I have been told about a strange scenario on one of our deployed systems. A networking issue caused an outage isolating some hosts from their storage for a period of time.
When the hardware was recovered and the workloads were being started up, most was recovered OK, but some VMs couldn't be started because some files were locked. Most of those were recovered in one way or another, two are left unresolved.
Our systems are highly customised, there is no user access to things such as vCenter, DCUI, commandlines, PowerCLI. SSH and the like. All anticipated work flows are provided for via custom system management software and installation procedures. The installation procedures and scripts worked to recover most of the VMs, by deleting the old ones and deploying new ones.
The last couple of VMs cannot be deleted using the provided tools for some reason. An engineer was sent to have a quick look and reported that NVRAM files for these VMs are locked and therefore cannot be deleted.
We want to improve the management software to recover from this situation, but in order to do this development and test it, we need a procedural way to create similar file locks without resorted to randomly 'breaking' things to try and achieve it.
So the question is: Can a file lock be put into place manually somehow? If so how?