

k8s-ci-robot added the approved Indicates a PR has been approved by an approver from all required OWNERS files. label Jan 27, 2020

k8s-ci-robot merged commit 04143df into kubernetes-sigs:master Jan 27, 2020

rfranzke deleted the feature/run-as-controller branch January 27, 2020 18:48

joejulian mentioned this pull request Mar 6, 2020

chore: update awsebscsiprovisioner chart mesosphere/charts#492

Merged

This was referenced Apr 9, 2020

Allow disabling controller/node service via command line flags kubernetes-sigs/azuredisk-csi-driver#354

Closed

[cinder-csi-plugin] Allow disabling particular driver services kubernetes/cloud-provider-openstack#1021

Closed

jacobwolfaws mentioned this pull request May 12, 2023

Adopting Multiple CSI Driver improvements kubernetes-sigs/aws-fsx-csi-driver#309

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Implement different driver modes and AWS Region override for controller service #438

Implement different driver modes and AWS Region override for controller service #438

rfranzke commented Jan 10, 2020 •

edited

Loading

k8s-ci-robot commented Jan 10, 2020

k8s-ci-robot commented Jan 10, 2020

coveralls commented Jan 10, 2020 •

edited

Loading

rfranzke commented Jan 10, 2020

leakingtapan commented Jan 11, 2020 •

edited

Loading

leakingtapan commented Jan 11, 2020

rfranzke commented Jan 11, 2020

rfranzke commented Jan 11, 2020

leakingtapan commented Jan 14, 2020

leakingtapan commented Jan 14, 2020

leakingtapan commented Jan 14, 2020 •

edited

Loading

rfranzke commented Jan 26, 2020

leakingtapan commented Jan 27, 2020

rfranzke commented Jan 27, 2020

rfranzke commented Jan 27, 2020

leakingtapan commented Jan 27, 2020

k8s-ci-robot commented Jan 27, 2020

Implement different driver modes and AWS Region override for controller service #438

Implement different driver modes and AWS Region override for controller service #438

Conversation

rfranzke commented Jan 10, 2020 • edited Loading

k8s-ci-robot commented Jan 10, 2020

k8s-ci-robot commented Jan 10, 2020

coveralls commented Jan 10, 2020 • edited Loading

Pull Request Test Coverage Report for Build 992

💛 - Coveralls

rfranzke commented Jan 10, 2020

leakingtapan commented Jan 11, 2020 • edited Loading

leakingtapan commented Jan 11, 2020

rfranzke commented Jan 11, 2020

rfranzke commented Jan 11, 2020

leakingtapan commented Jan 14, 2020

leakingtapan commented Jan 14, 2020

leakingtapan commented Jan 14, 2020 • edited Loading

rfranzke commented Jan 26, 2020

leakingtapan commented Jan 27, 2020

rfranzke commented Jan 27, 2020

rfranzke commented Jan 27, 2020

leakingtapan commented Jan 27, 2020

k8s-ci-robot commented Jan 27, 2020

rfranzke commented Jan 10, 2020 •

edited

Loading

coveralls commented Jan 10, 2020 •

edited

Loading

leakingtapan commented Jan 11, 2020 •

edited

Loading

leakingtapan commented Jan 14, 2020 •

edited

Loading