OpenEBS Project RoadMap Discussion Notes
===
:::info
- **Location:** https://meet.google.com/vid-sfmt-igs?hs=122
- **Date:** Monthly on third Tuesday from 8:30pm to 09:00pm (IST)
- **Agenda**
1. Walk through the status update
2. Discussion topic

*Add your request, question or suggestion to our [issue list](https://github.com/openebs/openebs/issues)*

*If you have anything you’d like to put on the agenda, please do so below for the next meeting:*
:::

## Feb 15, 2022
Attendees: Pawan, Akhil, Harsh, Kiran
Topics/Notes:
* Review the updated ROADMAP (https://github.com/openebs/openebs/pull/3519)
* Migration to CNCF Infra
* Running manual tests - able to create clusters.
* Next step is to run the pipeline
* Website PRs failing
* Netlify builds are failing
* CNCF updates
* Have to sync-up with CNCF-TAG to check for updates.
* Community Health?
* Mostly setup and feature clarifications
* Response time has increased comparatively in the past 2/3 months
* Community tracker not updated.

## Jan 18, 2022

Note: Starting this month converting this into a monthly maintainer sync-up for
all OpenEBS projects.

* Review the ROADMAP.md (https://github.com/openebs/openebs/blob/main/ROADMAP.md)
* Brainstorming on 3.2
* How to progress on the PVC cleanup tool?
* Check the K8s 1.23 feature on STS PVC cleanup and how this tool can compliment that feature.
* Shared VG and Data Populator
* Migration of e2e infra to CNCF hosted machines
* Mayastor feature updates
* Days 2 operations around XFS quota
* CLI updates?
* Jiva CSI Driver GA
* Incubation update
* Need to come up with an approach on how to call the umbrella project as incubating when sub-projects in different stages of development.
* Learnings from of 3.1?
* Had issues with e2e pipeline
* The previous process followed in monthly releases was good for quarterly release.

## Oct 19, 2021

Local PV:
1. Shared VG Support for LVM: https://github.com/openebs/lvm-localpv/issues/134
- Migration of LVM localpv volumes across nodes.
2. Unified LocalPV CSI Driver
3. ZFS 2.0 support. (Test)
4. Device LocalPV (refactor and use NDM as a library)
5. Document (ZFS), support for 3rd party backup without using velero
6. Rawfile: support for capacity based scheduling
check for k8s deprecations and support for new flavours.

NDM:
1. Fixing client call bugs : https://github.com/openebs/node-disk-manager/pull/648
2. Supprot for partition table UUID as a feature
3. Creating a base image for building NDM images
4. Fixing the NDM operator to have proper requeue / retry intervals.
5. adding location attributes as BD labels: https://github.com/openebs/node-disk-manager/pull/618

## Aug 17, 2021

Attendees: Akhil, Kiran, Harsh, Tom, Pawan, Sahil

Agenda:
- Discussion on including support for Shared Volume Groups into LVM CSI Driver
- https://github.com/Seagate/propeller
- What is the topology of nodes and devices? How are devices shared?
- SAS - upto 6-8 hosts or 3-4 hosts
- affinity with manual
- 2 sets of nodes with different devices
- How is new lvmlockd deployed into K8s clusters?
- https://github.com/Seagate/csiclvm
- VG creation - outside the scope of driver
- Detection of devices
- Can it create VGs with a prefix pattern
- VG creation with shared
- Any changes to LV, thin provisioning, snapshot commands
- Will there be any changes required thin provisioning related commands?
- LV (-an)
- should be disabled
- on publish enable it
- How is the node topology on the local volume setup?
- 6 pack
- iSCSI or NVMe
- DRBD replication
- Resiliency cases
- Lock/lease timeout
- node failure or node failure / unreachable cases
- https://openebs.io/blog/logical-volume-management-lvm-on-kubernetes-with-openebs-localpv/
- https://github.com/openebs/lvm-localpv
- https://github.com/openebs/lvm-localpv/blob/master/pkg/driver/controller.go#L354

- Node Disk Manager
- https://github.com/openebs/node-disk-manager/
- Encryption at rest with Kubernetes Secret
- Disaggregated Storage Control Plane

- Review the OpenEBS 3.0 and 4.0 planned items
- OpenEBS 3.0 Pending Features
- Rawfile Local PV (CSI)
- (Not started) Configurable Backend Directory (Mehran)
- Hostpath Local PV
- (Inprogress) XFS Quota ( Abishek/Almas)
- https://github.com/openebs/dynamic-localpv-provisioner/pull/78
- Device Local PV (NDM)
- (Inprogress) Online Detection of Device Changes (Aditya/Akhil)
- Capacity detection (with udev)
- Unmount is detected
- [experimenting] wipefs detection
- (Done) Dashboard for NDM Metrics to display status of already exposed metrics [Abishek]
- https://github.com/openebs/monitoring/pull/53
- https://github.com/openebs/node-disk-manager/pull/612
- (Not started) Tag based on Drive Type
- https://github.com/openebs/dynamic-localpv-provisioner/issues/35
- https://github.com/openebs/node-disk-manager/issues/615
- (Not started) Warning/Error log when requested FSType doesn't match the hostpath FSType
- Covered by: https://github.com/openebs/dynamic-localpv-provisioner/issues/35
- OpenEBS 4.0
- Local (Consolidated) CSI
- ZFS Local PV
- ZFS 2.0 support
- Device Local PV (CSI)
- refactor NDM to use a library (instead of seperate pods/CRs)
- LVM Local PV
- Shared LVM
- Migration of Local PVs across nodes (on hold)
- Rawfile Local PV (CSI)
- Capacity Based Scheduling (Mehran)
- Hostpath Local PV
- [Design in progress] Backup / Restore via Datapopulator (Shovan)
- Device Local PV (NDM)
- Replaced with Local (Consolidated) CSI for devices
- NDM
- Handle devices shared between nodes
- Manage Storage Pool

## July 20, 2021

Attendees: Akhil, Kiran, Aditya

- Discussion Items
- OpenEBS 3.0 Pending Features
- Device Local PV (CSI)
- Helm Chart (Pushed into next version, as part of consolidation)
- Rawfile Local PV (CSI)
- Configurable Backend Directory (Mehran)
- Hostpath Local PV
- XFS Quota ( Abishek/Almas)
- Device Local PV (NDM)
- Online Detection of Device Changes (Aditya/Akhil)
- Dashboard for NDM Metrics to display status of already exposed metrics
- Tag based on Drive Type ( https://github.com/openebs/dynamic-localpv-provisioner/issues/35)
- Warning/Error log when requested FSType doesn't match the hostpath FSType
- OpenEBS 4.0
- Local (Consolidated) CSI
- ZFS Local PV
- Device Local PV (with reuse of NDM )
- LVM Local PV
- Rawfile Local PV (CSI)
- Capacity Based Scheduling (Mehran)
- Hostpath Local PV
- Backup / Restore via Datapopulator
- Device Local PV (NDM)
- Replaced with Local (Consolidated) CSI for devices
- NDM
- Handle devices shared between nodes
- Manage Storage Pool

- Suppor for multiple points
- should we consider displaying all this as part of the BD
- One use could be to add additional checks to avoid creating Device Local PV when there are multiple mount points

- Setup a program for attracting contributions
- GSoD
- YouTubers
-

####### Discussion Notes prior to making the meeting as LocalPV RoadMap discussion ######
## June 15, 2021

Attendees: Akhil, Harsh, Kiran, Aditya

- [Akhil]
- doc for vritual disk expansion. partition also needs to be expanded manually (not started)
- moved to github projects since this is not on priority
- LocalPV, passing custom node selector to localpv-device (already done for localpv-hostpath)
- 5/18: check if any nodes has labels specified in the label selector of BDC, then select BDs from those nodes.
- LocalPV , select drivetype. Ref: https://github.com/openebs/dynamic-localpv-provisioner/issues/35
- 5/18: one method is NDM take up a configuration and tag the BDs based on the configuration and can use the block-device-tag feature.
- NDM not able to detect cstor installed in the disk. Ref: https://github.com/openebs/openebs/issues/3389
- this issue can be fixed by restarting NDM pod. was not able to reproduce in the lab clusters

- [Harsh]
- display capacity.storage in correct units. (blocked)
- Can use resource.quantity from api-machinery (need to migrate CR to next version v1alpha1 to v1)
- 1/19: with CRD migration/validation done. This is now waiting on operator SDK upgrade. migrate BD and BDC to v1 / v1beta1 spec. ALso need to handle the conversion cases
- Upgrade to Operator SDK v1.0 - which is integrated into kube-builder (not started)
- sig-api => coverging on kube-builder and operatorsdk (follow this thread)
- The CR (BD and BDC) migration to v1 will be done after upgrading to operator SDK 1.0
- 1/19: no updates.
- 2/16: plan is to have operator SDK 1.0 (or latest version) for this release.
- 3/16: https://github.com/openebs/node-disk-manager/pull/552. operatorSDK mirgation and CRDs to v1
- 4/20: PR is in review. https://github.com/openebs/node-disk-manager/issues/558
- 5/18: Akhil to review the PR about patching the resources., blocking the unit tests cases.
- 6/15: unit tests are fixed, working on integration tests.

- [Aditya]
- explore on using polling in for size changes (in progress )
- did some initial research on what to use: explored inotify, and kmesg parser
- need to check the kmesg parser used in node problem detector
- 11/17: Explore the kmesg parser way of using as used in node problem detector
- 12/17: no updates
- 1/19: no updates. Can ask to sig-node team to get a clear idea on implementation.
- 2/16: Akhil will try to explore on this issue. need to modify the kmesg parser package and reuse it.
- 3/16: Submitted in Gsoc as part of CNCF mentoring.
- 4/20: In review phase
- 5/18: findmnt command can be explored, to check for mountpoints of the device.
- 6/15: will be starting integration tests for the mount change detection
- mulitple mount points for blockdevice needs to be updated in the blockdevice resource.

Discussion:
- Concluded that this meeting going forward will be about roadmap discussion of localpv (CSI and non-CSI) in general. The format of this doc
will also be changed moving forward.

## May 18, 2021

Attendees: Akhil, Harsh, Pawan, Kiran, Aditya

- [Akhil]
- Add a note in NDM docs / OpenEBS install section about Maystor and ZFS Local PV not yet integrated with NDM. (not-started)
- 5/18: revamping docs, will be added as part of the docs update
- doc for vritual disk expansion. partition also needs to be expanded manually (not started)
- ci can be used for staging. created an issue for integration testing in github actions (done)
- 4/20: cstor, libcstor, istgt the PRs have been raised to migrate to github actions.Ref : https://github.com/openebs/openebs/issues/3352
- 5/18: Done. All repositories migrated to github actions
- LocalPV, passing custom node selector to localpv-device (already done for localpv-hostpath)
- 5/18: check if any nodes has labels specified in the label selector of BDC, then select BDs from those nodes.
- LocalPV , select drivetype. Ref: https://github.com/openebs/dynamic-localpv-provisioner/issues/35
- 5/18: one method is NDM take up a configuration and tag the BDs based on the configuration and can use the block-device-tag feature.
- NDM not able to detect cstor installed in the disk. Ref: https://github.com/openebs/openebs/issues/3389

- [Harsh]
- display capacity.storage in correct units. (blocked)
- Can use resource.quantity from api-machinery (need to migrate CR to next version v1alpha1 to v1)
- 1/19: with CRD migration/validation done. This is now waiting on operator SDK upgrade. migrate BD and BDC to v1 / v1beta1 spec. ALso need to handle the conversion cases
- Upgrade to Operator SDK v1.0 - which is integrated into kube-builder (not started)
- sig-api => coverging on kube-builder and operatorsdk (follow this thread)
- The CR (BD and BDC) migration to v1 will be done after upgrading to operator SDK 1.0
- 1/19: no updates.
- 2/16: plan is to have operator SDK 1.0 (or latest version) for this release.
- 3/16: https://github.com/openebs/node-disk-manager/pull/552. operatorSDK mirgation and CRDs to v1
- 4/20: PR is in review. https://github.com/openebs/node-disk-manager/issues/558
- 5/18: Akhil to review the PR about patching the resources., blocking the unit tests cases.

- [Aditya]
- explore on using polling in for size changes (in progress )
- did some initial research on what to use: explored inotify, and kmesg parser
- need to check the kmesg parser used in node problem detector
- 11/17: Explore the kmesg parser way of using as used in node problem detector
- 12/17: no updates
- 1/19: no updates. Can ask to sig-node team to get a clear idea on implementation.
- 2/16: Akhil will try to explore on this issue. need to modify the kmesg parser package and reuse it.
- 3/16: Submitted in Gsoc as part of CNCF mentoring.
- 4/20: In review phase
- 5/18: findmnt command can be explored, to check for mountpoints of the device.
- mulitple mount points for blockdevice needs to be updated in the blockdevice resource.

- [Mehran] (no updates)
- Implement CSI storage capacity tracking (rawfile-localpv).(no updates)
- Main goal: prevent overcommitting resources
- Blocked by distributed external resizer. Ref: https://github.com/kubernetes-csi/external-resizer/issues/142
- Successfully used btrfs snapshotting features for creating consistent backups.
- Need ideas on how it may be done on other filesystems.

- [Pawan]
- Unified CSI Driver for LocalPVs (in design and architecture phase.)

- [Kiran] (no updates)
- StatefulSet with Local PV - on nodes that are recreated during upgrades
- STS pods is running Local PV - 2TB - with ML
- When node restarts - the Pod waits and begins running.
- When node upgrade happens - the platform - addes a new node with a new disk with upgraded version.
- Deletes old node with old k8s version
- STS pod on old node-name gets stuck
- STS pods should move to the new node.
- Delete the pod and its PVC - expect the stateful set to recreate the pod and Local PV on new node.
- If we use the custom_toplogy key, add the add to the new replaced node. This will automtically bring up the pod - with empty data - rebuild will start.

- volume populator to migrate data ? Check with Shovan

Discussions:
- Device Plugin in k8s (discuss with sig-node)
-

## Apr 20, 2021

Attendees: Akhil, Harsh, Peeyush, Mehran

- [Akhil]
- Add a note in NDM docs / OpenEBS install section about Maystor and ZFS Local PV not yet integrated with NDM. (not-started)
- doc for vritual disk expansion. partition also needs to be expanded manually (not started)
- ci can be used for staging. created an issue for integration testing in github actions
- 4/20: cstor, libcstor, istgt the PRs have been raised to migrate to github actions.Ref : https://github.com/openebs/openebs/issues/3352
- check configmap is auto loaded without restarting pod(no updates/ not started).
- implement used by probe for jiva(not yet started).
- add support for md array (not started)
- LocalPV, passing custom node selector to localpv-device (already done for localpv-hostpath)
- LocalPV , select drivetype. Ref: https://github.com/openebs/dynamic-localpv-provisioner/issues/35
- Split the openebs/charts to use the NDM helm charts (draft PR is raised, still discussion are going on)
- explore rancher dependency manager for helm charts
- Some additional functionalities of NDM is required to use it with CSI LocalPV variants for preparing the disks.(in design phase)
- remove CRD installation from NDM operator, upgrade CRDs from v1beta1 to v1(done)

- [Harsh]
- display capacity.storage in correct units. (blocked)
- Can use resource.quantity from api-machinery (need to migrate CR to next version v1alpha1 to v1)
- 1/19: with CRD migration/validation done. This is now waiting on operator SDK upgrade. migrate BD and BDC to v1 / v1beta1 spec. ALso need to handle the conversion cases
- Upgrade to Operator SDK v1.0 - which is integrated into kube-builder (not started)
- sig-api => coverging on kube-builder and operatorsdk (follow this thread)
- The CR (BD and BDC) migration to v1 will be done after upgrading to operator SDK 1.0
- 1/19: no updates.
- 2/16: plan is to have operator SDK 1.0 (or latest version) for this release.
- 3/16: https://github.com/openebs/node-disk-manager/pull/552. operatorSDK mirgation and CRDs to v1
- 4/20: PR is in review. https://github.com/openebs/node-disk-manager/issues/558

- [Aditya]
- explore on using polling in for size changes (in progress )
- did some initial research on what to use: explored inotify, and kmesg parser
- need to check the kmesg parser used in node problem detector
- 11/17: Explore the kmesg parser way of using as used in node problem detector
- 12/17: no updates
- 1/19: no updates. Can ask to sig-node team to get a clear idea on implementation.
- 2/16: Akhil will try to explore on this issue. need to modify the kmesg parser package and reuse it.
- 3/16: Submitted in Gsoc as part of CNCF mentoring.
- 4/20: In review phase

- [Mehran]
- Implement CSI storage capacity tracking (rawfile-localpv).
- Main goal: prevent overcommitting resources
- Blocked by distributed external resizer. Ref: https://github.com/kubernetes-csi/external-resizer/issues/142
- Successfully used btrfs snapshotting features for creating consistent backups.
- Need ideas on how it may be done on other filesystems.

Discussions:
- Device Plugin in k8s (discuss with sig-node)
-

## Mar 16, 2021

Attendees: Akhil, Harsh

- [Akhil]
- Add a note in NDM docs / OpenEBS install section about Maystor and ZFS Local PV not yet integrated with NDM. (not-started)
- doc for vritual disk expansion. partition also needs to be expanded manually (not started)
- ci can be used for staging. created an issue for integration testing in github actions. (work not started)
- 12/17: no updates
- 1/19: cstor related repos has ben migrated to github actions with minikube. So a similar approach needs to be used
- 2/16: successfully ran the NDM integration tests in minikube in github actions. need to find out a way to test built images before pushing. release scripts also need to be migrated to github actions
- 3/16: NDM, localPV, LVM localPV, zfslocalPV, upgrade, exporter, cstor-csi, cstor-operators, velero-plugin have all been moved from travis to github actions, including the integrations tests. Only the dataplane componenents are now remaining for migration
- check configmap is auto loaded without restarting pod(not started).
- 12/17: no updates
- 1/19: no updates
- 2/16: no updates
- 3/16: no updates
- implement used by probe for jiva(not yet started).
- 12/17: no updates
- 1/19: no updates
- 2/16: no updates
- 3/16: no updates
- add support for md array (not started)
- 12/17: no updates
- 1/19: no updates
- 2/16: no updates
- 3/16: no updates
- LocalPV, passing custom node selector to localpv-device (already done for localpv-hostpath)
- 2/16: no updates
- 3/16: no updates
- Split the openebs/charts to use the NDM helm charts
- Ref: https://github.com/openebs/openebs/discussions/3340
- 3/16: https://github.com/openebs/charts/pull/198. openebs/charts will be making use of the dependent charts only.
- Some additional functionalities of NDM is required to use it with CSI LocalPV variants for preparing the disks.
- remove CRD installation from NDM operator, upgrade CRDs from v1beta1 to v1

- [Harsh]
- display capacity.storage in correct units. (blocked)
- Can use resource.quantity from api-machinery (need to migrate CR to next version v1alpha1 to v1)
- 1/19: with CRD migration/validation done. This is now waiting on operator SDK upgrade. migrate BD and BDC to v1 / v1beta1 spec. ALso need to handle the conversion cases
- Upgrade to Operator SDK v1.0 - which is integrated into kube-builder (not started)
- sig-api => coverging on kube-builder and operatorsdk (follow this thread)
- The CR (BD and BDC) migration to v1 will be done after upgrading to operator SDK 1.0
- 1/19: no updates.
- 2/16: plan is to have operator SDK 1.0 (or latest version) for this release.
- 3/16: https://github.com/openebs/node-disk-manager/pull/552. operatorSDK mirgation and CRDs to v1

- explore on using polling in for size changes (in progress )
- did some initial research on what to use: explored inotify, and kmesg parser
- need to check the kmesg parser used in node problem detector
- 11/17: Explore the kmesg parser way of using as used in node problem detector
- 12/17: no updates
- 1/19: no updates. Can ask to sig-node team to get a clear idea on implementation.
- 2/16: Akhil will try to explore on this issue. need to modify the kmesg parser package and reuse it.
- 3/16: Submitted in Gsoc as part of CNCF mentoring.

- [Kiran] (no updates)
- Cleaning up of Stale NDM BDC CR in case of BD failed to bind and LocalPV-PVC gets deleted (not started)
- ownership / next release
- 3/16: @niladrih has taken up and is working on this.
- Handle the case for cleanup jobs getting scheduled to nodes when nodes is out of the cluster (not started)
- checkout cassandra and postgress operator
- Pass the RunAsUser option via the StorageClass
- Format type in StorageClass will be silently ignored if devices are already formatted. (not started)
- next release
- Refactor/review the errors displayed when BD is not bound due to a FS being present on BDs. (not started)
- Shared hostpath (not started) - PR will be closed. Functionality in k8s-sigs/nfs
- 2/16: dyna:ic NFS provisioner can be used now.
- custom node selectors for local pv (in-progress)
- need to do for device
- importing a localPV when cluster failure happens (not started)
- importing PV after etcd corruption. Make use of the same LocalPV. data disks after the cluster is recreated.
- https://github.com/kubernetes-retired/external-storage/issues/755
- (Case for create volume from data source - importing data from PV available on disk. Helper job or NDM).
- Follow up on sig-node inclusion of NDM project (not started)

## Feb 16, 2021

Attendees: Akhil, Harsh

- [Akhil]
- Add a note in NDM docs / OpenEBS install section about Maystor and ZFS Local PV not yet integrated with NDM. (not-started)
- doc for vritual disk expansion. partition also needs to be expanded manually (not started)
- ci can be used for staging. created an issue for integration testing in github actions. (work not started)
- 12/17: no updates
- 1/19: cstor related repos has ben migrated to github actions with minikube. So a similar approach needs to be used
- 2/16: successfully ran the NDM integration tests in minikube in github actions. need to find out a way to test built images before pushing. release scripts also need to be migrated to github actions
- check configmap is auto loaded without restarting pod(not started).
- 12/17: no updates
- 1/19: no updates
- 2/16: no updates
- implement used by probe for jiva(not yet started).
- 12/17: no updates
- 1/19: no updates
- 2/16: no updates
- add support for md array (not started)
- 12/17: no updates
- 1/19: no updates
- 2/16: no updates
- LocalPV, passing custom node selector to localpv-device (already done for localpv-hostpath)
- 2/16: no updates
- Split the openebs/charts to use the NDM helm charts
- Ref: https://github.com/openebs/openebs/discussions/3340

- [Harsh]
- display capacity.storage in correct units. (blocked)
- Can use resource.quantity from api-machinery (need to migrate CR to next version v1alpha1 to v1)
- 1/19: with CRD migration/validation done. This is now waiting on operator SDK upgrade. migrate BD and BDC to v1 / v1beta1 spec. ALso need to handle the conversion cases
- Upgrade to Operator SDK v1.0 - which is integrated into kube-builder (not started)
- sig-api => coverging on kube-builder and operatorsdk (follow this thread)
- The CR (BD and BDC) migration to v1 will be done after upgrading to operator SDK 1.0
- 1/19: no updates.
- 2/16: plan is to have operator SDK 1.0 (or latest version) for this release.

- explore on using polling in for size changes (in progress )
- did some initial research on what to use: explored inotify, and kmesg parser
- need to check the kmesg parser used in node problem detector
- 11/17: Explore the kmesg parser way of using as used in node problem detector
- 12/17: no updates
- 1/19: no updates. Can ask to sig-node team to get a clear idea on implementation.
- 2/16: Akhil will try to explore on this issue. need to modify the kmesg parser package and reuse it.

- [Kiran] (no updates)
- Cleaning up of Stale NDM BDC CR in case of BD failed to bind and LocalPV-PVC gets deleted (not started)
- ownership / next release
- Handle the case for cleanup jobs getting scheduled to nodes when nodes is out of the cluster (not started)
- checkout cassandra and postgress operator
- Pass the RunAsUser option via the StorageClass
-
- Format type in StorageClass will be silently ignored if devices are already formatted. (not started)
- next release
- Refactor/review the errors displayed when BD is not bound due to a FS being present on BDs. (not started)
- Shared hostpath (not started) - PR will be closed. Functionality in k8s-sigs/nfs
- 2/16: dyna:ic NFS provisioner can be used now.
- custom node selectors for local pv (in-progress)
- need to do for device
- importing a localPV when cluster failure happens (not started)
- importing PV after etcd corruption. Make use of the same LocalPV. data disks after the cluster is recreated.
- https://github.com/kubernetes-retired/external-storage/issues/755
- (Case for create volume from data source - importing data from PV available on disk. Helper job or NDM).
- Follow up on sig-node inclusion of NDM project (not started)

- Discussion
How NDM can help localpV: aim is to create volume groups (lvm localPV), zpool (zfs pv)
- can either be done via the gRPC API or a new controller in the provisioner, that watches for a CR and sends grpc messages to setup the node qith the devices.

## Jan 19, 2021

Attendees: Akhil, Harsh

- [Akhil]
- Add a note in NDM docs / OpenEBS install section about Maystor and ZFS Local PV not yet integrated with NDM. (not-started)
- doc for vritual disk expansion. partition also needs to be expanded manually (not started)
- ci can be used for staging. created an issue for integration testing in github actions. (work not started)
- 12/17: no updates
- 1/19: cstor related repos has ben migrated to github actions with minikube. So a similar approach needs to be used
- check configmap is auto loaded without restarting pod(not started).
- 12/17: no updates
- 1/19: no updates
- implement used by probe for jiva(not yet started).
- 12/17: no updates
- 1/19: no updates
- add support for md array (not started)
- 12/17: no updates
- 1/19: no updates
- add Unknown driveType to Blockdevice resource (done)
- add support for software RAID(done)
- LocalPV, passing custom node selector to localpv-device (already done for localpv-hostpath)

- [Harsh]
- display capacity.storage in correct units. (blocked)
- Can use resource.quantity from api-machinery (need to migrate CR to next version v1alpha1 to v1)
- 1/19: with CRD migration/validation done. This is now waiting on operator SDK upgrade. migrate BD and BDC to v1 / v1beta1 spec. ALso need to handle the conversion cases
- explore on using polling in for size changes (in progress )
- did some initial research on what to use: explored inotify, and kmesg parser
- need to check the kmesg parser used in node problem detector
- 11/17: Explore the kmesg parser way of using as used in node problem detector
- 12/17: no updates
- 1/19: no updates. Can ask to sig-node team to get a clear idea on implementation.
- Upgrade to Operator SDK v1.0 - which is integrated into kube-builder (not started)
- sig-api => coverging on kube-builder and operatorsdk (follow this thread)
- The CR (BD and BDC) migration to v1 will be done after upgrading to operator SDK 1.0
- 1/19: no updates.

- [Kiran] (no updates)
- Cleaning up of Stale NDM BDC CR in case of BD failed to bind and LocalPV-PVC gets deleted (not started)
- ownership / next release
- Handle the case for cleanup jobs getting scheduled to nodes when nodes is out of the cluster (not started)
- checkout cassandra and postgress operator
- Pass the RunAsUser option via the StorageClass
-
- Format type in StorageClass will be silently ignored if devices are already formatted. (not started)
- next release
- Refactor/review the errors displayed when BD is not bound due to a FS being present on BDs. (not started)
- Shared hostpath (not started) - PR will be closed. Functionality in k8s-sigs/nfs
- custom node selectors for local pv (in-progress)
- need to do for device
- importing a localPV when cluster failure happens (not started)
- importing PV after etcd corruption. Make use of the same LocalPV. data disks after the cluster is recreated.
- https://github.com/kubernetes-retired/external-storage/issues/755
- (Case for create volume from data source - importing data from PV available on disk. Helper job or NDM).
- Follow up on sig-node inclusion of NDM project (not started)

- Discussion
How NDM can help localpV: aim is to create volume groups (lvm localPV), zpool (zfs pv)
- can either be done via the gRPC API or a new controller in the provisioner, that watches for a CR and sends grpc messages to setup the node qith the devices.

## Dec 15, 2020

Attendees: Akhil, Harsh, Kiran

- [Akhil]
- Status for BD changes periodically (https://github.com/openebs/openebs/issues/3051). Handling case where bad disks are there on the node, and NDM can't partition them.(done, fixed in 1.0.1)
- Verify CRD migration and validation changes. (testing on the final changeset is required)
- Add a note in NDM docs / OpenEBS install section about Maystor and ZFS Local PV not yet integrated with NDM. (not-started)
- doc for vritual disk expansion. partition also needs to be expanded manually (not started)
- ci can be used for staging. created an issue for integration testing in github actions. (work not started)
- 12/17: no updates
- check configmap is auto loaded without restarting pod(not started).
- 12/17: no updates
- implement used by probe for jiva(not yet started).
- 12/17: no updates
- add support for md array (not started)
- 12/17: no updates
- add Unknown driveType to Blockdevice resource

- [Harsh]
- display capacity.storage in correct units. (blocked)
- Can use resource.quantity from api-machinery (need to migrate CR to next version v1alpha1 to v1)
- explore on using polling in for size changes (in progress )
- did some initial research on what to use: explored inotify, and kmesg parser
- need to check the kmesg parser used in node problem detector
- 11/17: Explore the kmesg parser way of using as used in node problem detector
- 12/17: no updates
- Upgrade to Operator SDK v1.0 - which is integrated into kube-builder (not started)
- sig-api => coverging on kube-builder and operatorsdk (follow this thread)
- The CR (BD and BDC) migration to v1 will be done after upgrading to operator SDK 1.0

- [Kiran]
- Cleaning up of Stale NDM BDC CR in case of BD failed to bind and LocalPV-PVC gets deleted (not started)
- ownership / next release
- Handle the case for cleanup jobs getting scheduled to nodes when nodes is out of the cluster (not started)
- checkout cassandra and postgress operator
- Pass the RunAsUser option via the StorageClass
-
- Format type in StorageClass will be silently ignored if devices are already formatted. (not started)
- next release
- Refactor/review the errors displayed when BD is not bound due to a FS being present on BDs. (not started)
- Shared hostpath (not started) - PR will be closed. Functionality in k8s-sigs/nfs
- custom node selectors for local pv (in-progress)
- done for hostpath
- need to do for device
- importing a localPV when cluster failure happens (not started)
- importing PV after etcd corruption. Make use of the same LocalPV. data disks after the cluster is recreated.
- https://github.com/kubernetes-retired/external-storage/issues/755
- (Case for create volume from data source - importing data from PV available on disk. Helper job or NDM).
- Follow up on sig-node inclusion of NDM project (not started)

- Backlogs
- South bound provisioner - shift from mayadata-io to openebs, refactor to integrate into NDM, Auto-scaling
- NDM - resize, partition, lvm in co-ordination with CSI driver

## Nov 17, 2020

Attendees: Akhil, Harsh, Kiran

- [Akhil]
- add support for device mapper and mpath(done)
- Status for BD changes periodically (https://github.com/openebs/openebs/issues/3051). Handling case where bad disks are there on the node, and NDM can't partition them.
- Verify CRD migration and validation changes. (spec validation might break some existing functionality)
- Add a note in NDM docs / OpenEBS install section about Maystor and ZFS Local PV not yet integrated with NDM. (not-started)
- doc for vritual disk expansion. partition also needs to be expanded manually (not started)
- ci can be used for staging. created an issue for integration testing in github actions. (work not started)
- check configmap is auto loaded without restarting pod(not started).
- implement used by probe for jiva(not yet started).
- add support for md array (not started)

- [Harsh]
- display capacity.storage in correct units. (blocked)
- Can use resource.quantity from api-machinery (need to migrate CRD to next version)
- explore on using polling in for size changes (in progress )
- did some initial research on what to use: explored inotify, and kmesg parser
- need to check the kmesg parser used in node problem detector
- 11/17: Explore the kmesg parser way of using as used in node problem detector
- Upgrade to Operator SDK v1.0 - which is integrated into kube-builder (not started)
- sig-api => coverging on kube-builder and operatorsdk (follow this thread)

- [Kiran]
- Cleaning up of Stale NDM BDC CR in case of BD failed to bind and LocalPV-PVC gets deleted (not started)
- Handle the case for cleanup jobs getting scheduled to nodes when nodes is out of the cluster (not started)
- Pass the RunAsUser option via the StorageClass
- Format type in StorageClass will be silently ignored if devices are already formatted. (not started)
- Shared hostpath (not started) - PR will be closed. Functionality in k8s-sigs/nfs
- custom node selectors for local pv (not started)
- importing a localPV when cluster failure happens
- importing PV after etcd corruption. Make use of the same LocalPV. data disks after the cluster is recreated.
- https://github.com/kubernetes-retired/external-storage/issues/755
- (Case for create volume from data source - importing data from PV available on disk. Helper job or NDM).
- Refactor/review the errors displayed when BD is not bound due to a FS being present on BDs. (not started)
- Follow up on sig-node inclusion of NDM project (not started)

## October 15, 2020

Attendees: Akhil, Harsh, Kiran

- [Akhil]
- -ci can be used for staging. created an issue for integration testing in github actions. (work not started)
- implement used by probe for jiva(not yet started).
- check configmap is auto loaded without restarting pod(not started).
- add additional check in NDM operator to check for block-device-tag (done)
- Add a note in NDM docs / OpenEBS install section about Maystor and ZFS Local PV not yet integrated with NDM. (not-started)
- config based tag for blockdevices (done)
- add unit test cases for partition handling(done)
- add support for device mapper, md array and mpath(in progress)
- doc for vritual disk expansion.- partition also needs to be expanded manually (not started)
- https://github.com/openebs/openebs/issues/3051
- Verify CRD migration and validation changes. (spec validation might break some existing functionality)

- [Harsh]
- Metrics (done)
- read/ write bytes
- latency (working)
- existing metrics from seachest is now exported.
- display capacity.storage in correct units. Can use resource.quantity from api-machinery (need to migrate CRD to next version)
- explore on using polling in for size changes (in progress )
- did some initial research on what to use: explored inotify, and kmesg parser
- need to check the kmesg parser used in node problem detector

- [Kiran]
- Format type in StorageClass will be silently ignored if devices are already formatted. (not started)
- Shared hostpath (not started) - PR will be closed. Functionality in k8s-sigs/nfs
- custom node selectors for local pv (not started)
- Cleaning up of Stale NDM BDC CR in case of BD failed to bind and LocalPV-PVC gets deleted (not started)
- Handle the case for cleanup jobs getting scheduled to nodes when nodes is out of the cluster (not started)
- Refactor/review the errors displayed when BD is not bound due to a FS being present on BDs. (not started)
- Follow up on sig-node inclusion of NDM project
- importing a localPV when cluster failure happens
- importing PV after etcd corruption. Make use of the same LocalPV. data disks after the cluster is recreated.
- https://github.com/kubernetes-retired/external-storage/issues/755

- Open Items:
- https://github.com/openebs/openebs/discussions/3139 ( blank for BD Tag)
- https://github.com/openebs/openebs/discussions/3140 (OpenEBSCtl for NDM)
- gRPC transcoded within REST and use the kubectl plugin (ndmctl) access ndm API via REST over K8s
- kubectl wrapper for executing commands within ndmctl shipped within ndm-operator
- Explore cloud endpoints - ESP for transcoding http/json <-> grpc
- anti-affnitiy for blockdevices and blockdevice claims- similar to scheduling operatorion, predicates work.
- south bound provisioner
- kubectl apply runtime transformation for CR fields, KEP raise for this.
- backstage(spotify) - NDM integration

## Sept 15, 2020
-----

## August 25, 2020

Attendees: Akhil, Kiran

- [Akhil]
- -ci can be used for staging. created an issue for integration testing in github actions. (work not started)
- implement used by probe for jiva(not yet started).
- check configmap is auto loaded without restarting pod(not started).
- add additional check in NDM operator to check for block-device-tag (in review)
- Add a note in NDM docs / OpenEBS install section about Maystor and ZFS Local PV not yet integrated with NDM. (not-started)
- config based tag for blockdevices (in review)
- add unit test cases for partition handling(in progress)
- add support for device mapper, md array and mpath(not started)
- doc for vritual disk expansion.- partition also needs to be expanded manually (not started)

- [Harsh]
- add grpc developer docs(done)
- Metrics (review)
- read/ write bytes
- latency (working)
- existing metrics from seachest is now exported.
- display capacity.storage in correct units. Can use resource.quantity from api-machinery (need to migrate CRD to next version)
- explore on using polling in for size changes

- [Kiran]
- Format type in StorageClass will be silently ignored if devices are already formatted. (not started)
- Shared hostpath (not started) - PR will be closed. Functionality in k8s-sigs/nfs
- custom node selectors for local pv (not started)

### Discussion Notes
- How to migrate CRDs in NDM.
- Suggested approach: Instead of using a conversion webhook which can be lot of work for a single NDM deployment, the NDM operator can use both CRDs at the same time, and then apply the conversion logic and save back to the server. Depending on whether to keep the old CRDs or not, can be decided depending if CRD deletion causes any issues.
- dmesg / polling for checking resize of disk.
- First phase: will manually trigger the rescan via grpc API, and then the partition will be expanded
- Second phase: instead of rescan a polling or event based mechanism can be used to auto trigger the operation
- moving away from operator sdk to client go
- Decided on to move away from using operator sdk in NDM. Instead will use native client-go
- Related refactoring related changes has to go in by 2.2 release.

## August 18, 2020

Attendees: Akhil, Harsh, Kiran, Prateek, Shubham

- [Akhil]
- -ci can be used for staging. created an issue for integration testing in github actions. (work not started)
- implement used by probe for jiva(not yet started).
- check configmap is auto loaded without restarting pod(not started).
- add additional check in NDM operator to check for block-device-tag (in review)
- Add a note in NDM docs / OpenEBS install section about Maystor and ZFS Local PV not yet integrated with NDM. (not-started)
- config based tag for blockdevices (in review)
- add unit test cases for partition handling(started)
- add support for device mapper, md array and mpath(not started)
- doc for vritual disk expansion.- partition also needs to be expanded manually

- [Harsh]
- add grpc developer docs(done)
- Metrics (review)
- read/ write bytes
- latency (working)
- existing metrics from seachest is now exported.
- display capacity.storage in correct units. Can use resource.quantity from api-machinery (need to migrate CRD to next version)
- explore on using polling in for size changes

- [Kiran]
- Format type in StorageClass will be silently ignored if devices are already formatted. (not started)
- Shared hostpath (not started) - PR will be closed. Functionality in k8s-sigs/nfs
- custom node selectors for local pv (not started)

### Discussion Notes
- How to migrate CRDs in NDM.
- Suggested approach: Instead of using a conversion webhook which can be lot of work for a single NDM deployment, the NDM operator can use both CRDs at the same time, and then apply the conversion logic and save back to the server. Depending on whether to keep the old CRDs or not, can be decided depending if CRD deletion causes any issues.
- dmesg / polling for checking resize of disk.
- First phase: will manually trigger the rescan via grpc API, and then the partition will be expanded
- Second phase: instead of rescan a polling or event based mechanism can be used to auto trigger the operation
- moving away from operator sdk to client go
- Decided on to move away from using operator sdk in NDM. Instead will use native client-go
- Related refactoring related changes has to go in by 2.2 release.

## August 11, 2020

Attendees: Akhil, Harsh

- [Akhil]
- -ci can be used for staging. create an issue for integration testing in github actions. (not started)
- implement used by probe for jiva(not yet started).
- adding the labels to blockdevice on zfs localPV and mayastor used disks (done).
- Support for upgrading to unique uuid (done)
- https://github.com/openebs/openebs/issues/3042
- 1.9 block devices with cstor. The file system written by ZFS is not visible till the next system reboot or sysfs rescan. (May get resolved with sampling probe).
- check configmap is auto loaded without restarting pod(not started).
- add additional check in NDM operator to check for block-device-tag (PR is up, need some furter changes)
- Add a note in NDM docs / OpenEBS install section about Maystor and ZFS Local PV not yet integrated with NDM. (not-started)
- config based tag for blockdevices (in progress)

- [Harsh]
- grpc API (done)
- implementation is almost done, working on unit tests. code review is pending
- running grpc server within ndm daemon (done)
- version, nodename, listblockdevicedetails (stats, hierarchy), iscsi status.
- rescan (will start after adding server to daemon)
- hugepages : setting hugepages is done. Restarting kubelet - have to explore more on approaches.

- add grpc developer docs
- Metrics (review)
- read/ write bytes
- latency (working)
- existing metrics from seachest is now exported.
- display capacity.storage in correct units.(in explore / also migrating to v1alpha2)

- [Kiran]
- Format type in StorageClass will be silently ignored if devices are already formatted. (not started)
- Shared hostpath (not started) - PR will be closed. Functionality in k8s-sigs/nfs
- custom node selectors for local pv (not started)

## August 04, 2020

Attendees: Akhil, Harsh, Kiran
PR reviews

## July 28, 2020

Attendees: Akhil, Harsh, Kiran

- [Akhil]
- -ci can be used for staging. create an issue for integration testing in github actions. (not started)
- implement used by probe for jiva(not yet started).
- adding the labels to blockdevice on zfs localPV and mayastor used disks (in review).
- Support for upgrading to unique uuid (in progress)
- https://github.com/openebs/openebs/issues/3042
- 1.9 block devices with cstor. The file system written by ZFS is not visible till the next system reboot or sysfs rescan. (May get resolved with sampling probe).
- check configmap is auto loaded without restarting pod(not started).
- add additional check in NDM operator to check for block-device-tag (PR is up, need some furter changes)
- Add a note in NDM docs / OpenEBS install section about Maystor and ZFS Local PV not yet integrated with NDM. (not-started)

- [Harsh]
- grpc API (in review)
- implementation is almost done, working on unit tests. code review is pending
- running grpc server within ndm daemon (not started)
- version, nodename, listblockdevicedetails (stats, hierarchy), iscsi status.
- rescan (will start after adding server to daemon)
- hugepages : setting hugepages is done. Restarting kubelet - have to explore more on approaches.

- add grpc developer docs
- Metrics (started)
- read/ write bytes
- latency (working)
- existing metrics from seachest is now exported.
- display capacity.storage in correct units.(in explore / also migrating to v1alpha2)

- [Kiran]
- Format type in StorageClass will be silently ignored if devices are already formatted. (not started)
- Shared hostpath (not started) - PR will be closed. Functionality in k8s-sigs/nfs
- custom node selectors for local pv (not started)

### Discussion Topics

- Review Project Tracker: https://github.com/orgs/openebs/projects/2

## July 21, 2020

Attendees: Akhil, Harsh, Kiran

- [Akhil]
- -ci can be used for staging. create an issue for integration testing in github actions. (not started)
- implement used by probe for jiva(not yet started).
- adding the labels to blockdevice on zfs localPV and mayastor used disks (in progress).
- Support for upgrading to unique uuid (in progress)
- https://github.com/openebs/openebs/issues/3042
- 1.9 block devices with cstor. The file system written by ZFS is not visible till the next system reboot or sysfs rescan. (May get resolved with sampling probe).
- check configmap is auto loaded without restarting pod(not started).
- add additional check in NDM operator to check for block-device-tag (PR is up, need some furter changes)
- codecov fix for removing token and adding unit test to github actions
- Add a note in NDM docs / OpenEBS install section about Maystor and ZFS Local PV not yet integrated with NDM. (not-started)

- [Harsh]
- update go version(done)
- grpc API (in review)
- implementation is almost done, working on unit tests. code review is pending
- running grpc server within ndm daemon (not started)
- version, nodename, listblockdevicedetails (stats, hierarchy), iscsi status.
- rescan (will start after adding server to daemon)
- hugepages : setting hugepages is done. Restarting kubelet - have to explore more on approaches.

- add grpc developer docs
- Metrics (started)
- read/ write bytes
- latency (working)
- existing metrics from seachest is now exported.

- [Kiran]
- Format type in StorageClass will be silently ignored if devices are already formatted. (not started)
- Shared hostpath (not started) - PR will be closed. Functionality in k8s-sigs/nfs
- custom node selectors for local pv (not started)

### Discussion Topics

- Review Project Tracker: https://github.com/orgs/openebs/projects/2

## July 14, 2020

Attendees: Akhil, Harsh, Kiran

- [Akhil]
- -ci can be used for staging. create an issue for integration testing in github actions. (not started)
- implement used by probe for jiva(not yet started).
- adding the labels to blockdevice on zfs localPV and mayastor used disks (in progress).
- Support for upgrading to unique uuid (in progress)
- https://github.com/openebs/openebs/issues/3042
- 1.9 block devices with cstor. The file system written by ZFS is not visible till the next system reboot or sysfs rescan. (May get resolved with sampling probe).
- check configmap is auto loaded without restarting pod(not started).
- fixing the cleanup issue for cstor. issue when `wipefs` is called on cstor (done)
- add additional check in NDM operator to check for block-device-tag (PR is up, need some furter changes)
- codecov fix for removing token and adding unit test to github actions

- [Harsh]
- update go version(in review)
- remove gox (done)
- make rescan functionality thread safe(done)
- grpc API (in review)
- implementation is almost done, working on unit tests. code review is pending
- running grpc server within ndm daemon (not started)
- version, nodename, listblockdevicedetails (stats, hierarchy), iscsi status.
- rescan (will start after adding server to daemon)
- hugepages : setting hugepages is done. Restarting kubelet - have to explore more on approaches.

- updating the NDM developer docs (done)
- new BUILD.md file for building anf installing NDM(done).
- Metrics (started)
- read/ write bytes
- latency (working)
- existing metrics from seachest is now exported.

- [Kiran]
- Format type in StorageClass will be silently ignored if devices are already formatted. (not started)
- Shared hostpath (not started)
- custom node selectors for local pv (not started)
- Add a note in NDM docs / OpenEBS install section about Maystor and ZFS Local PV not yet integrated with NDM. (not-started)
- code cov

### Discussion Topics

- Review Project Tracker: https://github.com/orgs/openebs/projects/2

## July 7, 2020

Attendees: Akhil, Harsh, Kiran
PR walkthrough

## Jun 30, 2020

Attendees: Akhil, Harsh

- [Akhil]
- -ci can be used for staging. create an issue for integration testing in github actions. (not started)
- NDM events PR (done)
- Implement sampling probe (done)
- Add a sampling probe to determine if the block device is being used by cStor, Mayastor or even other well known userspace filesystems
- Looking at reading the super block to determine the type of fileystem on the BD
- Detect Mayastor or ZFS and add it as a discovered property
- While saving in etcd, add the discovered property in annotation, possibly add it as a block-device-tag
- implement used by probe for jiva(not yet started).
- adding the labels to blockdevice on zfs localPV and mayastor used disks (in review).
- Support for upgrading to unique uuid (in review)
- https://github.com/openebs/openebs/issues/3042
- 1.9 block devices with cstor. The file system written by ZFS is not visible till the next system reboot or sysfs rescan. (May get resolved with sampling probe).
- check configmap is auto loaded without restarting pod(not started).
- fixing the cleanup issue for cstor. issue when `wipefs` is called on cstor (started)

- [Harsh]
- grpc API (in review)
- implementation is almost done, working on unit tests. code review is pending
- running grpc server within ndm daemon (not started)
- version, nodename, listblockdevicedetails (stats, hierarchy), iscsi status.
- rescan (will start after adding server to daemon)
- hugepages : setting hugepages is done. Restarting kubelet - have to explore more on approaches.
- updating the NDM developer docs (review)
- new BUILD.md file for building anf installing NDM(review).
- Metrics (started)
- read/ write bytes
- latency (working)

- [Kiran]
- Format type in StorageClass will be silently ignored if devices are already formatted. (not started)
- Shared hostpath (not started)
- custom node selectors for local pv (not started)
- Add a note in NDM docs / OpenEBS install section about Maystor and ZFS Local PV not yet integrated with NDM. (not-started)
- code cov

### Discussion Topics

- Review Project Tracker: https://github.com/orgs/openebs/projects/2

## Jun 23, 2020

Attendees: Akhil, Harsh

- [Akhil]
- Multi-arch build (done)
- -ci can be used for staging. create an issue for integration testing in github actions. (not started)
- NDM events PR (in-review)
- Implement sampling probe (done)
- Add a sampling probe to determine if the block device is being used by cStor, Mayastor or even other well known userspace filesystems
- Looking at reading the super block to determine the type of fileystem on the BD
- Detect Mayastor or ZFS and add it as a discovered property
- While saving in etcd, add the discovered property in annotation, possibly add it as a block-device-tag
- implement used by probe for jiva(not yet started).
- adding the labels to blockdevice on zfs localPV and mayastor used disks (progress).
- Support for upgrading to unique uuid (in progress)
- https://github.com/openebs/openebs/issues/3042
- 1.9 block devices with cstor. The file system written by ZFS is not visible till the next system reboot or sysfs rescan. (May get resolved with sampling probe).
- check configmap is auto loaded without restarting pod.
- Metrics (not started)

- [Harsh]
- -o wide should show the filesystem (done)
- grpc API (in progress)
- implementation is almost done, working on unit tests. code review is pending
- running grpc server within ndm daemon (not started)
- version, nodename, listblockdevicedetails (stats, hierarchy), iscsi status.
- rescan (will start after adding server to daemon)
- hugepages : setting hugepages is done. Restarting kubelet - have to explore more on approaches.
- updating the NDM developer docs (started)
- new BUILD.md file for building anf installing NDM.

- [Kiran]
- Format type in StorageClass will be silently ignored if devices are already formatted. (not started)
- Shared hostpath (not started)
- custom node selectors for local pv (not started)
- Add a note in NDM docs / OpenEBS install section about Maystor and ZFS Local PV not yet integrated with NDM. (not-started)
- code cov

### Discussion Topics

- Review Project Tracker: https://github.com/orgs/openebs/projects/2

## Jun 16, 2020

Attendees: Akhil, Harsh

- [Akhil]
- Multi-arch build (in-progress)
- NDM events PR (in-review)
- Implement sampling probe (done)
- Add a sampling probe to determine if the block device is being used by cStor, Mayastor or even other well known userspace filesystems
- Looking at reading the super block to determine the type of fileystem on the BD
- Detect Mayastor or ZFS and add it as a discovered property
- While saving in etcd, add the discovered property in annotation, possibly add it as a block-device-tag
- implement used by probe for jiva(not yet started).
- adding the labels to blockdevice on zfs localPV and mayastor used disks (progress).
- Support for upgrading to unique uuid (in progress)
- https://github.com/openebs/openebs/issues/3042
- 1.9 block devices with cstor. The file system written by ZFS is not visible till the next system reboot or sysfs rescan. (May get resolved with sampling probe).
- check configmap is auto loaded without restarting pod.
- Metrics (not started)

- [Harsh]
- -o wide should show the filesystem (done)
- grpc API (in progress)
- implementation is almost done, working on unit tests. code review is pending
- running grpc server within ndm daemon (not started)
- version, nodename, listblockdevicedetails (stats, hierarchy), iscsi status.
- rescan (will start after adding server to daemon)
- hugepages : setting hugepages is done. Restarting kubelet - have to explore more on approaches.
- updating the NDM developer docs (started)

- [Kiran]
- Format type in StorageClass will be silently ignored if devices are already formatted. (not started)
- Shared hostpath (not started)
- custom node selectors for local pv (not started)
- Add a note in NDM docs / OpenEBS install section about Maystor and ZFS Local PV not yet integrated with NDM. (not-started)
- Add a release note about avoiding updates to claimed block devices. useful for cleaning. Local PV device clearing will not happen. (done)
- code cov

### Discussion Topics

- Review Project Tracker: https://github.com/orgs/openebs/projects/2

## Jun 9, 2020

Attendees: Akhil, Kiran

### Status Updates

- [Akhil]
- PR on builds/shell check (done)
- Follow-up on move towards go-modules for NDM (done)
- Multi-arch build (in-progress)
- NDM events PR (in-review)
- Implement sampling probe (done)
- Add a sampling probe to determine if the block device is being used by cStor, Mayastor or even other well known userspace filesystems
- Looking at reading the super block to determine the type of fileystem on the BD
- Detect Mayastor or ZFS and add it as a discovered property
- While saving in etcd, add the discovered property in annotation, possibly add it as a block-device-tag
- implement used by probe for jiva.
- adding the labels to blockdevice on zfs localPV and mayastor used disks.
- Support for upgrading to unique uuid (in progress)
- https://github.com/openebs/openebs/issues/3042
- 1.9 block devices with cstor. The file system written by ZFS is not visible till the next system reboot or sysfs rescan. (May get resolved with sampling probe).

- -o wide should show the filesystem (TODO harsh)
- check configmap is auto loaded without restarting pod.

- Metrics (not started)

- [Kiran]
- Format type in StorageClass will be silently ignored if devices are already formatted. (not started)
- Shared hostpath (not started)
- custom node selectors for local pv (not started)
- Add a note in NDM docs / OpenEBS install section about Maystor and ZFS Local PV not yet integrated with NDM. (not-started)
- Add a release note about avoiding updates to claimed block devices. useful for cleaning. Local PV device clearing will not happen. (done)
- code cov

### Discussion Topics

- Review Project Tracker: https://github.com/orgs/openebs/projects/2

## Jun 2, 2020

Attendees: Akhil, Kiran

### Status Updates

- [Akhil]
- PR on builds/shell check (in-review)
- Follow-up on move towards go-modules for NDM (in-review)
- Multi-arch build (in-progress)
- NDM events PR (in-review)
- Implement sampling probe (coding-in-progress)
- Add a sampling probe to determine if the block device is being used by cStor, Mayastor or even other well known userspace filesystems
- Looking at reading the super block to determine the type of fileystem on the BD
- Detect Mayastor or ZFS and add it as a discovered property
- While saving in etcd, add the discovered property in annotation, possibly add it as a block-device-tag
- Support for upgrading to unique uuid (in progress)
- https://github.com/openebs/openebs/issues/3042
- 1.9 block devices with cstor. The file system written by ZFS is not visible till the next system reboot or sysfs rescan. (May get resolved with sampling probe).

- Metrics (not started)

- [Kiran]
- Format type in StorageClass will be silently ignored if devices are already formatted. (not started)
- Shared hostpath (not started)
- custom node selectors for local pv (not started)
- Add a note in NDM docs / OpenEBS install section about Maystor and ZFS Local PV not yet integrated with NDM. (not-started)
- Add a release note about avoiding updates to claimed block devices. useful for cleaning. Local PV device clearing will not happen. (not started)
- code cov

### Discussion Topics

- Review Project Tracker: https://github.com/orgs/openebs/projects/2

## May 19, 2020

Attendees: Akhil, Kiran

### Status Updates

- [Akhil]
- Helm charts for 1.10 with option to add NDM feature gates (done)
- Implement sampling probe (design-in-progress)
- Add a sampling probe to determine if the block device is being used by cStor, Mayastor or even other well known userspace filesystems
- Looking at reading the super block to determine the type of fileystem on the BD
- Detect Mayastor or ZFS and add it as a discovered property
- While saving in etcd, add the discovered property in annotation, possibly add it as a block-device-tag
- Support for upgrading to unique uuid (in progress)
- https://github.com/openebs/openebs/issues/3042
- 1.9 block devices with cstor. The file system written by ZFS is not visible till the next system reboot or sysfs rescan. (May get resolved with sampling probe).
- Follow-up on move towards go-modules for NDM (in-progress)
- Metrics (not started)
-

- [Kiran]
- Format type in StorageClass will be silently ignored if devices are already formatted. (not started)
- Shared hostpath (not started)
- custom node selectors for local pv (not started)
- Add a note in NDM docs / OpenEBS install section about Maystor and ZFS Local PV not yet integrated with NDM. (not-started)
- Add a release note about avoiding updates to claimed block devices. useful for cleaning. Local PV device clearing will not happen. (not started)

### Discussion Topics

- Review Project Tracker: https://github.com/orgs/openebs/projects/2

## May 12, 2020

Attendees: Akhil, Kiran

### Status Updates

- [Akhil]
- Finalizers for claimed BDs (done)
- Helm charts for 1.10 with option to add NDM feature gates (review)
- Support for upgrading to unique uuid (in progress)
- 1.9 block devices with cstor. The file system written by ZFS is not visible till the next system reboot or sysfs rescan.
- Can NDM issue a rescan prior to initial discover/probe?
- Add a sampling probe to determine if the block device is being used by cStor, Mayastor or even other well known userspace filesystems
- Metrics (not started)

- [Kiran]
- Format type in StorageClass will be silently ignored if devices are already formatted. (not started)
- Shared hostpath (not started)
- custom node selectors for local pv (not started)

### Discussion Topics

- Review Project Tracker: https://github.com/orgs/openebs/projects/2

## May 05, 2020

Attendees: Akhil, Kiran

### Status Updates

- [Akhil]
- Exclude openebs volumes from discovery by default (done)
- Invalid BD entries are created (done)
- Partition UUID changes (done)
- Finalizers for claimed BDs (not started)
- Helm charts for 1.10 with option to add NDM feature gates
- Support for upgrading to unique uuid (in progress)
- 1.9 block devices with cstor. The file system written by ZFS is not visible till the next system reboot or sysfs rescan.
- Can NDM issue a rescan prior to initial discover/probe?
- Add a sampling probe to determine if the block device is being used by cStor, Mayastor or even other well known userspace filesystems
- Metrics (not started)

- [Kiran]
- Issue with hostpath volume not getting created when nodeName is used (done)
- Format type in StorageClass will be silently ignored if devices are already formatted. (not started)
- Shared hostpath (not started)
- custom node selectors for local pv (not started)

### Discussion Topics

- Review Project Tracker: https://github.com/orgs/openebs/projects/2
- Prioritize the finalizer on BD
- Clean up and sort the near term goals

## April 28, 2020

Attendees: Akhil, Giri, Kiran

### Status Updates

- [Akhil]
- BD/BDC with namespace scope (done)
- Exclude openebs volumes from discovery by default (Needs review)
- Invalid BD entries are created (Working on it)
- Partition UUID changes (Needs Review)

- [Kiran]
- Local PV with custom node selector (Not started)
- Local PV hostpath fails when application spec has nodeName specified (Started)
- Local PV hostpath and device user guides (done)

- [Giri]
- E2e for Local PV with Block Device Tags (In-progress)
-

### Discussion Topics

- [Giri]
- Functional Testing on Local PV
- Provisioning of Local PV hostpath and with Device. Verify setting up and cleaning of various resources (done)
- Provisioning Block Device with XFS (instead of default Ext4)
- Running Percona/MySQL on Raw Block Device with Local PV Device
- Running Cassandra on Raw Block Device with Local PV Device
- Running MongoDB on dedicated block devices using tagging feature.
- Reusing a block device for different applications - by changing the filesystem format. Percona with Ext4 and then run MongoDB with XFS
- Backup and Restore of MinIO using Local PV hostpath and device
- Application pods with Cross AZ functionality
- Chaos Testing on Local PV
- Local PV Provisioner, NDM operator and daemon set chaos (podkill) during provisioning and deletion of PVCs.
- Application pod kill
- Node Failure/Restart
- Node Delete (or changing the hostname of the node) and handle the graceful application pod rescheduling/creating new pod.
- Scale Testing on Local PV
- Test for capacity threshold crossing
- Test for underlying device running out of capacity
- Test for 50, 100, 200 volumes
- Soak Testing for Local PV
- Setup long running clusters - soak test bed
- Cassandra on hostpath and Kafka and mongo on device
- Use data generators to keep pumping data
- Introduce occasional choas as part of build/release pipelines into the soaktest bed cluster on these applications.