We’ve combined OpenNebula with Parallels on MacOS. However, we tend to get random timeouts executing run_monitord_client on the Macs. The processes exists when listing all processes, so apparently they get stuck.
Running run_monitord_client manually to see what happens doesn’t fly, since it always just stops at "STDIN=cat -. Figured that one by adding “set -ex” at the start of the script and running it manually.
How is that script supposed to work and how can I debug the problem of the frozen process? Give it enough time and it will proceed and at some point monitoring will resume as normal and the hosts come back to ON from being in ERROR state.
Running OpenNebula 6.0.0.2, but at least that run_monitord_client is still the same in 6.6.0.
Could you explain a bit more how you are trying to use OpenNebula? On what device are you running OpenNebula (some virtualized OS using Parallels)? What are the target hypervisor Hosts (MacOS)?
We have done our own Parallels drivers, that are pretty much based on existing qcow drivers. OpenNebula is running safely on a Ubuntu 22.04 controlling a myriad of rack servers which run Windows and different Linux distros in Virtual machines via KVM on Dell rack servers. But the case with MacOS is a bit different. Here we have a fleet of MacMinis which naturally can’t run KVM. So here Parallels comes into play. So we run MacOS on MacMinis with Parallels taking care of the hypervisor part to run MacOS in virtual machines.
The problem we’re facing is why the host monitoring goes and freezes itself when we monitor the MacMini hardware.