Persistent Volume of Type NFS Share cannot be created with access mode "Many Nodes Read-Write" (aka ReadWriteMany / RWX)

[Remigius2011]

Rancher versions:
rancher/server or rancher/rancher: v2.0.0-beta4
rancher/agent or rancher/rancher-agent: v2.0.0-beta4

Infrastructure Stack versions:
healthcheck:
ipsec:
network-services:
scheduler:
kubernetes (if applicable): v1.10.1-rancher2

Docker version: (docker version,docker info preferred) 17.03.2-ce

Operating system and kernel: (cat /etc/os-release, uname -r preferred) ubuntu 16.04.04 4.4.0-119-generic

Type/provider of hosts: (VirtualBox/Bare-metal/AWS/GCE/DO) VMWare Workstation 14 Pro 14.1.1 build-7528167

Setup details: (single node rancher vs. HA rancher, internal DB vs. external DB) single node rancher / single node k8s cluster

Environment Template: (Cattle/Kubernetes/Swarm/Mesos)

Steps to Reproduce:

• add new persistent volume
• enter name "test"
• select volume plugin "NFS Share"
• enter valid data for path and server
• check only access mode "Many Nodes Read-Write"
• click "save"

Results:

The following error message appears:

"Validation failed in API: PersistentVolume "test" is invalid: spec.accessModes: Required value"

[Remigius2011]

Follow-Up: Sometimes it seems to work (e.g. after logout / login). Also, creating a PVC under Volumes fails when only "Many Nodes Read-Write" is checked.

[tfiduccia]

@Remigius2011 - I believed we have fixed this in the latest rancher/rancher:master. If you see this problem in our next release, please reopen.

[Remigius2011]

@tfiduccia still the same behavior in GA - sorry, I can't reopen the issue. Actually, some of my (admittedly rather mickey mouse type) workloads require access to a shared storage on a NAS and the best solution I could so far come up with is an NFS share. RWX is convenient because it avoids having to create an inflationary quantity of volumes. As long as this isn't fixed I'll have to stick with my swarm cluster.

[nathanweeks]

I'm also seeing the same error in rancher/rancher:2.0.0

[tfiduccia]

Version - v2.0.0
Was finally able to reproduce - had to create a few Persistent Volumes before it would error. Just create 3 and you should hit the error. Make sure for all of them, in Customize, the Many Nodes Read-Write is the only Access Mode checked.

[Remigius2011]

@tfiduccia exactly (and see also my instructions), if you select another mode, it will not complain, as another mode satisfies the validation requirements and you will only notice a problem when you try to create a claim with RWX.

[tfiduccia]

Version - 2.0 master 5/8
Verified fixed

@Remigius2011 & @nathanweeks - The issue is fixed in our latest master build and will be going into the 2.0.1 build.

[Remigius2011]

I can confirm it is working in v2.0.1-rc4 / k8s v1.10.1. I can creeate a volume and an RWX volume claim and access it from at least two pods (one read-only, the other writable).