A couple of months ago when upgrading one of my Nutanix clusters to vSphere 5.5 U1 I started to see what appeared to be random loss of connection to my NFS datastores. All VMs on the datastores would become inaccessible, then a few minutes later, access would be restored. As it turns out, this All Path Down (APD) bug introduced in vSphere 5.5 U1 and affected most any storage vendor using NFS. VMware wrote a KB article about the problem, which you can read here.
I am now pleased to announce that with vSphere 5.5 express patch 04, VMware has said they resolved the issue. You can download the patch here. I haven’t yet tried out the patch, but I’m very glad it is now out. For those of you still on vSphere 5.5 GA and using NFS, I would encourage you in a *lab* environment to thoroughly test express patch 04 and validate for your environment the issue has been resolved.
A big thanks to VMware for staying on top of this issue and coming out with a patch. I’ll be testing this in my lab very shortly.