Closed CharlieLaughton closed 6 years ago
If you reboot the head node everything falls apart, as the /shared filesystem does not automatically remount. As a result the task scheduler cannot find the dsf.json file.
This needs a mount command added to fstab at the when the cluster is launched. This can be added to the user data that the VM receives when booting.
This has been addressed and fixed in the commit 4ddca47.
If you reboot the head node everything falls apart, as the /shared filesystem does not automatically remount. As a result the task scheduler cannot find the dsf.json file.