[FEATURE]Native RWX with Longhorn volume

[yasker]

We want to support RWX on Longhorn. Probably using NFS protocol in someway.

[SoutisV]

With the current nfs-provisioner it takes around 90sec for the failover. Would be supported in v1.1.0 a HA setup for RWX?

[yasker]

@SoutisV Unfortunately it's not going to be HA RWX. It's very hard to do zero downtime HA RWX with Longhorn which is designed to be a distributed block storage system instead of a distributed filesystem.

[longhorn-io-github-bot]

Pre-merged Checklist

• Does the PR include the explanation for the fix or the feature?

• Is the backend code merged (Manager, Engine, Instance Manager, BackupStore etc)?
  The PR is at
  1470 - single volume from fork longhorn-share-manager#3
  1470 rwx single volume longhorn-manager#778
  Transfer share manager ownership when a node is being drained longhorn-manager#791

• Is the reproduce steps/test steps documented?

• Which areas/issues this PR might have potential impacts on?
  Area: csi, volumes, remount, node failure
  Issues

• If the fix introduces the code for backward compatibility Has an separate issue been filed with the label release/obsolete-compatibility?
  The compatibility issue is filed at

• If labeled: area/ui Has the UI issue filed or ready to be merged?
  The UI issue/PR is at
  [FEATURE] Option of creating RWX volume from the Longhorn UI #2048
  [BUG] Update Access mode should be disabled if volume is already bound to a PV/PVC #2094

• if labeled: require/doc Has the necessary document PR submitted or merged?
  The Doc issue/PR is at Add documentation for native rwx workloads website#240

• If labeled: require/automation-e2e Has the end-to-end test plan been merged? Have QAs agreed on the automation test case?
  The automation skeleton PR is at
  The automation test case PR is at

• if labeled: require/automation-engine Has the engine integration test been merged?
  The engine automation PR is at

• if labeled: require/manual-test-plan Has the manual test plan been documented?
  The updated manual test plan is at Add manual test for RWX feature longhorn-tests#496

[zbialik]

Is there really no plan to make the longhorn-nfs-provisioner available in a HA setup? Having a single point of failure for a shared persistent volume for my RWX workloads is not ideal. A 90 second fail over is a deal breaker.

If its the longhorn storage-class that is limiting the longhorn-nfs-provisioner from replicating the storage quick enough for zero downtime, could another storage-class be leveraged? I know that the Service ClusterIP gets hardcoded and the docs say it must be linked to a single pod, but is there any reason why this Service ClusterIP couldn't front 2 or 3 pods?

[yasker]

@zbialik It's complicated to get a HA NFS working for a block storage system like Longhorn. The cluster IP can front multiple Pods, but those Pods must have the same data to keep it consistent across the board. Also, each NFS server has an internal state machine that needs to keep consistent across the whole setup. And in the case of losing one NFS server, we need to handle the traffic on the fly as well to prevent data corruption.

We've spent a lot of effort trying to piece everything together, but unfortunately, we still face technical difficulties to get it right in this release. We will investigate more after the v1.1.0 release.

[zbialik]

Thanks for the explanation!

[khushboo-rancher]

Verified the RWX feature with Longhorn v1.1.0-rc1

Validation - Pass

Executed test cases from longhorn/longhorn-tests#496.

[khushboo-rancher]

Closing this as the validation is completed.

[taxilian]

The documentation is still pointing to old information; I'm currently trying to figure out how to do this, hoping the lack of docs means to just use the flag?

[yardenshoham]

@taxilian it looks updated to me. https://longhorn.io/docs/1.1.0/advanced-resources/rwx-workloads/

[taxilian]

Huh; maybe I was looking at an old version of the docs and didn't realize it. Thanks for the link! Richard

…

On Sat, Jan 2, 2021 at 1:47 AM Yarden Shoham ***@***.***> wrote: @taxilian <https://github.com/taxilian> it looks updated to me. https://longhorn.io/docs/1.1.0/advanced-resources/rwx-workloads/ — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#1470 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AABWYTW2W6UQJ5GXX7M3G53SX3MRTANCNFSM4NYUHLOQ> .