Description of problem:
The North American SA group runs a cloudforms instance with hooks into 3 different datacenters (as well as EC2). One of the datacenters went offline. The NFS mount when unaccessable and problems started. I didn't do a full test suite, but I know programs that tried to stat the filesystem list would hang.
Unfortunately when it happened I was more interested in getting df and other system commands to return, so I forcibly umounted it with 'umount -f'.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. Add a Cloud Resource Provider (rhev based, with an NFS backed export domain)
2. Interrupt communication between CF & RHEV (turn off NFS for example)
Some way for CF to detect the CRP has gone offline and/or taken approprite actions to umount or disable the NFS store.
I'll admit I may not fully understand the full workflow of the product still. But I feel this may need at least another revisit to see if there's a better way to handle this exception.
This is for hard NFS mounts. I went in and manually disabled the Cloud Resource Provider.
RHEVM Cloud Resource Provider, that is