I've followed this article to setup a NFS server on GKE so that I can mount a ReadWriteMany pvc on many pods.

These pods are jenkins-slave agents that are created and destroyed continuously as needed.

The point is sometimes I can see several pods mounting the same pvc against the nfs server, working fine. However very often there are a bunch of pods stuck in Pending state. Describing them I see:

Unable to mount volumes for pod "jenkins-slave-rqsb7_mynamespace(044f1362-5925-4812-875c-723f32d5e6fc)": timeout expired waiting for volumes to attach or mount for pod "mynamespace"/"jenkins-slave-rqsb7". list of unmounted volumes=[volume-2]. list of unattached volumes=[volume-0 volume-2 volume-1 workspace-volume default-token-qdlhx]

Where volume-2 is the one corresponding to the pvc agains the NFS backed volume.

What could be the problem here?